diff --git "a/src/all_text.txt" "b/src/all_text.txt"
new file mode 100644--- /dev/null
+++ "b/src/all_text.txt"
@@ -0,0 +1 @@
+1 Disproportionality in Misconduct Cases in the Metropolitan Police Service January 2023 Emanus Prospere, Rachael Parker, Chloe Iliesa and Zoe Hobson MOPAC Evidence and Insight Appendix B Contents Executive Summary ..................................................................................................................... 4 Introduction ................................................................................................................................ 6 Methodology .............................................................................................................................. 8 Results ........................................................................................................................................ 9 Literature Review ........................................................................................................................ 9 In-depth analysis of MPS misconduct data ............................................................................... 15 Vignette Study .......................................................................................................................... 24 Discussion ................................................................................................................................. 27 References ................................................................................................................................ 31 Appendix 1 – Vignette Example ................................................................................................ 33 Appendix 2 – Regression Tables ................................................................................................ 34 Background Since 2016 MOPAC Evidence & Insight (E&I) has been undertaking research exploring the issue of racial disparity within the police misconduct system and has previously reported that police officers from Black and/or other Minority Ethnic communities were twice as likely as white officers to receive an allegation of misconduct within the Metropolitan Police Service (MPS). This is a critical area for the Mayor and MOPAC – the Mayor’s Action Plan to improve transparency, accountability, and trust in the Metropolitan Police Service was published in November 2020 which included a commitment by the MPS to continue to significantly reduce disproportionality within the misconduct processes by 2024. This issue has further been brought to the fore by the recent Casey report on the police misconduct system (2022) which also identified this concern. This has formed a growing body of evidence on the over-representation of staff from Black and/or other Minority Ethnic communities in MPS disciplinary proceedings, yet less is known on the specific factors driving the disproportionality. This report combines research from two phases – Phase 1 was completed in August 2021 and included a review of the key literature, updated the previous analytics conducted, as well as a statistical attempt to explore the drivers of this disproportionality. Phase 2 conducted a vignette experiment to explore the disproportionality in misconduct decision-making processes by MPS supervisors. This was completed in September 2022. While it is important that the MPS has the power to apply disciplinary procedures to ensure staff behave in a professional manner, the evidence to date indicates that disciplinary procedures are disproportionately applied to police officers from Black and/or other Minority Ethnic communities. These issues are not new and there has been no significant change in the overall disproportionality gap since the original MOPAC analysis was conducted in 2016. The updated analysis demonstrates disproportionality is most stark for male, short-service police officers from Black and/or other Minority Ethnic communities who are working in front-line roles, and that disparity starts at the beginning of the misconduct process when officers are first referred into the misconduct system by line managers and supervisors. The disparity was also found to be specific to the misconduct process and not replicated in other reporting mechanisms. Findings from the vignette study pointed to a lack of training, as well as inconsistency and subjectivity in decision-making in relation to officer conduct, at the stage where a supervisor decides whether a scenario is suitable for the formal misconduct process or not. This suggests a greater need for enhanced training and the need for more consistency in training for supervisors on this topic. Key Findings Disproportionality in Misconduct Cases in the Metropolitan Police Service Executive Summary More detailed results are outlined below: • There is strong and consistent evidence of disparity in misconduct allegations against police officers from Black and/or other Minority Ethnic communities. There has been no change in the disproportionality gap since the original MOPAC analysis in 2016. Police officers from Black and/or other Minority Ethnic communities remain twice as likely (x1.9) as white officers to be subject to a misconduct allegation. • The disparity is reflected across all minority ethnic groups, although the rate is highest for Black Police Officers (x2.1 times more likely), followed by Asian and ‘Other’ Ethnicity officers (both x1.9 times more likely) and Mixed officers (x1.4 times). • A disproportionality gap is reflected across all staff types. However, police officers account for the large majority (83%) of individuals subject to a misconduct allegation so volumes are low for all other staff types. • The most significant risk-factors for police employees to be subject to a misconduct allegation are gender, ethnicity, and length of service. Logistic regression indicates that being a Police Officer was the strongest predictor to be subject to a misconduct allegation, followed by (in order of strength) gender (male), Ethnicity (being Black, Asian or Mixed, with the strongest association for Black employees), and length of service (less than 5 years’ service). • Role-related variables such as rank, working pattern, business group or Basic Command Unit were not found to have an impact. However, working within a frontline role was found to be a predictor for those subject to a misconduct allegation, with individuals on the frontline being x1.3 times more likely to be subject to a misconduct allegation. • The results from a logistic regression on misconduct outcomes found that ethnicity was not an independent predictor of misconduct outcomes or of the type of formal action sanction issued. This would appear to confirm it is the entry into the system that is central. • Racial disparity is specific to the misconduct system, as compared to Right Line reporting (no difference by ethnicity) and public complaints (in which Black, Asian and Mixed Ethnicity police officers received a higher rate of public complaints per 100 Officers as compared to White but the disproportionality gap is smaller). This suggests a key driver for disproportionality is related to the police misconduct process itself. • Contrary to the main findings - qualitative research to explore and identify the drivers of disproportionality via a vignette experiment found that the ethnicity of the officer in a given scenario presented to supervisors did not significantly influence the responses to the questions in the vignette. Ethnicity of the officer (as a single variable) was not found to be linked to how a supervisor would respond to an allegation of misconduct, or the decision on whether to refer the allegation into the formal misconduct process. It should be noted that limitations of the vignette methodology may have impacted on results. • Vignette findings did highlight a lack of training as well as a high inconsistency and subjectivity in decision-making in relation to officer conduct. Just over a third (38%) of respondents reported they had completed training on the misconduct process, illustrating inconsistent training provision for supervisors on this topic. Conclusion • The report continues the ongoing analysis into the issue of disproportionality in police misconduct that MOPAC began in 2016. The current findings advance learning towards the underlying drivers and raises new directions for ongoing work: • Entry into the system should be the focus - findings from this report indicate the disproportionate use of internal misconduct proceedings starts at the beginning of the process when officers are first referred into the misconduct system by supervisors. Novel work exploring this could be case reviews, more detailed scenario-based decision-making exercises, or behavioural decision research. • More nuanced understanding of the drivers but gaps still remain – on top of ethnicity, factors such as length of service (under 5 years), gender (male), and age (under 25 years) were significant predictors of misconduct allegation. Given the workforce changes being seen in policing at present, future work to establish necessary support and supervision for this group would be appropriate. On a related point, the overall predictive model was relatively weak – suggesting a wider range of variables would be beneficial. Replicating the analysis with more ‘on the job’ variables (i.e., stop and searches conducted, arrests/investigations conducted, public interactions, entry route to the police, staff sickness, previous complaints or conduct matters) or additional personal characteristics would allow greater sophistication. • The need for improved training - a lack of consistent training for supervisors on performance management and/or the misconduct process will undermine any efforts to improve standards of behaviour. Improving these will help prevent misconduct in the first place and help some low-level cases from escalating, as well as encouraging quicker and more effective action where necessary. • Testing of interventions – the lack of evidence base as to what type of intervention may address the issue is a clear gap. It would be recommended that the MPS turn their focus to the development (and robust assessment) of interventions to counter the issue. • The issue is police service wide, yet the response to identifying and removing disparity within misconduct has been sporadic, resulting in an inconsistent approach to disproportionality in misconduct across the UK police service. It is hoped this research can make a useful contribution to the debate. Introduction Background Since 2016 MOPAC Evidence & Insight (E&I) has been undertaking research exploring and evidencing the issue of racial disparity within the police misconduct system. This is a critical area for the Mayor and MOPAC – the Mayor’s Action Plan to improve transparency, accountability, and trust in the Metropolitan Police Service (MPS) was published in November 20201 and this included a commitment by the MPS to continue to significantly reduce disproportionality within the grievance and misconduct processes by 2024. This issue has also been brought to the fore by the recent Dame Casey report on the police misconduct system (2022) where one of the key issues highlighted was the racial disparity throughout the Met’s misconduct system. These reviews have helped raise the awareness of the over-representation of staff from Black and/or other Minority Ethnic communities in MPS disciplinary proceedings, yet there remains a need to identify if there are any specific factors driving the disproportionality identified. MOPAC E&I 2016 Review The initial review (Wunsch et al., 2016) found that a number of studies evidence the existence of ethnic disproportionality in disciplinary proceedings in other police forces as well as other professions. This was supported further by the analysis of five years’ worth of MPS misconduct data (2010 – 2015)2 which recognised that police officers from Black and/or other Minority Ethnic communities were twice as lightly as white officers to be subject to misconduct allegations, and that these allegations were more likely to be substantiated for Black, Asian and Mixed Ethnicity officers. The 2016 MOPAC review also conducted qualitative research consisting of interviews with key staff involved in the assessment and investigation of misconduct in the MPS. Those interviewed for the 2016 review perceived MPS misconduct processes as fair and robust, and reported that the disproportionality evident in the data did not reflect their experiences. In talking about possible causes, most commonly talked about was a fear of being accused of racism, leading to supervisors in particular being less comfortable and therefore less likely to resolve issues informally. MOPAC E&I 2018 Review Analysis of three years’ worth of MPS misconduct data (financial year 2015-16 to financial year 2017-18) was undertaken as part of a commitment to conduct analysis following the implementation of the 2017 Action Plan to address disproportionality in misconduct proceedings in the MPS. This analysis concluded that a disproportionality gap was still evident, although noted that it had started to close very slightly over the three years, with police officers from Black and/or other Minority Ethnic communities 1.8 times more likely than white officers to be subject to misconduct allegations in 2017-18. Where more significant progress had been made was around outcomes of the allegations, with little difference found by ethnic group in the outcome of the allegation between 2015 and 2018. The proportion of allegations that were substantiated 1 https://www.london.gov.uk/sites/default/files/action_plan_- _transparency_accountability_and_trust_in_policing.pdf 2 The analysis focused on internal conduct matters pertaining to officers only. accounted for 30% of finalised allegations against Black, Asian and Mixed Ethnicity officers, and 27% of allegations against White officers (see Cain, 2019 for an overview of the results). These results indicated that progress has been made at the ‘outcome end’ of the misconduct process, and that the disproportionate representation of officers from Black and/or other Minority Ethnic communities starts at the beginning of the misconduct process. Research Objectives The overarching aim of the current research was to update the previous work and further explore and identify the drivers behind racial disparity in the misconduct system within the MPS. The report presents insights from a wider research programme undertaken in 2021 and 2022. The overall objectives for the research were to: • Update and explore any changes since the 2016 and 2018 MOPAC E&I research was conducted. This would enable continued tracking of disproportionality trends and evidence any changes. • Build a greater understanding of factors influencing disproportionality within the misconduct process. • To conduct detailed analysis of factors that could be influencing police officer misconduct using data mining and regression techniques. • To identify whether disproportionality is reflected across multiple reporting mechanisms and data sources. The analysis was conducted in two phases: • Phase 1 aimed to supplement the MOPAC analysis of misconduct data that has already been conducted. The research expanded on this analysis with key learning from the wider literature and social research evidence as well as utilising additional data sources and analytical techniques. • Phase 2 conducted a vignette experiment to further explore the potential drivers of disproportionality in misconduct decision-making processes. The key objective was to explore if the ethnicity of an officer is a driver for disproportionate referrals into the misconduct process within the MPS, as compared to other case-specific factors. It also examined allegations / concerns that could be resolved informally through performance management processes to better understand how decision-making processes may differ to those that are formally recorded via the misconduct process. Methodology Phase 1 of this project encompassed four elements: • An update of the literature review conducted as part of the 2016 original review, with focus on ‘what works’, or interventions that have been tried or are ongoing in the areas of misconduct and ethnic disproportionality. Published peer reviewed literature or published organisational reports from both the UK and internationally, up to the end of 2020, were identified through internet searches using a variety of search terms such as ‘misconduct’, ‘complaints’, ‘disproportionality’, ‘disciplinary hearings’, and ‘interventions’ for both the police and other public sector organisations such health care. • A review of the 2018 report and updating the results with analysis of MPS data on internal officer misconduct (using 2018-2020 data) to explore any changes following analysis conducted in 2018. Data was sourced from the MPS Centurion database. • Analysis of Right Line Data and Public Complaints data (using 2018-2020 data) to identify if there are any patterns of referral or disproportionality in relation to other reporting mechanisms and data sources. Data was sourced from the MPS Right Lines referral recording system and MPS Centurion database. • Use of data mining and regression techniques to identify if there are specific factors that can be used to predict and explain the disproportionality found for recorded misconduct allegations and the outcome received. Logistic regression3 was used to predict and explain the impact of specific factors (such as gender, staff type, business group, ethnicity, age, and length of service) on the number of misconduct allegations received over the 12-month period between January 2020 and December 2020. Phase 2 Vignette Experiment This vignette experiment employed short, systematically varied descriptions (i.e., vignettes) of police officer misconduct as well as matters that did not reach the threshold for formal disciplinary action in order to elicit the beliefs and likely behaviours of respondents with respect to the presented scenarios. Please see Appendix 1 for an example of a vignette utilised within this study. The study utilised text vignettes administered online via an online survey platform. For the vignette scenarios to appear plausible and real to participants they were constructed by MPS colleagues working within the Directorate of Professional Standards and Deputy Commissioner’s Delivery Group based on real-life situations. A total of five scenarios were presented (2 misconduct, 2 that did not reach the threshold for formal disciplinary action, and 1 designed to be more unclear and open to interpretation). The five identical scenarios were presented to each participant, with the ethnicity of the officer involved randomised. This randomised factor within the vignette gave the survey a unique capability to investigate the effect of ethnicity in the decisions and actions taken following the 3 Logistic regression is an analytic method that enables us to look at the effect of several predictor variables together upon an outcome. In this way, it is possible to understand the most important risk-factors that make an outcome more likely. initial reporting of police officer misconduct allegations. All other factors in the vignette were held constant (gender, rank, length of service). Following each scenario, associated questions were asked on how participants would respond and the actions they would take if they were in that situation as a line manager in real life. Demographic information was also collected from respondents at the end of the survey to help in investigating response behaviours. Participants were MPS police officers in line management roles (Sergeants, and Inspectors) across 4 Borough Command Units (BCU) - Central East, North Area, South East, and South Area. These BCUs were already taking part in a pilot focussing on performance management processes (focusing on the use of reflective practice) - so a survey on this topic would seem more logical to participants. The underlying focus of the research to explore disproportionality was not divulged to participants as this could have modified their responses. However, E&I worked closely with the MPS on design of the survey and the work received MPS ethnical approval. The vignette survey achieved an overall response rate of 47% (n=282). In terms of analytics, Chi- squared tests were carried to test if the randomised ethnicity had any link to the decision making of either an informal4 or formal misconduct route. Who responded to the survey? A third of respondents were from the South East BCU (33%, n=69/2095), followed by South Area (27%, n=57), and Central East and North Area (both 19%, n=41). Most respondents to the survey were Police Sergeants (51%, n=107/209), followed by Inspectors (21%, n=43) and Detective Sergeants (18%, n=37). Most had been in their current roles for five years or less (59%, n=119/202). Almost three-quarters of participants were White (73%, n=154/212), 10% (n=22) were Black, Asian or Mixed Ethnicity, and 17% preferred not to answer this question (n=36). This is in line with the wider cohort, in which 11% of all Inspectors and Sergeants in the MPS are Black, Asian or Mixed ethnicity6. Results This section of the report presents the results from across the two phases. This begins with a refresh of the literature, presentation of MPS misconduct data and other reporting mechanisms, before presenting the results of statistical modelling and finally sharing results from the vignette exercise. Literature Review Disproportionality in misconduct investigations is well documented (for example Cain, 2019; MOPAC 2016, Angiolini, 2020; Arhibong, Klin, Eshareturi & McIntosh, 2019). The three main 4 Either speaking to officer, keeping an informal record, or doing nothing. 5 There is a lower base size depending on the number of people who responded to certain questions in the survey. 6 As of end of January 2022. academic theories of reasons for disproportionality in internal misconduct proceedings that were identified in the MOPAC 2016 report remain salient today, including a fear of being labelled racist, implicit bias and failures to address differences: 1. Disproportionality as the result of a fear of being accused of racism, leading to managers and supervisors to shy away from trying to find informal solutions A number of studies (e.g. Cain 2019; Smith et al., 2012; Chapman, 2014) have found that low level conduct allegations are referred to Professional Standards Department (PSD) for officers from Black and/or other Minority Ethnic communities rather than managed by supervisors. An examination into misconduct allegations in West Midlands Police also identified that the primary driver of disproportionality is too many allegations being initially recorded at a higher level than they should be (West Midlands PCC, 2016). The supervisors described a lack of confidence challenging colleagues from Black and/or other Minority Ethnic communities for fear of being labelled racist or a fear of ‘getting it wrong’ (Ghaffur, 2004) and being subject to misconduct investigations or tribunals themselves, so passed the responsibility onto the PSD (Cain, 2019). Indeed, Smith et al (2012) argues that officers from Black and/or other Minority Ethnic communities saw these practices as a subtle form of racism. However, by drawing back from addressing minor performance and conduct issues means that these officers are often deprived of the learning and support that their White colleagues receive. 2. Disproportionality as the result of a failure of organisations to understand and deal with difference A failure of forces to understand and deal with difference has been identified as an explanation for the disproportionality in misconduct. More specifically, disproportionality exists because forces fail to deal with difference, particularly officers being ignorant to other cultures and people doing things differently (Chapman, 2014). Through workshops and one-to-one sessions with members of UK police forces, Cain (2019) found officers from Black and/or other Minority Ethnic communities highlighted a lack of cultural competence in Professional Standards Departments as well as a failure to consider culture when assessing and conducting misconduct investigations. Smith et al’s. (2012) qualitative research details how Black and/or other Minority Ethnic officers felt there was no attempt to understand differences by colleagues or senior officers and that colleagues preferred to believe it was wrong to be different. An independent review into police disciplinary proceedings concluded that the police often struggle to manage difference in their workforce, resulting in minor issues that are settled informally with White officers often end up with full disciplinary hearings for officers from Black and/or other Minority Ethnic communities (Chapman, 2014). 3. Disproportionality as the result of implicit bias Finally, disproportionality needs to be understood in terms of implicit bias. According to the implicit bias theory, we associate people we do not know with stereotypes related to their group (e.g., Fridell & Lim, 2016). These associations are implicit as they often occur outside of conscious awareness yet can impact on our behaviour. However, whilst there are numerous studies evidencing the existence of bias within policing, for example police intervening disproportionately with black people (EHRC, 2016), it must be noted that no quantitative research has been found which examined whether conscious or unconscious bias could explain the disproportionality in misconduct. The results of recent research conducted by the National Police Chiefs Council (NPCC) (Cain, 2019) concur with many findings from previous research. Cain (2019) found a primary driver of disproportionality was the failure of supervisors to deal with the low-level matters at the earliest opportunity or proportionately, placing officers under unnecessary lengthy investigations which can have a negative impact on officer’s health, career progression, family and the community of that officer. Once conduct allegations are passed onto the Professional Standards Department (PSD), cultural factors and working practices were then found to be inconsistently applied and the wider context is often ignored - in particular, a failure to explore if a ‘trigger incident’, e.g., whistle blowing or an incidence of racism, had occurred prior to the alleged misconduct. Disproportionality in misconduct is also not unique to policing. For example, previous research with the NHS has found that Black, Asian and minority ethnic staff are twice as likely to be disciplined compared to white staff (Archibong and Darr, 2010). Further research into disproportionality within NHS disciplinary proceedings identified additional factors at both an individual and organisational level that contribute to the over-representation of staff from Black and/or other Minority Ethnic communities, including (Arhibong, Klin, Eshareturi & McIntosh, 2019). These were: • Closed Culture – Black, Asian and minority ethnic staff were likely to get involved in disciplinaries as a consequence of being closed to the culture of the organisation. Reports of NHS Trusts fostering cultures that were hard to challenge, with individuals criticised if their behaviour did not conform to the accepted organisational norms. • Subjective attitudes and behaviour – where Black, Asian and minority ethnic staff attitudes/behaviours did not conform to the accepted norms, these individuals were likely to be involved in disciplinaries even if these behaviours did not reflect a breach in performance standards. Black, Asian and minority ethnic staff were also not given the opportunity through informal mentoring, training or discussions to understand the accepted behaviours prior to disciplinary proceedings commencing. This is fuelled by a lack of cultural competence in senior staff. • Unfair decision making – staff felt there was unfair decision making at different levels of the disciplinary process, which related to a lack of Black, Asian and minority ethnic senior managers being involved within the process. • Poor disciplinary support and discipline policy misapplication – poor support in the workplace for Black, Asian and minority ethnic staff who face disciplinary actions. For example, some staff are more comfortable discussing disciplinary issues with their peers, often in a comfortable language, but this could be seen as inappropriate. Disciplinary polices are not properly applied and are applied in the first instance to Black, Asian and minority ethnic staff where an informal approach would be appropriate. The arbitrary application of rules and a lack of culturally competent qualified managers have been identified as a root causes for poor disciplinary support. Whilst these triggers are not new and have been consistently evidenced in the literature (Smith et al., 2012; Sehmi, 2015), the difficulty is the reactiveness of practice to these triggers, reinforcing the need for clear policies advocating for informal resolution as a first step and to ensure staff involved in proceedings are culturally competent. Taking the literature into consideration, the fear of reprisal or of being accused of racism impacts on the levels of disparity. Inadvertently, this avoidance of dealing with low level matters at the earliest opportunity is magnifying those levels of distrust, resulting in the exact consequence those supervisors are seeking to avoid (Cain, 2019; Sehmi, 2015; Arhibong, Klin, Eshareturi & McIntosh, 2019). ‘What Works’ to address disproportionality? Whilst previous research has been able to clearly evidence the existence of disproportionality in police officer misconduct proceedings, a conclusive argument as to its causes or ‘what works’ in responding to it has yet to be identified. The issue is service wide yet the response to identifying and removing disparity within misconduct has been sporadic. This has resulted in an inconsistent approach to misconduct across the UK police service (Cain, 2019). However, good practice and recommendations for future changes have been identified from police forces across the UK which may offer solutions to address disproportionality in the misconduct process. For example: Consistency: • New legislation around Police Integrity Regulations provides an opportunity to develop a standardised approach across the service (Cain, 2019). • Making sure that PSD’s are more consistent in their approach on the use of guidance or working practise to understand cultural differences (Cain, 2019, Arhibong, Klin, Eshareturi & McIntosh, 2019). • Address the inconsistent methods used by forces/departments to capture data around protected characterises (Angiolini, 2020). Training • Provide comprehensive support for all staff, including a framework and standards for the provision of feedback (Atewologun, Kline & Ochieng, 2019). • Training needs to move away from traditional NCALT training, towards an engaging, interactive experience in a safe environment, for topics such as cultural awareness, emotional intelligence and having difficult conversations. Thus, encouraging learning through the lived experiences of officers from Black and/or other Minority Ethnic communities (Cain, 2019). • Management training for supervisors is required to provide them with the confidence to take responsibility and address misconduct at the earliest opportunity rather than passing onto the PSD (Angiolini, 2020; Arhibong, Klin, Eshareturi & McIntosh, 2019). Positively, the EHRC (September 2016) investigation into the MPS describes a clear appetite for further training, information and sharing of learning amongst managers and they were open to improving performance in this area. However, the report also recognised that an open and reflective culture is not established, with senior officers often not acknowledging when mistakes are made in misconduct cases which limits the MPS’ ability to identify learning. • A need for supervisors to receive more training around cultural competence and awareness, such as Wellbeing Passports. The National Police Chiefs Council (Cain, 2019) are looking at investment in the comprehensive cultural awareness training the MPS currently run – ‘Leading for London’ which works to develop an understanding of localised cultural awareness of communities being served by the force and supports line managers to make difficult decisions and to promote learning rather than penalising officers and staff (MPS Action Plan 2017). Organisational Learning & Leadership • Ensure PSD’s reflect the workforce and communities they serve and at present many do not. In 2019, out of all the Home Office forces, 63% (25 PSD) had no Black and/or other Minority Ethnic police officers or staff working within them (Cain, 2019). For those departments that did have Black and/or other Minority Ethnic representation, this was greater than the Black and/or other Minority Ethnic population percentage for police officers within their respective force. Within the MPS, officers from Black and/or other Minority Ethnic communities account for 15% of the PSD officer workforce7 (as compared to 16% of the total police workforce, and the London population of 44%)8. • There needs to be a better development of cultural understanding across PSDs and bespoke positive action plans to improve representation within departments (Cain, 2019; Angiolini, 2020). • There is a need for strong and ethical leadership at all ranks to provide a supportive, inclusive and high performing team that feel supported when challenging how they’ve been treated. • Promoting people who understand and value difference is pivotal to ensuring those BAME colleagues entering the service are supported and nurtured to realise their potential. In addition, a recent review into referrals of medical staff to the General Medical Council has raised recommendations on how to address disproportionality that focus particularly on organisational leadership (Atewologun, Kline & Ochieng, 2019). For example: • Ensuring senior leaders engage regularly with staff, listening to and taking action in response to concerns regarding fairness and inclusion, providing positive leadership more consistently across the organisation. • Identify and address systemic issues that may affect individuals’ professional performance. When undertaking a performance assessment or responding to a concern, take into account the context in which they work with a focus on learning not blame. • Senior leaders to engage regularly with all staff, listening to and taking action in response to concerns regarding fairness. Implement a strategy of active inclusion and mechanisms to mitigate the risk of disproportionality in discipline and referral processes. • Establish a UK-wide mechanism to share good practice and undertake ongoing monitoring of data concerning misconduct disproportionality. 7 MPS Workforce Data Report (August 2021). 8 Based on GLA 2016-based Round of Demographic Projections for 2020 (Housing-led ethnic group projections). It is important to note however, that at this stage these are useful recommendations, but it remains unclear ‘what works’ to reduce racial disparity in the misconduct system. Hence this ongoing research programme of work sought to further explore and identify the key drivers behind disproportionality in misconduct cases within the MPS. The learning and insights gained via this research will be important in identifying any potential barriers or enablers to fair misconduct processes in the MPS and be used to drive change within the misconduct system. Baroness Casey Review – Interim Report on Misconduct (2022) Since the literature review was conducted in 2021 the MPS appointed Baroness Louise Casey to lead an independent review of its culture and standards of behaviour. The interim findings published in October 2022 focussed on the internal misconduct system. One of the key issues highlighted was evidence of racial disparity throughout the Met’s misconduct system. Consistent with conclusions found in several research pieces that precede this report - the main finding was that Black, Asian, and Mixed Ethnicity officers and staff were substantially more likely than their White colleagues to receive a misconduct allegation and that this was consistent in every year of data (2013-2022). In the most recent financial year (2021-22), Black officers and staff were 81% more likely to receive a misconduct allegation than their White colleagues, Asian officers/staff were 55% more likely and Mixed Ethnicity officers/staff 41% more likely. Whilst useful, the interim report did not include any statistical modelling to explore or explain the emerging findings. The report also found that up until 2018-19 Black officers and staff were more likely than their White colleagues to have a misconduct allegation given a case to answer decision, and therefore have a sanction justified. After 2018-19 this disparity reduces. This was also reflected in MOPAC analysis which preceded this report - the initial MOPAC review in 2016 also found that police officers from a Black or minority ethnic background were significantly more likely to have a misconduct allegation substantiated as compared to officers from a white background. However, the MOPAC 2018 update report found that the disproportionality gap had closed and that once reported and investigated there was no longer significant disproportionality in the type of outcome received. As reported in previous studies, several reasons are cited for this disparity in the misconduct system. This included the concern that raising issues relating to racism or other discrimination and wrongdoing often led to being labelled a troublemaker, which then led to unfair disciplinary action. It was also reported that the misconduct system is not sufficiently robust with White officers who breach professional standards, but there is a lower threshold for Black, Asian and Mixed Ethnicity officers and staff. There may also be a reluctance among supervisors to address low level incidents for fear of being labelled a racist and being 'passed up’ into the misconduct system. In-depth analysis of MPS misconduct data One of the objectives for this research was to revisit the 2016 and 2018 MOPAC analysis and explore any changes in the findings of these research reports. The analysis was conducted in June- August 2021 and enabled continued tracking of disproportionality trends over 2018 - 2020. This section will update the previous results with analysis of MPS data on internal officer misconduct allegations, and then move to an exploration of finalised misconduct outcomes. Misconduct allegations Analysis of three years’ worth of MPS misconduct data (January 2018 – December 2020) was undertaken. Over the three-year period, the number of MPS police officers subject to an allegation of misconduct was highest in 2018 and has since fluctuated, as shown in Figure 1 below. Figure 1: Total police officers subject to a misconduct allegation (2018-2020) A rate per 100 population has been used throughout the rest of the report to account for changes in allegation levels and police officer strength over time.9 Disproportionality Rate MOPAC analysis of disproportionality in police misconduct conducted in 2016 and 2018 found that Black, Asian and Mixed Ethnicity officers were twice as likely as White officers to be subject to a misconduct allegation. As shown in Figure 2 below, there has been no change in the disproportionality gap over the last three years - with Black, Asian and minority ethnic (BAME) officers still x1.9 times more likely than White officers to be subject to a misconduct allegation. An average rate of 4.9 BAME Police Officers per 100 workforce were subject to a misconduct allegation between 2018 and 202010, as compared to a rate of 2.6 for White Police Officers. 9 Monthly workforce figures from the MPS Monthly Workforce Report have been aggregated to represent the average staffing levels for the individual years reported. 10 Rate of 5 BAME Police Officers per 100 and 2.6 White Police Officers per 100 in 2018 report. 1,625 1,241 1,493 1,090 805 973 2018 2019 2020 Allegation Count Officer Count Figure 2: Police officer misconduct rate per 100 officers Table 1: Police officer misconduct allegations disproportionality rate (2018-2020) Misconduct Allegations 2018 2019 2020 Ratio BAME to White 1.9 1.8 1.9 Although BAME Police Officers were x1.9 times more likely to be subject to a misconduct allegation, the rate was highest for Black Police Officers (x2.1 times more likely) followed by Asian and ‘Other’ Ethnicity officers11 (both x1.9 times more likely), and Mixed officers (x1.4 times). As shown in Figure 3 below, the misconduct rate is consistently higher for Black Police Officers, except for 2020 in which officers from an “Other” ethnic background recorded the highest rate per 100 officers. It should be noted that the category of “Other” is small (2% of total police officer workforce) so rates should be interpreted with caution. Low numbers mean that it is impossible to distinguish random fluctuation from true changes in the underlying risk of being subject to a misconduct allegation. The rate of misconduct allegations per Black officer has remained stable over the last three years. Figure 3: Police officer misconduct rate per 100 officers by ethnic group The misconduct rate by staff type is shown in Figure 4 below. A disproportionality gap is reflected across all staff types. The difference was statistically significant across all staff types, but the largest disparity is for BAME Police Staff (x2.4 times more likely to be subject to a 11 The ‘Other’ ethnicity grouping includes the categories Chinese, Other Groups, Greek and Greek Cypriot, Turkish and Turkish Cypriot. 5.4 4.2 4.9 2.9 2.4 2.6 2018 2019 2020 BAME rate per 100 officers White rate per 100 officers 5.4 4.6 5.2 5.6 5.7 5.5 5.2 2.7 3.3 5.4 3.0 6.3 2.9 2.4 2.6 2018 2019 2020 Asian Black Mixed Other White misconduct allegation as compared to White Police Staff). The lowest disparity was recorded for BAME Special Constables (MSC) who were x1.6 times more likely to be subject to a misconduct allegation as compared to White Special Constables. However, it should be noted that 83% of individuals subject to a misconduct allegation over the last three years were Police Officers so volumes are low for other staff types - 10% (342) were Police Staff, 3% (86) were PCSOs, and 4% (139) were Special Constables (MSC). For those police officers subject to a misconduct allegation, two-thirds (66%) were working in frontline roles, (21%) in Met Operations and (8%) in Specialist Operations. Figure 4: Misconduct rate per 100 workforce by staff type (2018-2020) Table 2: Misconduct disproportionality rate by staff type (2018-2020)12 Rate Police Officer Police Staff PCSO MSC Ratio BAME to White 1.9 2.4 2.1 1.6 Misconduct Outcomes This section of the report focuses on misconduct outcomes, to explore if disparity exists within the decision-making end of the misconduct process. The analysis was conducted on outcomes of allegations for Police Officers only as they account for the majority of the workforce who were subject to a misconduct allegation and outcome in the given period. Any additional findings of significance related to other staff type will be noted. Allegation Outcomes The initial review in 2016 found that police officers from a BAME ethnic background were significantly more likely to have a misconduct allegation substantiated as compared to officers from a white background (48% BAME as compared to 39% white), whereas white officers were more likely to have the allegation against them unsubstantiated. The 2018 analysis found that the disproportionality gap for misconduct outcomes was largest between 2012-13 and 2013-14 before closing (by 6 percentage points) between 2015-16 and 2017-18. 12 The differences for each staff type were all significant at p < 05. 4.9 2.2 3.4 3.1 2.6 0.9 1.6 2.0 Police Officer Police Staff PCSO MSC BAME White Figure 5 below shows that between 2018 and 2020 the converse has been found, with 36% of White Police Officers subject to a misconduct allegation having their allegation substantiated, as compared to 32% of BAME Police Officers. This means that although BAME Police Officers were more likely to be subject to a misconduct allegation, they were proportionally less likely to have it substantiated. These results corroborate findings from the 2018 analysis and indicate that progress in reducing disproportionality has been made at the ‘outcome end’ of the misconduct process. The majority of misconduct allegations are unsubstantiated (accounting for almost two thirds of police officer misconduct allegations for both BAME and white officers). Figure 5: Outcomes of finalised police officer misconduct allegations (2018-2020) As seen in Figure 6 below, the overall proportion of finalised allegations that are substantiated has dropped for both BAME and White officers. Although it is important to note that this analysis was conducted in June 2021, and at that point over half (58%) of all allegations recorded in 2020 were yet to be finalised at the time of the analysis. Figure 6: Proportion of finalised allegations that are substantiated (police officers) More detailed analysis of police officer ethnicity shows that this is largely driven by lower substantiated rates for ‘Other’ and Mixed ethnicity police officers. 5% 32% 63% 3% 36% 61% Discontinuance Substantiated Unsubstantiated BAME White 37% 34% 22% 38% 39% 30% 2018 2019 2020 % BAME % White Table 3: Proportion of finalised allegations that are substantiated by ethnicity (police officers) Police officer ethnicity 2018 2019 202013 Asian 33.7% 38.6% 16.3% Black 38.2% 34.7% 36.7% Mixed 44.9% 31.8% 11.8% Other 28.0% 13.3% 21.1% White 38.0% 38.9% 29.7% Write Off Method For those misconduct allegations that are substantiated, a higher proportion of BAME Police Officers received a Formal Action as compared to White Police Officers (67% vs. 58%), as shown in Figure 7. Within BAME Police Officers, Asian (65%), Black (63%), Mixed (71%) and Other (85%) had a higher rate of Formal Action when compared to White Police Officers (58%). White Police Officers were more likely to receive Management Action when compared to all other ethnicities. These differences across write off method were not found to be statistically significant14. A higher proportion of White Police Officers received Management Action (20% vs. 16%) and No Action (12% vs. 7%) as compared to BAME Police Officers. This difference is not statistically significant15. Figure 7: Write off method for substantiated allegations (police officers) 2018 - 2020 The use of Formal Action for substantiated allegations has dropped notably as compared to the 2018 analysis. In the previous report, 79% of White Police Officers received a Formal Action compared to only 58% between 2018-2020. Similarly, 78% of BAME Police Officers received a Formal Action in the 2018 report, as compared to 67% between 2018-2020. 13 Over half (58%) of all allegations recorded in 2020 are still yet to be finalised at the time of this report. 14 Comparison of ethnicities and proportion of formal action received. The chi-square statistic is 6.8212. The p- value is .145644. The result is not significant at p < .05. 15 Comparison of BAME and White Police Officers of No Action received. The chi-square statistic is 2.9221. The p- value is .087374. The result is not significant at p < .05. 67% 16% 7% 6% 4% 58% 20% 12% 6% 4% Formal Action Management Action No Action Reflective Practice Meeting Retired/Resigned % BAME % White The reduction in the proportion of Formal Action being taken can be explained by the increased use of Management Action, with a higher proportion of use for substantiated allegations. In the 2018 report the use of Management Action for substantiated allegations between 2015-2018 was evenly split between BAME and White Police Officers (7% BAME vs. 8% White), however, uplift in this method has also resulted in some disparity. Between 2018 – 2020 a higher proportion of White Police Officers received Management Action for substantiated allegations as compared to BAME Police Officers (16% BAME vs. 20% White)16. This difference was not found to be statistically significant. Figure 8 shows there is no significant difference between BAME and White Police Officers in the write off method for unsubstantiated allegations. The overall use of management action for unsubstantiated allegations has significantly reduced as compared to previous analysis. The initial review in 2016 found that management action was used in 34% of all unsubstantiated allegations (as compared to 15% between 2018-2020). The new category of a Reflective Practice Meeting is proportionally used equally between BAME and White Police Officers. Figure 8: Write off method for unsubstantiated allegations (police officers) 2018-2020 Formal Action Sanctions As shown in Figure 9 below, BAME police officers were most likely to receive a Written Warning, and this was at a higher rate than White officers (34% vs. 26%). The dismissal rate for both BAME and White officers was similar (12% vs. 11%). White officers were most likely to receive Management Advice (26%). No Action was taken for a higher proportion of White Police Officers (10% for White Police Officers as compared to 4% for BAME Police Officers). 16 Comparison of BAME and White Police Officers of Management Action received. The chi-square statistic is 1.0034. The p-value is .316489. The result is not significant at p < .05. 68% 15% 14% 1% 1% 1% 71% 15% 13% 1% 1% 0% No Action Management Action Reflective Practice Meeting UPP Retired/Resigned Other Informal Action % BAME % White Figure 9: Formal action outcomes (police officers) 2018-2020 There was a total of 53 dismissals from the MPS in the period between 2018 and 2020 as a result of a substantiated misconduct allegation. Of the 53, 44 of those were Police Officers and 9 were Special Constables. Of the Police Officers dismissed 32 were White (73%) and 11 BAME (25%). Of the BAME Police Officers, 3 were Asian, 3 Black and 5 Mixed. Analysis of other reporting mechanisms Analysis of Right Line Data and Public Complaints data (using 2018-2020 data) was conducted to identify if there are any patterns of referral or disproportionality in relation to other reporting mechanisms and data sources as compared to the misconduct system. The analysis was conducted in 2021. Right Line Referrals The MPS has an external reporting system for MPS employees. If a member of staff wishes to anonymously report wrongdoing, they are able to do so through the Right Line telephone number, or electronically through Right Line On-Line. Right Line Referrals does not distinguish by staff type. There were 1,287 Right Line Referrals disclosed in the period between 2018 and 2020, which is lower than the total misconduct allegations over the same period (4,359). This may suggest a lack of awareness of the Right Line reporting mechanism. The rate of Right Line referrals peaked in 2019 - which is a different trend to internal misconduct allegations in which the rate was highest in 2018 and lowest in 2019. Of the total Right Line Referrals recorded between 2019 and 2020, almost three quarters (72%) of individuals were White and 16% BAME. As shown in Figure 10 below, the disproportionality evident in misconduct allegations is not reflected in confidential Right Line referrals. The rate of Right Line referrals per 100 workforce did not differ by ethnicity17. 17 Referrals are mostly for CARMS misuse/not working hours/overtime concerns/working from home concerns. Figure 10: Right Line referrals rate per 100 workforce Public Complaints MPS public complaint data (Figure 11) shows that in the period between 2018 and 2020 BAME Police Officers received a higher rate of public complaints per 100 Officers as compared to White Police Officers. This difference is statistically significant18. However, when compared to internal misconduct allegations the disproportionality gap is smaller - BAME Police Officers were 1.3 times more likely to receive a public complaint when compared to White Police Officers in the period between 2018 and 2020. The rate of complaints has increased each year, which is a different trend to internal misconduct allegations in which the rate was highest in 2018 and lowest in 2019. Figure 11: Police officer public complaints rate per 100 officers Table 4: Police officer public complaints disproportionality rate Rates 2018 2019 2020 Ratio BAME to White 1.3 1.3 1.3 18 Comparing misconduct and public complaints for BAME v White police officers: The chi-square statistic is 78.3789. The p-value is < .00001. The result is significant at p < .05. 0.7 1.1 0.8 0.8 1.1 0.8 2018 2019 2020 BAME White 30.2 36.2 41.1 23.9 28.0 31.8 2018 2019 2020 BAME White Modelling to understanding the differences One of the new analytic directions within the present research was an exploration of the drivers behind the disproportionally observed in the misconduct system. Previous analyses have evidenced the over-representation of staff from Black and/or other Minority Ethnic communities in MPS disciplinary proceedings, yet there remains a need to identify if there are any specific factors driving the disproportionality identified. To this end the research undertook a series of regression analytics and other methods to attempt to delve deeper into the drivers behind disproportionality in the misconduct system. A logistic regression was carried out to understand if variables based on employee demographic and role- related data could significantly predict whether an officer was subject to a misconduct allegation (see Appendix for the full model). A logistic regression was also carried out on misconduct outcomes to determine if any of these variables were key predictors of substantiated allegations or formal action sanctions. Key factors predicting Misconduct Allegation As shown in Table 5 below, the regression model identified being a police officer (rather than other staff types), being male, of Black or Asian ethnicity and having under 5 years length of service as the strongest predictors for being subject to a misconduct allegation, but relatively speaking only explained a small proportion of the total variance.19 To this effect, replicating this analysis with a wider array of variables, especially ones related to ‘on the job’ and other situations variables (i.e., stop and searches conducted, investigations conducted, public interactions, entry route to the police, staff sickness, previous complaints or conduct matters) or additional personal characteristics would enable a more sophisticated overall analytic. Table 5: Risk-factors for police employees to be subject to a misconduct allegation Factor Likelihood of inclusion Staff Type (Police Officer) x 2.4 more likely Male x 2.1 more likely Black ethnicity x 1.95 more likely Staff Type (MSC) x 1.92 less likely Length of service: (3 to 5 years) x 1.8 more likely Asian ethnicity x 1.7 more likely Length of service: (under 2 years) x 1.6 more likely Age under 25 years x 1.5 more likely Mixed ethnicity x 1.4 more likely Length of service: (6-10 years) x 1.38 more likely Front line role x 1.3 more likely 19 The strength of this model is reflected in the Nagelkerke R Square statistic which was 0.05. This is a low R-squared value but the independent variables are still statistically significant. Statistically significant coefficients continue to represent the odds of the outcome variable occurring holding other variable constant. Substantiated Misconduct Allegations Analysis then sought to replicate the previous analysis to determine if any of these variables were key predictors of misconduct outcomes (likelihood of an allegation being substantiated or issued a formal action sanction). The strongest predictor in predicting increased likelihood of a misconduct allegation being substantiated was age20. Those aged under 25 years were 1.6 times more likely to have their misconduct allegation substantiated. The regression also found that allegations triaged for the severity of gross misconduct was 1.4 times less likely to be substantiated. Ethnicity was not found to be a predictor of substantiated allegations. Formal Action Ethnicity was not found to be a predictor of a formal action sanction. Gross misconduct was the only significant predictor in determining whether a substantiated misconduct allegation would lead to a formal action taking place. Gross misconduct was 2.1 times more likely to receive formal action from a substantiated misconduct allegation. Vignette Study Results In order to further explore the initial misconduct decisions, the research sought to include a qualitative aspect where vignettes were used to explore decision making. Responses to Informal Scenarios Two of the vignettes in the study were designed with the expectation of being more likely to have ‘informal’ action taken, to explore if decision-making processes may differ to those that should be formally recorded via the misconduct process. This may include informal action such as Reflective Practice which is recorded locally, Reflective Practice Review Process (RPRP), or via the relevant performance framework. For this scenario, results showed that the majority of respondents decided to take informal action for these scenarios (see Table 6); a quarter of respondents (25%, n=71) to Informal Vignette 1 selected to refer the case to Professional Standards as a misconduct referral (via submitting an MM1 form), despite the scenario being designed as appropriate for ‘informal’ action. This is reflective of the apparent high subjectivity in the decision-making process. Results indicate that ethnicity of the officer in these scenarios did not significantly influence the response. Regression analysis was also conducted to explore if ethnicity was a significant driver when considered in combination with other factors such as participant length of service, rank, own ethnicity, and previous experience in misconduct matters. Analysis did not find ethnicity to be a significant driver. Cases involving Black or Asian individuals were not significantly different in terms of conduct decisions. 20 The Nagelkerke R Square statistic was 0.025 which is not a strong predictive model. Table 6: Responses to ‘Informal action’ Vignettes Informal Vignette 1 White Black Asian Total Informal action 76 68 67 211 78% 72% 75% Formal Action (misconduct) 22 27 22 71 22% 28% 25% Total 98 95 89 282 Informal Vignette 2 White Black Asian Total Informal action 85 72 72 229 97% 97% 100% Formal Action (misconduct) 3 2 0 5 3% 3% 0% Total 88 74 72 234 Responses to Formal Scenarios Two vignettes were designed with the expectation of being more likely to have formal action and suitable for a misconduct investigation. While the results showed that most respondents to these scenarios did decide to take formal action (see Table 7 below); a third of respondents to one of the vignettes still selected informal action despite the scenario being designed as appropriate for the formal misconduct process. Again, this illustrates the apparent high subjectivity in the decision-making process. Again, there were minimal and non-significant differences between each ethnicity; and the ethnicity of the officer in the scenario did not significantly influence the responses to the questions. Ethnicity was not found as a significant driver when considered as both a separate category and in combination with other factors. Table 7: Responses to ‘Formal Action’ Vignettes Formal Vignette 1 White Black Asian Total Informal action 10 11 8 29 13% 13% 9% Formal Action (misconduct) 70 77 83 230 88% 88% 91% Total 80 88 91 259 Formal Vignette 2 White Black Asian Total Informal action 23 23 32 78 32% 29% 37% Formal Action (misconduct) 50 57 55 162 68% 71% 63% Total 73 80 87 240 One vignette was designed to be more unclear and open to interpretation. Most respondents (87%, n=198) decided to take formal action for this scenario (see Table 8 below). Again, ethnicity was not found to be a significant influencing factor in decision-making for this scenario. Table 8: Vignette 5 results White Black Asian Total Informal action 11 13 6 30 15% 15% 8% Formal Action (misconduct) 62 71 65 198 68% 71% 63% Total 73 84 71 228 The link between training and decision-making The ‘Leading for London’ programme is a leadership training package within the MPS that aims to build the skills, capability and positive behaviour the MPS needs to create an inclusive workforce. Included in the final set of questions on participant characteristics, respondents were asked whether they had completed this training programme. Out of the 212 respondents who answered this question, 74% (n=157) said they had completed it. Some significant relationships were found here when examining this variable in relation to the action that respondents chose to take. It was found to be a significant relationship in two of the five vignettes where if participants had completed the Leading for London training, they were more likely to choose a formal misconduct route compared to an informal one (p<0.05) and this was on both the vignette designed to elicit a formal response but also on the one more open to interpretation. Table 9: Training and conduct decision-making Formal Vignette Completed Leading for London training Yes No Total Informal 13 10 23 8% 18% Formal Action (misconduct) 144 45 189 92% 82% Total 157 55 212 Vignette 5 Completed Leading for London training Yes No Total Informal 15 11 26 10% 20% Formal Action (misconduct) 141 44 185 90% 80% Total 156 55 211 Finally, the survey identified asked some more general questions about the level of training received on the topic of misconduct matters. Results here are noteworthy – with only 38% (n=79/210) noting that they had completed training on misconduct matters, with over half (55%, n=115) specifying they had not been offered any training. Of those who had completed previous training on handling misconduct matters, 18% thought it was poor (n=14), 46% (n=36) satisfactory, 33% (n=26) good, and 4% (n=3) excellent. No significant findings arose when examining the relationship between previous misconduct training and decision-making in the vignettes. In terms of previous experience in handling misconduct matters, very few respondents believed they had either ‘lots’ or ‘extensive’ experience in dealing with this, with almost half of respondents believing they had ‘some’ experience (48%, n=103). Discussion While it is important that the MPS has the power to apply disciplinary procedures, ensuring staff behave in a professional and appropriate manner, the evidence to date indicates that disciplinary procedures are disproportionately applied to police officers from Black and/or other Minority Ethnic communities. These issues are not new, not unique to policing and as a topic, this been explored analytically since 2016 by MOPAC Evidence and Insight. The overarching aim of the current research was to monitor wider trends on the issue and further explore and identify the drivers behind racial disparity in the misconduct system within the MPS. The report presents insights from a wider research programme undertaken in 2021 and 2022. Summary of the results Analysis of misconduct data between 2018 and 2020 found that police officers from Black and/or other Minority Ethnic communities are still twice as likely (1.9) as white officers to be subject to a misconduct allegation. The disparity is reflected across all minority ethnic groups, although the rate is highest for Black Police Officers (2.1 times more likely). The gap is also reflected across all staff types. Indeed, there has been no change in this disproportionality gap since the original MOPAC analysis in 2016. The more recent publication of the Baroness Casey Review Interim Report on Misconduct (2022) also found that this disparity was still evident in every year between 2013-14 and 2021-22. The disproportionality gap is most pronounced for misconduct referrals as compared to other reporting pathways. The disproportionality gap recorded in misconduct cases is not replicated for confidential Right Line Referrals and is smaller for public complaints made against officers. The annual trends also vary across the different reporting mechanisms and pathways for reporting misconduct, indicating that they have different drivers and that the format and process of the reporting mechanism has an influence on how it is used. One of the key objectives of this report was to analyse the factors that are influencing the disparity in the police officer misconduct system using data mining and regression techniques. An anonymised dataset of all employees working within the Metropolitan Police Service was used to examine which factors statistically matter most for misconduct referrals and outcomes. The results show that being a police officer was associated with increased likelihood of being subject to a misconduct allegation followed by (in order of strength) gender (male), ethnicity (being Black, Asian or Mixed, with the strongest association for Black employees), and length of service (less than 5 years’ service). Other role-related variables were not found to have an impact such as rank, working pattern, business group or Basic Command Unit of the employee. A vignette study was conducted to explore if the ethnicity of an officer was an independent driver for disproportionate referrals into the initial misconduct process within the MPS. In contrast to the wider statistics, the research found that the ethnicity of officers within the presented scenarios did not significantly influence the responses to the questions in the vignette. That is, ethnicity of the officer – as a single variable – was not found to be linked to how a supervisor would respond to an allegation of misconduct, or the decision on whether to refer the allegation into the formal misconduct process. The vignette did identify that those participants to have completed the Leading for London management training were significantly more likely to choose a formal misconduct compared to an informal procedure. The other notable aspect from the vignette were the officer perceptions towards training, which indicated over half had not received relevant training. Various aspects above are consistent with the previous research and wider evidence base, others that are original are raised as deeper areas for discussions. Entry into the system not outcome is the issue Results demonstrated that the disparity starts at the beginning of the misconduct process when officers are first referred into the misconduct system by supervisors. For example, Ethnicity (being Black, Asian or Mixed) was an independent predictor for the likelihood of being subject to a misconduct allegation - which shows that even when gender, length of service and other variables were taken into account, significant differences remain between against employees from Black and minority ethnic groups and White employees. However, ethnicity was not found to be a predictor of substantiated allegations or a formal action sanction. This indicates that improvements are likely required within line management and supervision within the MPS – when officers are first referred into the misconduct system by supervisors. These entry points become critical opportunities in setting, embedding and enforcing professional standards. This further corroborates findings from previous research studies that suggest disproportionality arises as a consequence of conduct allegations being referred to the Professional Standards Department for officers from Black and/or other Minority Ethnic communities, whereas similar behaviour of white officers is dealt with informally by divisional supervisors (e.g., Cain 2019; Smith et al., 2012; Chapman, 2014). Findings from both MOPAC research (2016) and the more recent Casey Review (2022) also found that there may be a reluctance among supervisors to address low level incidents for fear of being labelled a racist and are therefore being 'passed up’ into the misconduct system. It was also reported that the misconduct system is not sufficiently robust with White officers who breach professional standards, but there is a lower threshold for Black, Asian and Mixed Ethnicity officers and staff. More nuanced understanding of the drivers but gaps still remain The research findings from both Phase 1 and Phase 2 underline the value of accounting for multiple categories in any action plans or work programmes to reduce the disparity in misconduct allegations against Black, Asian and minority ethnic employees within the MPS. The current work undertook statistical modelling to explore the wider factors that may also contribute to the issue of disproportionality. As outlined earlier, ethnicity was a predictor for being subject to a misconduct allegation, along with wider factors such as the Police officer role, being Male, length of service (under 5 years) and/or age (under 25 years). Whilst interesting the overall predictive model was relatively weak – suggesting a wider range of variables would be beneficial in order to understand to a greater level. Indeed, ideally this analytic could be repeated with additional detail – moving into ‘on the job’ variables (i.e., stop and searches conducted, arrests, investigations conducted, public interactions, entry route to the police, staff sickness, previous complaints or conduct matters) or additional personal characteristics. None-the-less, a greater appreciation of the findings that did emerge would also be useful to help forward planning. For example, MPS workforce statistics indicate police workforce is becoming younger – under 30s now account for a quarter (25.1%) of all MPS police officers, an increase of 6 percentage points over the last 5 years.21 The increase in recruitment over the last two years also means that a third of the workforce have under five years’ service (33.6%). In this way, new recruits may be at increased risk of being subject to a misconduct allegation and may require additional support and supervision and this can be built into current governance. The need for improved training Findings from the vignette study pointed to a high inconsistency and subjectivity in decision- making in relation to officer conduct, and at the stage where a supervisor decides whether a scenario is suitable for the formal misconduct process or not. On a related point there was a clear gap in the training available for staff. Just over a third (38%) of respondents reported they had completed training on the misconduct process, illustrating inconsistent training provision for supervisors on this topic. These are stark findings that illustrate a deficiency that needs to be addressed organisationally. This lack of consistent training for supervisors on performance management and the misconduct process will likely undermine any efforts to improve decision making going forward. Similarly, the Casey Review (2022) also highlighted that Human Resources does not have enough of a role in the misconduct system or in supporting line managers. Improving these will help prevent misconduct in the first place and help some low- level cases from escalating, as well as encouraging quicker and more effective action where necessary. Testing of interventions Whilst research to date has been able to evidence the existence of disproportionality, no studies have so far been able to put forward conclusive evidence as to ‘what works’ in responding to it. The issue is police service wide, yet the response to identifying and removing disparity within misconduct has been sporadic. This has resulted in an inconsistent approach to misconduct across the UK police service. With this in mind it would be recommended that the MPS turn their focus to the development (and robust assessment) of interventions to counter the issue. Limitations We should acknowledge that the results presented in this report are limited to the information captured on the MPS Centurion system and therefore variables that may impact decision making will not all be included in the data. The available factors to explore risk factors for misconduct allegations and outcome categories were limited to employee demographic data and role related factors. It is important to note that we cannot draw conclusions about how and why the factors we discuss may influence outcomes. The analyses we present in the report is only able to illustrate 21 MPS Workforce Data Report. the relationships between the various factors and outcomes and the strength of each factor’s unique effect. Whilst the vignette approach was an innovative attempt to explore officer decision making – the method does have limitations. The scenarios were carefully designed to represent real-life conduct situations, however one of the main concerns of the methodology is the overall validity and inability to capture all elements of reality or the full complexity of misconduct assessments. While participants were not fully informed about the focus on disproportionality, it was possible that respondents guessed the purpose behind the study, which may have influenced their responses. This participant bias would mean that responses given to vignettes may be socially desirable and could be a different response to how people may react to real life situations. Conclusion The report continues the ongoing analysis into the issue of disproportionality in police misconduct that MOPAC began in 2016. The current findings advance learning towards the underlying drivers and raises new directions for ongoing work – the focus on the beginning of the process, the need for additional variables to better explore risk factors, the role of training, and the need for the new design and robust testing of interventions. The issue is police service wide, yet the response to identifying and removing disparity within misconduct has been sporadic, resulting in an inconsistent approach to disproportionality in misconduct across the UK police service. It is hoped this research can make a useful contribution to the debate. References Archibong, U., & Darr, A. (2010). The involvement of Black and minority ethnic staff in NHS disciplinary proceedings. Centre for Inclusion and Diversity, University of Bradford, Yorkshire, UK. Archibong, U., Kline, R., Eshareturi, C., & McIntosh, B. (2019). Disproportionality in NHS disciplinary proceedings. British Journal of Healthcare Management, 25 (4). Angiolini, E. (2020). Policing - complaints handling, investigations and misconduct issues: Independent review. Cabinet Secretary for Justice. Atewologun, D., Kline, R., & Ochieng, M. (2019). Fair to refer? Reducing disproportionality in fitness to practise concerns reported to the GMC. Retrieved from: https://www.ouh.nhs.uk/education-centres/training/documents/fair-to-refer.pdf Baroness Casey Review (2022). Baroness Casey Review – Interim Report on Misconduct. Metropolitan Police. Retrieved from: https://www.met.police.uk/police-forces/metropolitan- police/areas/about-us/about-the-met/bcr/baroness-caseys-report-misconduct/ Barter, CA., & Renold, E. (1999). The Use of Vignettes in Qualitative Research. Social Research Update, 25. Cain, P. (2019). NPCC: Understanding disproportionality in police complaint and misconduct cases for BAME police officers and staff [online]. National Police Chiefs Council. Available from: https://www.npcc.police.uk/documents/NPCC%20Understanding%20Disproportionality%20in %20Police%20Complaint%20Misconduct%20Cases%20for%20BAME%20Police%20Officers% 20and%20Staff%202019.pdf Chapman, C. (October 2014). An independent review of the police disciplinary system in England and Wales. Home Office. Retrieved from https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/385911/An _Independent_Review_of_the_Police_Disciplinary_System_-_Report_-_Final....pdf Commission on Race and Ethnic Disparities (CRED) (2021). Independent report: Crime and Policing. CRED. Retrieved from: https://www.gov.uk/government/publications/the-report-of- the-commission-on-race-and-ethnic-disparities/crime-and-policing Equality and Human Rights Commission (2016). Stop and think: A critical review of the use of stop and search powers in England and Wales. Equality and Human Rights Commission. Retrieved from https://www.equalityhumanrights.com/en/publication-download/stop-and- think-critical-review-use-stop-and-search-powers-england-and-wales Equality and Human Rights Commission (2016). Section 20 investigation into the Metropolitan Police Service. Equality and Human Rights Commission. Retrieved from https://www.equalityhumanrights.com/en/publication-download/section-20-investigation- metropolitan-police-service Erfanian, F., Roudsari, R.L., Haidari, A., & Bahmani, M.N.D. (2020). A narrative on using vignettes: its advantages and drawbacks. Journal of Midwifery & Reproductive Health, 8(2), 2134-2145. Fridell, L., & Lim, H. (2016). Assessing the racial aspects of police force using the implicit-and counter-bias perspectives. Journal of Criminal Justice, 44, 36-48. doi: 10.1016/j.jcrimjus.2015.12.001 Ghaffur, T. (2004). Thematic review of race and diversity in the metropolitan police service. Metropolitan Police. London: Metropolitan Police Service. House of Commons (May 2016). Police diversity. First Report of Session 2016–17. House of Commons. Retrieved from http://www.publications.parliament.uk/pa/cm201617/cmselect/cmhaff/27/27.pdf Naqvi, H., Razaq, S. & Piper, J. (2016). 2015 Data Analysis Report for NHS Trusts. NHS Workforce Race Equality Standard Retrieved from https://www.england.nhs.uk/wp- content/uploads/2014/10/WRES-Data-Analysis-Report.pdf Prison Reform Trust (2006). Experiences of minority ethnic employees in prison. Prison Reform Trust. Retrieved from http://www.prisonreformtrust.org.uk/Portals/0/Documents/experiences%20of%20minority% 20ethnic%20employees%20in%20the%20prison%20service.pdf Sehmi, H. S. (2015). An ethnographic study exploring the over-representation of black and minority ethnic (BAME) employees in the disciplinary process in a National Health Service (NHS) Trust. [Unpublished doctoral thesis], University of Surrey, Guildford. Smith, G., Hagger Johnson, H., & Roberts, C. (2012). Disproportionality in police professional standards [online]. Greater Manchester Police. Retrieved from: https://www.escholar.manchester.ac.uk/api/datastream?publicationPid=uk-ac-man- scw:170650&datastreamId=FULL-TEXT.PDF West Midlands Police and Crime Commissioner (2016). Disproportionality in police misconduct proceedings. West Midlands Police and Crime Commissioner. Retrieved from https://www.westmidlands-pcc.gov.uk/wp-content/uploads/2019/05/SPCB-20160105- Disproportionality-in-Police-Misconduct-Proceedings.pdf Wunsch, D., Hughes, C., Hobson, Z., & Yesberg, J. (2016). Disproportionality in Misconduct Cases in the Metropolitan Police Service. MOPAC Evidence and Insight. Retrieved from: https://www.london.gov.uk/sites/default/files/misconduct_research.pdf Appendix 1 – Vignette Example Scenario 1 (Informal) Officer A approaches you, as their line manager, and requests leave on Saturday 21st August 2021, a night shift. You are forced to refuse the request as the team are already below minimum strength through sickness and other abstractions. Officer A subsequently reports sick on Friday 20th August 2021 and resumed on Thursday 26th August 2021, the first day of the next set of shifts. You are approached by a disgruntled team member. They state that it is clear from social media that Officer A was at a family gathering on Saturday evening and not sick. You decide to speak to Officer A. Here, OFFICER A becomes upset and concedes they were at the gathering and called in sick to facilitate their attendance. However, Officer A explains that their partner had threatened to leave them if they did not attend the gathering, as this would have been the third successive family gathering, they had missed through work. As their line manager you are aware that Officer A had recently had other annual leave requests refused and has been experiencing problems in their relationship. Officer A, a 37 year old [ethnicity] male officer is otherwise a diligent, hardworking officer with a good attendance record over his 14 years of service. Appendix 2 – Regression Tables Logistic Regression of Misconduct Allegations B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B) Lower Upper Is_male 0.70 0.09 60.09 1.00 0% 2.01 1.68 2.39 UPP2020 1.74 1.09 2.56 1.00 11% 5.72 0.68 48.36 PO 0.87 0.12 55.23 1.00 0% 2.39 1.90 3.01 SPO 1.86 1.04 3.20 1.00 7% 6.41 0.84 49.14 MSC -0.65 0.26 6.46 1.00 1% 0.52 0.32 0.86 Asian 0.53 0.11 22.16 1.00 0% 1.70 1.37 2.13 Black 0.67 0.14 23.84 1.00 0% 1.95 1.49 2.55 Mixed 0.37 0.16 5.04 1.00 3% 1.44 1.05 1.99 Other 0.37 0.20 3.45 1.00 6% 1.44 0.98 2.12 U25 0.40 0.16 6.30 1.00 1% 1.49 1.09 2.03 LoS22 U2 0.49 0.15 10.34 1.00 0% 1.63 1.21 2.20 LoS 3to5 0.60 0.10 33.88 1.00 0% 1.82 1.49 2.22 LoS 6to10 0.32 0.09 12.69 1.00 0% 1.38 1.15 1.64 Frontline 0.27 0.08 11.23 1.00 0% 1.30 1.12 1.52 Constant -5.55 0.13 1964.58 1.00 0.00 0.00 Substantiated Allegations B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B) Lower Upper Is Male 0.14 0.11 1.66 1.00 20% 1.15 0.93 1.41 Is Gross Misconduct -0.33 0.09 14.80 1.00 0% 0.72 0.61 0.85 Is Asian -0.14 0.13 1.02 1.00 31% 0.87 0.67 1.14 Is Black 0.16 0.15 1.12 1.00 29% 1.18 0.87 1.59 Is Mixed -0.02 0.21 0.01 1.00 92% 0.98 0.65 1.47 Is Other -0.67 0.26 6.42 1.00 1% 0.51 0.31 0.86 Is Senior PO -20.75 20072.24 0.00 1.00 100% 0.00 0.00 . Has under 3 years LoS 0.12 0.12 0.93 1.00 34% 1.12 0.89 1.43 Is under 25 0.47 0.14 11.84 1.00 0% 1.61 1.23 2.11 Frontline - ref 0.69 2.00 71% Met Operations -0.03 0.13 0.05 1.00 82% 0.97 0.76 1.25 Other 0.06 0.14 0.18 1.00 67% 1.06 0.80 1.41 Constant -0.56 0.14 15.25 1.00 0% 0.57 22 Length of Service (LoS) Formal Action B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B) Lower Upper Is Male -0.01 0.18 0.00 1.00 96% 0.99 0.69 1.42 Is Gross Misconduct 0.74 0.15 23.59 1.00 0% 2.10 1.56 2.83 Is Asian 0.15 0.23 0.42 1.00 52% 1.16 0.74 1.83 Is Black 0.07 0.25 0.07 1.00 79% 1.07 0.66 1.75 Is Mixed 0.74 0.38 3.69 1.00 6% 2.09 0.99 4.42 Is Other 0.59 0.51 1.34 1.00 25% 1.80 0.66 4.89 Is under 25 0.18 0.24 0.60 1.00 44% 1.20 0.76 1.90 Police Officer - ref 3.09 3.00 38% MSC 0.07 0.26 0.08 1.00 78% 1.08 0.65 1.78 PCSO 0.71 0.45 2.52 1.00 11% 2.03 0.85 4.88 Other Staff 0.13 0.43 0.10 1.00 76% 1.14 0.49 2.66 Frontline - ref 1.99 2.00 37% Met Operations 0.01 0.22 0.00 1.00 98% 1.01 0.65 1.55 Other 0.28 0.25 1.29 1.00 26% 1.32 0.82 2.13 0-2 years - ref 5.83 5.00 32% 3-5 years 0.80 0.66 1.47 1.00 23% 2.22 0.61 8.05 6-10 years 1.05 0.65 2.62 1.00 11% 2.87 0.80 10.27 11-20 years 0.84 0.65 1.67 1.00 20% 2.31 0.65 8.27 21-30 years 1.10 0.64 3.00 1.00 8% 3.00 0.87 10.41 31 years or more 0.82 0.66 1.55 1.00 21% 2.27 0.62 8.23 Constant -1.05 0.67 2.47 1.00 12% 0.35 1 Beneath the Numbers An exploration of the increases of recorded Domestic Abuse and Sexual Offences MOPAC Evidence and Insight 2 Executive Summary Both domestic abuse and sexual offending represent substantial volumes of recorded crime and demands on the Metropolitan Police Service (MPS) and one that has been increasing over recent years. Within the Mayor’s Police and Crime Plan1 2017-2021, a specific commitment was made in relation to understanding the increases seen in these recorded crimes. Analysis was conducted across a range of data sources. To explore what is beneath the recorded numbers, we have gathered information from a variety of sources – but we have primarily utilised Metropolitan Police Service recorded crime data and Crime Survey for England and Wales (CSEW) data published by Office for National Statistics (ONS) to drive our insights. Our main findings indicate: Has there been an increase in these crimes? • The two primary data sources related to domestic abuse and sexual offences contradict each other in terms of any potential increase. Police recorded offences have significantly increased over time, whilst prevalence as measured by CSEW shows little change • Levels of recorded Domestic abuse and sexual offences across Police Forces in England and Wales has increased in recent years • Prevalence estimates using London level CSEW data indicates that the number of recorded crimes is well below the potential number of victims per year • Wards which are more vulnerable to community stability issues, as per the Vulnerable Localities Profile, experience higher levels of recorded domestic abuse than less vulnerable areas • There are no statistical correlations between population density and volumes of offending for either sexual offending or domestic abuse across the capital • Following the publication of HMICFRS Crime Data Integrity reports in 2014, there was an increase in the levels of recorded violence and sexual offences – two areas that were found to have been previously under-recorded by police across England and Wales Are these increases due to better police recording? • Following the launch of Operation Yewtree there was a slight increase in overall recorded sexual offences 1 https://www.london.gov.uk/sites/default/files/mopac_police_and_crime_plan_2017-2021.pdf 3 • However, there was a clear increase in the number of historic sexual offences coming to the attention of the police following the launch of Operation Yewtree with historic offences recorded by police have increased by 134% over five years • The addition of previously non-notifiable offences such as malicious communications has had a clear impact on recorded domestic abuse • There has been a reduction in the level of ‘no-criming’ of rape offences and a slight increase for other sexual offences with the impact of ‘no-criming’ deemed to be negligible • Younger adult victims appear under-represented in the police data when compared to CSEW estimates • According to CSEW estimates, 1 in 8 single adult women living with children were victims of domestic abuse in the preceding 12 months to the survey compared with 6.4% of those living in a household with no children • For sexual assaults, women who were separated or divorced (7.7%) were the most likely to be victims • There are clear indications that vulnerabilities such as long-term illness or disability and low-income impact on the likelihood of victimisation Is there evidence that victims are more willing to come forward to the police? • Based on the most reliable information from the CSEW, there is no compelling evidence of an increase in victim confidence to report • For domestic abuse and sexual offences, the level of response to CSEW that indicates that people have told police remains consistent over time, with a slight reduction in the case of domestic abuse in the most recent results • Reasons for not reporting are complex and often linked to the seriousness of the offence or the relationship with the offender This report is the culmination of an analytic exercise exploring the increases in police recorded crime for both domestic abuse and sexual offences. Whilst it is not possible to infer causality, the weight of evidence provided by the premises we have highlighted allows us to infer that increases in recorded crime can be primarily attributed to increased compliance with crime data integrity. It is hoped that this analysis and research can drive further insights into the evidence base and inform decision making across London. 4 Introduction The Mayor’s Police and Crime Plan 2017-2021 (PCP) brought with it a specific focus on the victims of recorded crime – and more so those who are most vulnerable. Within the PCP both domestic abuse and sexual offences were the focus of considerable attention, with an additional focus on these crimes also being bolstered by the publication of the Mayor’s Violence Against Women and Girls Strategy 2018-2021. “(MOPAC will) commission specific work to look beneath reported crime statistics for rape and sexual and domestic violence and abuse so that we can say with much greater certainty what changes in figures mean" That is - can we explore these increases in recorded crime and seek to understand what is driving them. This essentially boils down to three questions. Has there been an increase in these crimes occurring? Are these increases due to better police recording? Is there evidence that victims are more willing to come forward to the police? This report explores each of these. Methodology The cross-government definition describes domestic abuse as any incident or pattern of incidents of controlling, coercive, threatening behaviour, violence or abuse between those aged 16 or over who are, or have been, intimate partners or family members regardless of gender or sexuality. The abuse can encompass, but is not limited to: • psychological • physical • sexual • financial • emotional This definition was updated to include coercive behaviour (December 2015) and the recognition that victims of this type of abuse may be aged under 18. Therefore, from April 2013 the way these offences are recorded includes victims aged 16 and above2. Sexual offences, as recorded by police, covers a range of different crime types. These range in severity from non-consensual crimes such as rape or sexual assault involving physical contact, crimes specifically against children including sexual abuse or grooming, and crimes that exploit others. The offences are broadly recorded into two groups: 2 https://www.gov.uk/guidance/domestic-violence-and-abuse 5 • Rape – “A person ‘A’ commits an offence if (a) he intentionally penetrates the vagina, anus or mouth of another person ‘B’ with his penis, (b) ‘B’ does not consent to the penetration, and (c) ‘A’ does not reasonably believe that ‘B’ consents." The classifications of rape recorded by police vary dependent on the gender and age of the victim. • Other Sexual Offences – This is a much broader crime grouping incorporating offences such as contact sexual assaults (various classifications based on age/gender), grooming, sexual activity with a child, child sexual exploitation, abuse of position of trust of a sexual nature, exposure and voyeurism. We have gathered information from a variety of sources – but we have primarily utilised police recorded crime data and Crime Survey for England and Wales (CSEW) data published by Office for National Statistics (ONS) to drive our insights. These are the two primary sources of information related to crime available to us – with each adding to our understanding. It is widely reported that information recorded and held by police can only provide a partial picture of the level of crime experienced by the public. This is particularly the case for domestic abuse and sexual offences where it is believed there is a high level of under- reporting to police. Police recorded crimes are also prone to changes in recording practice and legislation as well as the propensity of victims to come forward. Even with the understanding of the caveats of such data, it does allow us to glean valuable insights in relation to the demands these crimes have on police. The Home Office Counting Rules (HOCR) for recorded crime provides a national standard for the recording and counting of ‘notifiable’ offences recorded by police across England and Wales. The vision of these rules is that all police forces have a consistent approach that delivers accurate statistics that can be trusted by the public and puts the needs of victims at the core. The purpose3 of the HOCR is: • ensure that victims of crime receive the service they expect and deserve; • prioritise effective investigation of crime in keeping with national standards and the College of Policing’s Code of Ethics; • inform the public of the scale, scope and risk of crime in their local communities; • allow Police and Crime Commissioners, Police Forces and local partners to build intelligence on crime and criminal behaviour necessary for an efficient and effective response; 3 Vision and Purpose Statements for Crime Recording (NCRS and HOCR) – Home Office https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/387762/count- vision-december-2014.pdf 6 • enable Government, PCCs, Forces and their partners to understand the extent of demands made on them and the associated costs of service delivery; and • inform the development of Government policy to reduce crime and to establish whether those policies are effective. The CSEW is a longstanding survey across many crime and justice topics covering approximately 32,000 individuals across England and Wales – although has a far smaller sample at the London level. Information on domestic abuse and sexual offences in CSEW is collected through self-completion modules on tablet computers and face to face interviews. This source is hugely beneficial in covering and identifying crimes that are not reported to the police. The CSEW is a reliable source of estimated prevalence and the consistency in methodology remains unaffected by changes in recording practices, police activity or the propensity of victims to report such crimes to the police. To complement the above data and provide richer insights further qualitative research was undertaken with representatives from organisations working with survivors of domestic abuse and sexual violence in London. During August and September 2018 fifteen people were interviewed across several organisations. For the purposes of reporting, we have grouped these into the following categories: • IDVA: Independent Domestic Violence Advocates (IDVAs) from Victim Support, working in a variety of functions/locations across London • Domestic/Sexual Violence Support Organisations: Representatives from organisations supporting survivors domestic abuse and either historic or recent sexual abuse (RASASC, Into the Light, One in Four, Solace Women’s Aid) • Specialist Support Organisation: Representatives from organisations that provide tailored support focused on survivors from specific demographic groups (IKROW, GALOP, MANKIND, RESPOND) These organisations were identified following discussions with colleagues in the MOPAC Violence Against Women and Girls team, and efforts were made to approach organisations that provided support to a range of survivors of domestic abuse and sexual violence. However, it should be stressed that this was a small sample, comprising of those who agreed to participate in the research. Consequently, the themes and findings from these interviews are not presented as definitive conclusions but to add a more subjective view using these perceptions of service providers of domestic abuse and sexual offences in London. Interviewees were asked a series questions4 about their perceptions of trends in the reporting and recording of domestic abuse and sexual violence offences. Themes included respondents’ perceptions of whether there had been any increase in the number of offences of domestic abuse and sexual violence overall, whether there had been an increase in people’s willingness 4 the full interview schedule is appended. 7 to report these offences, whether the respondents had witnessed any changes in the number of survivors accessing their organisation’s services and, if so, why they believed this to be the case. They were also asked to identify the barriers that prevented someone who had experienced domestic abuse or sexual violence from reporting, and whether there were any specific groups of survivors who were less likely to seek help. We also asked for their perceptions of the response to domestic abuse and sexual offences provided by the police and other criminal justice agencies, and about any improvements they felt could be made to provide a better service to survivors. Key questions for the report To better understand what the changes in both police recorded domestic abuse and sexual offences mean, the following key areas were explored: 1. Has there been an increase in these crimes occurring? • Who experiences these crimes? Have the demographics changed of victims changed? • Are there areas of London where there is potential under recording/reporting of these types of offence? 2. Are these increases due to better police recording? • What impact has police activity or legislative change had on the recorded volume of offences? • What impact has reporting and recording of historic sexual offences had on the overall level of these crimes? 3. Is there evidence that victims are more willing to come forward to the police? • Are victims more confident to come forward? • Has demand on survivor organisations increased? 8 Section 1: Has there been an increase in these crimes? The data conundrum Immediately, upon starting this analysis, we encounter what we have termed the ‘data conundrum’ – essentially that the two primary data sources related to domestic abuse and sexual offences contradict each other. In the year ending March 2018, the CSEW indicated that there had been little change over recent years in the prevalence of domestic abuse5 across England and Wales. (see Chart 1) The most recently available data of this type for sexual assaults across England and Wales6 is to end of twelve months to March 2017. The long-term results of respondents aged 16-59 who reported experiencing sexual assault in the year prior to survey shows a slight decline over time from 3% in the twelve months to March 2006 to 2% in the year ending March 2017 (see Chart 2). The estimated prevalence of rape has remained consistent over the last twelve years. 5 Source: ONS Year ending March 2006 to year ending March 2018 Crime Survey for England and Wales - No data point is available for the year ending March 2008 because comparable questions on stalking, an offence that makes up the domestic abuse category, were not included in that year. 6 Source: ONS Year ending March 2006 to year ending March 2017 Crime Survey for England and Wales - The sample size is lower for the years ending March 2011 to March 2013 than for other years due to use of a split- sample experiment in these years. New questions were introduced into the survey from the year ending March 2013 and estimates from this year onwards are calculated using these new questions. Estimates for earlier years are calculated from the original questions with an adjustment applied to make them comparable to the new questions 9 This stands in stark contradiction to police recorded crime data. In terms of domestic abuse, incidents are defined as reports where, after initial investigation, no crime notifiable to the Home Office was committed. Notifiable offences are those that after investigation a crime has been committed. The ONS Domestic Abuse Data Tool7 provides incident and offence volumes across England and Wales between the years to March 2016 and March 2018. The data in this release shows that in the year to March 2018, the volume of domestic abuse offences recorded by police in England and Wales increased by 23% compared to year to March 2017. For the MPS area the increase in the year to March 2018 was lower than the national figure - with an increase of 5% on the preceding year. For sexual offences recorded by police, the ONS release to end of March 2017 showed that at that point there were a total of 121,187 sexual offences recorded by the police in England and Wales in that year. This was an increase of 14% compared to the preceding year and was at the highest level recorded since the introduction of the National Crime Recording Standards were introduced in 2002. In the year to March 2018, this had increased across England and Wales to 150,732 – an increase on the twelve months to March 2017 of 24%. This compares to an increase across the MPS area of 12% in the same period8. 7 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/datasets/domesticabuseinenglandand walesdatatool 8 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/crimeinenglandandwales/ye arendingmarch2018 10 By using the available MPS data over a longer period, between year ending March 2011 and March 2018 inclusive, we can see that recorded domestic abuse and sexual offences remained constant until twelve months to March 2014 at which point sharp increases in both were apparent (Domestic abuse offences showed a 20% increase at end of year to March 2016 compared to previous year, and total sexual offences +32%). Since this date increases have been less pronounced, although still on an upwards trend. Looking at domestic abuse incidents recorded by the MPS, these increased by 17% in the year ending March 2018 compared to the year ending March 2011. However, in the latest twelve months, the number of these recorded incidents marginally decreased compared to the preceding year - a reduction of 3% compared to year ending March 2017 (see Chart 3). For domestic abuse recorded notifiable offences, these have increased at a much higher rate than that seen for incidents. As at the year ending March 2018, total recorded offences were 63% higher than the level recorded in the year ending March 2011. Further, at the end of the last twelve months there was an increase of 5% recorded compared to the year ending March 2017. For sexual offences recorded by police, there is a clear upward trend for both recorded rape offences and other sexual offences over the long term. At the end of the twelve months to March 2018, the recorded total sexual offences recorded by the MPS was 96% higher than at the same point to March 2011 and 12% when compared to the period to March 2017. Rape offences record by the MPS in the twelve months to March 2018 were 140% higher than those recorded in the period to March 2011 (equivalent to over 4,600 more recorded offences) and 21% higher than the twelve months to March 2017. Regarding other sexual offences, the increase in recorded crime was 75% higher than that recorded in the twelve months to March 2011 and 7% higher than the period to March 2017. (see Chart 4). ONS 11 releases have also previously indicated that sexual offences recorded by police across England and Wales started to increase9 from 2012/13. By standardising the long-term data, it is possible to compare the domestic abuse incident and recorded offence trends together. Using a standard score (z-score10) we can explore how many standard deviations from the mean a value is (see Chart 5). 9 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/sexualoffencesinenglandandw ales/yearendingmarch2017#sexual-offences-recorded-by-the-police 10 The Z score is the number of standard deviations from the mean any given data point is. 12 Using this method further illustrates the rising trends for both series – the continued upward trajectory over the last eight rolling twelve-month periods for recorded offences in comparison to the reducing trend over the same period for incidents. We have found that the long-term trends at the rolling twelve-month level for both incidents and offences are statistically significant. In the case of domestic abuse incidents, the trend has been above the twelve-month average since mid-2014, whilst for domestic offences the above average trend is apparent from late-2014. There has also been a clear upward trend in both rape and other sexual offences recorded by the MPS (see chart 6). Using the same method as described above, the key point when recorded offence volumes started to increase was during 2014, with levels since early 2015 exceeding the long-term average. In recent twelve-month periods it is also clear that the trend for recorded Other Sexual offences has slowed, whereas the recorded levels of rape offences continued to increase on a rolling twelve month-basis. 13 Estimated prevalence of domestic abuse and sexual assault in London ONS have published London level data from CSEW on both domestic abuse and sexual assault. To increase base sizes, the sample for analysis covered the years ending March 2016 to March 2018 inclusive11. The estimates are based on the self-completion module on domestic abuse, sexual assault and stalking, which is restricted to adults aged 16 to 59 years living in households in London12. Using the Mid-Year 2017 population estimate 13 for London and the London level CSEW results, we can estimate that approximately 246,700 adults aged 16 to 59 years who live in London experienced any form of domestic abuse in the year prior to completing the survey. This equates to a prevalence rate of approximately 4 in 100 adults between these ages. Women were more likely to have experienced domestic abuse than men (5.9% compared with 2.9%). This equates to an estimated ~164,000 women and ~82,000 men. 11 Crime Survey for England and Wales, Office for National Statistics https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/adhocs/008805crimesurveyenglandan dwalesestimatesofsexualassaultanddomesticabuseexperiencedbyadultsaged16to59 12 The upper age limit for completing the self-completion module was increased to 74 from April 2017, but to ensure all data included in this analysis are comparable, only 16 to 59-year olds are included 13 https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/ populationestimatesforukenglandandwalesscotlandandnorthernireland 14 Further to this, and by using the same methodology we can estimate that 168,000 adults aged 16 to 59 years experienced specifically partner abuse14 in the year prior to survey, equating to a prevalence rate of 3 in 100 adults. Women, as seen with overall domestic abuse, are more likely to have experienced partner abuse than men (4.5% compared to 1.5%). This equates to an estimated ~125,000 women and ~42,000 men who had experienced partner abuse. For sexual assault, we can estimate that 151,000 adults aged 16 to 59 years experienced any form of sexual assault (including attempts) in the year prior to completing the survey, which is a prevalence rate of approximately 3 in 100 adults. Women were more likely to have experienced sexual assault than men (4.1% compared with 1.3%). This equates to an estimated ~114,000 female victims and ~37,000 male victims. Women were more likely to have experienced indecent exposure or unwanted touching in the year prior to survey (3.6% compared to 1.2% of men). This means that an estimated ~100,000 women and ~28,000 men had experienced indecent exposure or unwanted touching. The percentage of women and men that had experienced rape or assault by penetration in the preceding year were much smaller (1% of women and 0.1% of men). This means an estimated ~30,600 adults had experienced these offences in the year prior to the interview – ~27,800 women and ~2,800 men. We can use these estimates to illustrate the differences between this estimated prevalence and police recorded crime. By using the estimate of prevalence for the population of London aged between 16 and 59 years that experienced any form of domestic abuse in the year prior to survey (4 in 100); applying this to the estimated mid-year population for this cohort across the last five financial years and comparing to the recorded victims of domestic abuse 15 by police we can hypothesise the gap between the number recorded by the police and the number of victims that may have experienced this type of crime as inferred by the responses to CSEW (see Chart 7). This echoes what we know about under-reporting of these incidents. 14 non-physical abuse, threats, force, sexual assault or stalking committed by their partner 15 Data only from 2013/14 as this is the point that the definition of DA was extended to include 16 and 17 year olds 15 Has there been any change in where these recorded crimes occur? We know from previous analysis that by mapping the volume of offences across the 32 boroughs of London that this does not necessarily indicate the impact of offending on the population. The mid-year population estimate by ONS16 for 2016 shows a rounded figure of 8.7m residents living in the jurisdiction of the MPS. Across London boroughs there is wide variation in population levels – Barnet has a population of ~386,100 people compared to less than half of that number in the Royal Borough of Kensington and Chelsea (~156,700). Domestic abuse across the capital In the twelve months to March 2018, there were 9.0 Domestic Abuse Offences per 1,000 population across the 32 boroughs17 . Although Croydon and Newham boroughs recorded the highest volume of recorded offending for domestic abuse, their rate of offending per 1,000 population was lower than other boroughs such as Barking and Dagenham, Greenwich and Haringey. 16 https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates#publicati ons – this has been used as this allowed ward level comparison 17 Population figures based on 2016 mid-year estimates 16 Barking and Dagenham recorded the highest rate of offending in the year to March 2018, a rate of 12.6 offences per 1,000 population. Kingston Upon Thames has the lowest volumes of offending as well as the lowest rate. Since the twelve months to March 2016, Barking and Dagenham and Greenwich have had the highest rates of offending for Domestic Abuse in the MPS (see map 1) 18. While Greenwich had seen variations in the rate on a year on year basis, Barking and Dagenham had a reduction in year to March 2017 compared to the previous year, however this has since returned to similar levels as seen previously. When we explored the relationship between the volume of recorded domestic abuse and those electoral wards that are vulnerable to community stability issues (Vulnerable Localities Profile19) we found that wards with the highest volumes of recorded offending generally are those most vulnerable to these issues – that is to say that six in ten of the highest volume wards for recorded offences were in the ‘most vulnerable’ groupings20. Furthermore, for all offences recorded across the MPS, just under 30% happened in the most vulnerable wards compared to just 11% in the least vulnerable. Statistical analysis identified that across London, there was a strong correlation between the volume of offending and the 18 Both having rate of offending approximately 2 standard deviations from the mean 19 an index that incorporates different sets of data such as crime, deprivation, health, educational attainment and population density 20 Offending at least 2 standard deviations from the mean and VLP 0-20% Map 1 Map 1 17 level of vulnerability of wards. This correlation was also marginally stronger when vulnerability was compared to rate of offending per 1,000 population. In short, we can therefore surmise that wards which are more vulnerable to community stability issues experience higher levels of recorded domestic abuse offending. 21 We also tested if there was any relationship between population density in areas and the recorded offences – this found that there was generally no correlation between population density and offending levels. Sexual offending across the capital Sexual offences recorded across the MPS in the twelve months to March 2018 was highest in Westminster, both in terms of volume and rate per 1,000 population (see Map 2). These findings are believed to be influenced by the high level of footfall in this borough (workday population and the night time economy) and the relatively low residential population. This is further supported by other inner London boroughs having a significantly higher rate of offending when compared to outer London boroughs. The volume of sexual offences recorded by the MPS has a positive correlation with the population size and density of a borough. By attempting to minimise the effect of Westminster’s high volumes, the strength of this correlation increases when the borough is excluded from the calculation; this is true for both rape and other sexual offences. 21 Using Pearson’s Cor Test on rate of offending per 1,000 population: Df = 568; p-value < 0.05; cor = 0.74 Map 2 18 Across the 32 boroughs we observed that in the twelve months to March 2018 Hammersmith & Fulham, Hounslow and Lewisham all had reductions in recorded sexual offences compared to the previous year; only Lewisham had a recorded a reduction in rape offences. At the lower geographic level of electoral ward, the highest volume of recorded sexual offences occurred in areas within boroughs that have higher levels of footfall. Offences were highest in St. James’s and West End in Westminster, as well as Stratford and New Town in Newham and Fairfield in Croydon. These areas are town centres, generally popular for socialising, attracting residents from surrounding areas, tourists and others from further afield. However, in areas such as these, where offending levels are high, rates per population is an inaccurate method of establishing a rate of offending as population at times of generally high offending will be significantly different to the residential population of the area. There is a very weak to no statistical correlation between population density and volumes of offending across recorded sexual offending. Unlike the strong correlation seen with domestic abuse recorded by police, sexual offences recorded in the twelve months to March 2018 had a weak positive correlation to areas vulnerable to community stability issues. Each year after the period to March 2015 the boroughs experiencing the largest increase in Rape Offences differ, however the boroughs with the highest number of offences remained the same; these are Croydon, Lambeth, Newham and Westminster. In the twelve months to March 2018, Westminster recorded the highest volume of rape Offences, and the highest increase in volume from twelve months to March 2015. Boroughs such as Camden, Greenwich, Hackney and Islington also experienced significant increases in offending. Section one - Key Findings • The two primary data sources related to domestic abuse and sexual offences contradict each other. Police recorded offences have increased over time, whilst prevalence as measured by CSEW shows little change • Prevalence estimates using London level CSEW data indicates that the number of recorded crimes is well below the potential number of victims per year • Wards which are more vulnerable to community stability issues experience higher levels of recorded domestic abuse • There are no statistical correlations between population density and volumes of offending for either sexual offending or domestic abuse across the capital 19 Section 2 Are increases due to better police recording? Have improvements to police compliance with standards driven increases in police recorded crime? So far, analysis indicates the disparity between the increasing Police recorded crime and the stable findings of the CSEW. It is possible to explore the increases within Metropolitan Police Service recorded offences in more detail. As can be seen in charts 8 & 9, two major milestones – the launch of ‘Operation Yewtree' and an improvement in compliance with National Crime Recording Standards (NCRS) - may have had an impact across both domestic abuse and sexual offences recorded by police. In November 2014 Her Majesty’s Inspectorate of Constabulary (HMIC)22 released a report of an inspection into crime data integrity in police forces across England and Wales23. This report highlighted that, at that time, over 800,000 crimes reported to police go unrecorded each year – approximately one in five of all reported crimes. This was found particularly to affect victims of violence and sexual offences. For violence against the person (VAP) under- recording rates were assessed as to be as much as one third. This same report suggested that sexual offences were under-recorded by 26% across England and Wales. Indeed, as at the end of the twelve-months to March 2015, following the publication of the inspection report by HMIC, domestic abuse offences across the MPS increased by 21% and Sexual Offences by 32% when compared to the same period to March 2014. Follow up inspections by HMICFRS indicate that improvements have been made but varies force to force. It is therefore likely that any increases in relation to improved compliance with recording may continue for some time. 22 Now Her Majesty’s Inspectorate of Constabulary and Fire and Rescue Services (HMICFRS) 23 ‘Crime-recording: making the victim count’ © HMIC 2014 ISBN: 978-1-78246-660- 4https://www.justiceinspectorates.gov.uk/hmicfrs/wp-content/uploads/crime-recording-making-the-victim- count.pdf 20 Following the launch of Operation Yewtree, the investigation into historic sexual offences, in October 2012 there was a period whereby sexual offence volume did not see immediate increases across the MPS. However, by end of the twelve months to March 2013 these offences had increased by 13% when compared to the preceding year. 21 Additionally, in respect of sexual offences, ONS published detailed information24 on this topic in February 2018 which covered the period to end of March 2017. Within this publication, ONS stated that “high-profile coverage of sexual offences and the police response to reports of non-recent sexual offending is another factor that is likely to influence police recording of sexual offences. For example, Operation Yewtree, which began in 2012, and more recently, allegations by former footballers”. The report further states that this may influence victim’s willingness to come forward for both historic and more recent offences. As seen earlier in this report, standardised trends of rape and other sexual offences recorded by the MPS show a similar trend. Both types of offence increased since 2014 - however the increase in rape offences appears to pre-date the HMICFRS report being published (see chart 10). Rape offences have increased in volume since year to March 2013. Between then and the following period to March 2014 the volume of rape offences increased by almost a third; with the volume recorded in the period to March 2018 being almost two and a half times larger than that in the period to March 2013. As indicated, improvements in recording practices were implemented following the HMICFRS reporting the year to March 2015 and volumes recorded in the following year were 21% higher. Overall, volumes recorded in 2017/18 were 54% higher than those in 2014/15. It must also be recognised that the MPS Crime Data Integrity inspection 201825 conducted by HMICFRS ended with the service being judged as ‘Good’. The report stated that “the Metropolitan Police Service has made concerted efforts to improve crime-recording accuracy” and that the service has a “commitment to accurate crime recording that is victim- focused and free from performance pressures of any kind”. The MPS are now recording 87.6% of reported violence against the person accurately and 91% of reported sex offences. 24 ONS Sexual Offences in England and Wales: year ending March 2017 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/sexualoffencesinenglanda ndwales/yearendingmarch2017 25 https://www.justiceinspectorates.gov.uk/hmicfrs/publications/metropolitan-police-service-crime-data- integrity-inspection-2018/ 22 The impact on improved compliance on domestic abuse recorded by police HMICFRS report of November 2014 highlighted that Violence Against the Person (VAP) was a classification of crime that was significantly under recorded by police across England and Wales. We know from the police recorded crime to end of March 2018, that total VAP recorded by the MPS was 24% higher than at the same point to March 2015. The proportion of recorded domestic abuse offences identified as VAP accounted for 81% of the total recorded domestic abuse in year ending March 2018; this has remained consistent in each twelve-month period since year ending March 2012. Considering the impact that the increased compliance with recording has clearly had on overall VAP levels recorded by the MPS, it follows that this has also impacted on the recorded levels of domestic abuse VAP. The impact of changes to the Home Office Counting Rules We notice that although VAP accounted for most of the domestic abuse flagged offences over time, over the last eight twelve-month periods to end of March, the proportions that each minor classification contributes to domestic abuse VAP has shifted. Chart 11 illustrates the clear increase in the proportion that is attributed to the Harassment minor classification. In year ending March 2012, this category accounted for just 16% of all domestic abuse VAP compared to 25% in year ending March 2018. HOCR for recorded crime are constantly updated, whereby offences recorded by police that had previously been ‘non-notifiable’ become ‘notifiable’. This essentially means that crimes 23 that were previously not ‘counted’ are subsequently included in the police recorded crime data sent by forces to the Home Office. This is pertinent to the levels of recorded Harassment domestic abuse offences as several additional offences have been added as per HOCR in the past few years. Harassment, as a minor classification of VAP, incorporates numerous crime types. Presently these are: • Breach of a Restraining Order • Breach of the Conditions of an Injunction (protection from Harassment Act) • Causing Harassment, Alarm or Distress • Causing Intentional Harassment Alarm or Distress • Disclose private sexual photographs with intent to cause distress (added 2015) • Fear of Provocation of Violence • Offence of Harassment • Pursue Course of Conduct In Breach Of Sec 1 (1) Which Amounts To Stalking (added 2012). • Putting People in Fear of Violence • Sending letters etc with intent to cause distress or anxiety *includes electronic messages (added 2015- ‘Malicious Communications’) • Stalking Involving Fear of Violence (added 2013) • Stalking Involving Serious Alarm/Distress (added 2013) 24 The impact of the inclusion of previously non-notifiable offences is that at the end of the period to March 2018, the total number of the total recorded Harassment offences had increased by 166% (over 9,800 more recorded offences in twelve months) when compared to twelve-months to March 2012. Table 1 shows that the key driver in the increase across domestic abuse related harassment appears to have been the inclusion of ‘Malicious Communications’ as a notifiable offence. In the twelve months to March 2016, ‘Offence of Harassment’ accounted for 35% of all recorded offences within the harassment classification – previously this category accounted for over 75% of all harassment offences recorded by the MPS. In the same period, malicious communications accounted for 44% of all recorded harassment offences whereas previously this was not included. It is also clear that the inclusion of malicious communications offences has impacted not just recorded domestic abuse related harassment and, in effect, VAP; but also, the overall levels of domestic abuse offences recorded by the MPS. In the twelve months to the end of March 2018, this specific category of offence accounted for almost 9% of all domestic abuse recorded by the MPS. This demonstrates that legislative changes and improved compliance to recording standards has contributed to the increase in police recorded domestic abuse offences. Although respondents during our interviews of providers, were uncertain about increases in domestic abuse and sexual violence overall, several respondents felt from their experience that increases had occurred for specific offences. Although based on six responses, reference was made to on-line and social media offences, particularly offences of harassment and malicious communications, and that these were particularly marked for young people. Classification - expansion 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 Breach of a Restraining Order 8% 5% 6% 5% 5% 6% 6% Breach of the Conditions of an Injunction: Protection from Harrassment Act Only 3% 3% 2% 2% 2% 2% 2% Causing Harassment, Alarm or Distress 3% 1% 0% 0% 0% 0% 0% Causing Intentional Harassment Alarm or Distress 5% 6% 6% 6% 6% 5% 6% Disclose private sexual photographs with intent to cause distress 0% 0% 0% 0% 1% 2% 2% Fear of Provocation of Violence 6% 6% 5% 5% 5% 4% 4% Offence of Harassment 73% 78% 78% 76% 35% 39% 32% Pursue Course Of Conduct In Breach Of Sec 1 (1) Which Amounts To Stalking. 0% 0% 0% 1% 1% 1% 2% Putting People in Fear of Violence 2% 1% 2% 2% 1% 1% 0% Sending letters etc with intent to cause distress or anxiety 0% 0% 0% 0% 44% 40% 44% Stalking Involving Fear of Violence 0% 0% 0% 1% 0% 0% 1% Stalking Involving Serious Alarm/Distress 0% 0% 0% 1% 1% 1% 1% Total Harassment Offences 5,925 7,488 8,380 11,570 14,563 15,140 15,742 Table 1 - % of all harassment by classification expansion Year to March 2012 to March 2018 25 Although this is by no means definitive, this subjective information from those assisting people who have experienced domestic abuse, seems to reflect what we have seen following the inclusion of malicious communications as a notifiable offence. The impact of reclassification and ‘no-criming’ of sexual offences Through our analysis we wanted to test if there had been any change in the way in which sexual offences were recorded by police – here we were specifically testing whether reclassification occurred i.e., rape offences reclassified as other sexual offences and vice versa. What we found was that in the twelve months to March 2018 7% of sexual offences were reclassified compared to the initial classification that was attributed at the time of recording; broken down as 6.5% of Other Sexual Offences reassessed as Rape Offences, and 4.9% Rape offences reassessed as Other Sexual Offences. What we also saw was for both rape and sexual offences, reclassifications of this type remained unchanged since year ending March 2015. This indicates that the MPS appear to have remained consistent in their assessment and recording of such cases during this period. We also tested to see if there were any changes in the type of sexual offence recorded by police – however, we found that there has been no shift in this. The hypothesis here was to test if there had been an increase in non-contact sexual offences (voyeurism, exposure etc.) recorded by police that may have attributed to the overall increase. Since year to end March 2015 Sexual assault on a Female aged 13 and over and Rape of a Female aged 16 and over, account for more than half of all Sexual Offences recorded. The third highest is Exposure and Voyeurism, which accounts for approximately 10% of all recorded sexual offences. Although these types of offences have remained consistent as a proportion of total sexual offences, their volume from year to year have seen significant changes. Compared to end of March 2015, the volume of Sexual Assault on a Female aged 13 and over has increased by 26%, while the volume of Rape of a Female aged 16 and over have increased by 61%. Chart 12 – Proportion of rape by minor classification 26 As an illustration, between year to March 2015 and to March 2018 the largest category of rape offences was Rape of a Female aged 16 and over, accounting for approximately 7 in 10 of all rape offences. During this period, Rape of a Female aged 16 and over increased by 61%. This is significantly higher than any other category and accounted for 77% of all increases during this period. Police forces across England and Wales record some crimes which are then subsequently ‘no crimed’ where they have judged that no crime had taken place. The HOCR set out the circumstances where crimes may be ‘no crimed’. These include where a crime was perhaps recorded in error or where additional, verifiable information becomes available that determines that no crime had taken place. Crime reports that are ‘no crimed’ are not included in police recorded crime statistics and are therefore not part of the annual data return of crimes submitted by police to the Home Office. In the twelve months to end of March 2018, there were approximately ~1,100 recorded sexual offences that were subsequently ‘no-crimed’ by the MPS (see chart 13). This was made up of ~760 other sexual offences and ~330 recorded rape offences. As a proportion of the total initially recorded offences, ‘no crimed’ equated to 5% of all recorded sexual offences, 6% of all other sexual offences and 4% of rapes. This has reduced from 10% in the twelve months to March 2012 for all sexual offences, 8% for other sexual offences and 13% of rapes. 27 In terms of volumes, there has been a reduction in rape offences that are subsequently ‘no crimed’, yet a slight increase in ‘no crimed’ other sexual offences. Therefore, more offences of rape are remaining classified as such than in previous years, yet more other sexual offences have been judged to not have been notifiable offences. The reduction in ‘no crimed’ records for rape was clearly apparent from the end of twelve months to March 2015 – indeed, over the course of the following twelve months ‘no crimed’ rape offences reduced by 40%. As at the end of March 2018, ‘no crimed’ rape offences had reduced by a third when compared to the end of March 2012. We can therefore infer that, whilst a clear positive for survivors, the reduction in ‘no-crimed’ rape has had a negligible impact on the overall levels of recorded sexual offences by the MPS. This is further supported by the increase in recorded rape offences as at the end of March 2018 being 134% (over 4,000 more recorded offences) compared to March 2012. The impact of recorded historic sexual offences We know from previous ONS publications, that the launch of Operation Yewtree and the associated media coverage was considered to have had some form of impact on the propensity of victims of non-recent sexual offences to come forward. In the recent release by ONS into sexual offences in England and Wales26 there was an increase in non-recent sexual offences recorded by police (those that took place longer than 12 months ago) in the twelve months to March 2017 of 16% compared to the previous period. Further, over the last five years represented, the number of non-recent offences recorded by police had more than tripled. However, the rises seen in these non-recent offences is in line with all sexual offences recorded, with non-recent sexual offences representing a similar proportion of the total number recorded as in previous years. This means that recent sexual offences have also increased and are the main contributor to the total level of sexual offences recorded by police across England and Wales. Using MPS recorded crime data, we have explored the potential impact of historic reporting on the overall levels of recorded sexual offences across London. To identify these, the difference between ‘recorded’ and ‘committed on/from’ dates was used27. By defining non- recent sexual offences in the same way as ONS, i.e. those that had been recorded twelve months or more after the date committed, it is possible to indicate the changes in non-recent offences over time. Also, by comparing the most recently recorded data to March 2018 with the period to March 2013 (following the launch of Operation Yewtree) we can see that the number of recorded offences that occurred at least twelve months prior to the recorded date had increased by 134%. However, it is also apparent that over the same time recorded ‘recent’ offences (i.e. 26 ONS Sexual Offences in England and Wales: year ending March 2017 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/sexualoffencesinenglanda ndwales/yearendingmarch2017 27 Only those offences whereby a ‘committed on/from date’ was recorded were used for this analysis 28 those that had occurred less than twelve months previous) had also doubled with both recent and non-recent offences having a clear upward trend in terms of recorded crime. We also observed that the proportion of total recorded sexual offences that are non-recent has remained consistent since year ending March 2013 (see chart 14) and represents approximately one in five recorded offences. This is an increase in proportion from one in seven offences being non-recent in the period28 prior to the launch of Operation Yewtree. Therefore, the proportion of recent offences continue to make up the largest proportion of recorded sexual offences. It may be that the media coverage of high-profile historic sex offences, may have also impacted on the likelihood of victims to report recent offences. However, from the recorded crime series, it is not possible to categorically state this to be the case. The potential role of media coverage was echoed through our interviews with stakeholders - who generally reported to us that there had been an increase in such historic cases particularly in the number of men coming forward, with the perceived driving factor being increased coverage, particularly following the Savile case. Has there been any change to who experiences these offences? To better understand what is beneath the numbers for both domestic abuse and sexual offences, we need to better understand who are more likely to be victims of these crimes. As with the general data conundrum discussed at the beginning of this report, we also see some divergence between the people who indicate victimisation via the self-completion module of the CSEW and the victims recorded by police. 28 Twelve months to end of March 2012 29 As we know from the prevalence estimates, women are far more likely than men to experience both types of victimisation. This is also the case for the recorded domestic abuse and sexual offences, whereby the proportion of women and men that experienced these offences has remained consistent over time. However, as part of the publication29 of London level responses to CSEW for these, it is possible to further infer the personal characteristics of victims of domestic abuse and sexual assaults. What ages are more likely to experience domestic abuse and sexual offences? From CSEW results for London, younger age groups, for both men and women, experienced a higher prevalence of domestic abuse (see chart 15) 30. Particularly, women aged between 16 and 19 and between 20 and 24 were more likely to be victims of any domestic abuse (9.5% and 9.7% respectively) compared with those in the 55 to 59 aged group (4.1%). This can also be seen in the case of younger men, with the 16 to 19 age group having a higher prevalence than those in the 55-59 age group (5.7% compared to 3%). These findings compare to almost half of all victims (irrespective of gender) being aged between 20 and 34 years of age (45%) in the police recorded crime series in the last year. The largest proportion of these were in the 25 and 34 years of age bracket. This potentially reveals insights into specific age groups that require more support or focus to come forward. 29 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/adhocs/008805crimesurveyenglan dandwalesestimatesofsexualassaultanddomesticabuseexperiencedbyadultsaged16to59 30 Ibid 30 However, there is differentiation seen across the age ranges in relation to victim’s gender (see chart 16). This illustrates that as the age of victims increase, the proportions shift in that female victims account for a smaller proportion. The police recorded crime series also showed that despite the recognition of 16 and 17-year olds as victims of domestic abuse since year ending March 2014, this has had negligible impact on the overall volume of victims of recorded domestic abuse offences. Each year since March 2014, 16 and 17-year olds accounted for between 2% and 3% of all recorded victims of domestic abuse. This therefore indicates that domestic abuse recorded by police may not represent the true scale of the impact on these younger victims. 31 Similarly, to the domestic abuse findings from CSEW, women aged between 16 and 19 and between 20 and 24 were far more likely to be victims of any sexual assault (17.5% and 8.5% respectively) compared with those in the 55 to 59 aged group (1.9%). This is also case for victims of indecent exposure or unwanted touching with 16.8% of women aged between 16 and 19 and 7.7% of women aged between 20 and 24 experiencing this type of incident in the preceding year. For rape or assault by penetration (including attempts) 7.5% of women aged between 16 and 19 indicated they had experienced this type of offence (see chart 17). More than half of female victims of sexual offences recorded by police were aged between 17 and 34 years, yet more than a quarter of female victims of recorded rape offences were aged between 16 and 21 years old. This is like the CSEW findings whereby the highest proportion of respondents having experienced rape in the year prior to survey were younger. The CSEW cohort for this report is restricted to only those victims aged between 16 and 59 years of age. However, we know from the police recorded crime, that those aged 16 and under accounted for a quarter of all recorded victims of sexual offences recorded in the twelve months to March 2018. 32 What is the marital status of those who experience domestic abuse and sexual offences? Adults who were separated and divorced at the time of interview as part of the Crime Survey for England and Wales, were most likely to be victims of any domestic abuse in the preceding 12 months. This is true for both men and women. Data for those who had been widowed was not available (see chart 18). In terms of the relationships identified between domestic abuse offenders and victims (where data is available) from the police recorded crime, the majority are of an intimate or ex- intimate nature. Almost seven in ten relationships recorded during the twelve months to March 2018 were within this category. Almost a quarter of all identified relationships were where the offender was the ‘Ex-Boyfriend’ of the victim. Familial relationships were less frequent, with close family members most frequently recorded in this grouping– 28% of all familial relationships indicated that the offender was the son of the victim. This was reflected in responses during our research interviews, that service providers perceived an increase in violent offences by children towards parents and grandparents. For those who had experienced sexual assault (see chart 19) in the year preceding the survey also were most frequently separated or divorced (7.7%). However, this was closely followed by those who were single (7.1%). 33 Single women were slightly more likely to have experienced indecent exposure or unwanted touching followed by divorced women (6.3% and 5.7% respectively). No data was available for widowed victims. Although marital status is not an identifiable field within the recorded crime data it is possible to comment on the general findings of the relationships between suspects and victims for sexual offences. In the twelve months to March 2018 we can see that over half of all sexual offences, where a suspect has been identified, that the victim and suspect are known to each other in some way. Intimate or ex-intimate relationships were recorded in 5% of all recorded other sexual offences where a relationship was identified yet almost one in five for rape offences. What is the experience for those with long-standing illness or disability? For domestic abuse, the CSEW results indicate that those with a long-term illness or disability were more likely to be victims of any domestic abuse in the preceding 12 months than those without a long-term illness or disability. This was true for both men (4.7% compared with 2.7%) and women (18% compared with 4.5%). One in ten women who indicated a long-standing illness or disability had experienced any form of sexual assault compared to just 3 in 100 for those without long standing illness or disability. 34 Does household structure and income feature? The results from CSEW related to domestic abuse also allows us to comment on the household structure and income of the respondents who had experienced this type of abuse in the twelve months preceding interview. The findings show that just under 1 in 8 single adult women living with children (13.4%) were victims of domestic abuse in the preceding 12 months to the survey compared with 6.4% of those living in a household with no children and 3.7% of those living in a household with other adults and children. In terms of income, almost 4 times as many women in the lowest income bracket (less than £10,000 per annum) had experienced domestic abuse in the preceding 12 months (12.8%), compared with those in the highest household income bracket (£50,000+ per annum = 3.7%). Section 2 Key Findings • Following the publication of HMICFRS Crime Data Integrity reports in 2014, there was an increase in the levels of recorded violence and sexual offences – two areas that were found to have been under-recorded by police across England and Wales • Following the launch of Operation Yewtree there was a slight increase in recorded sexual offences • The addition of previously non-notifiable offences such as malicious communications has had a clear impact on recorded crime • There has been a reduction in the level of ‘no-criming’ of rape offences and a slight increase for other sexual offences with the impact of ‘no-criming’ deemed to be negligible. The MPS appear to have remained consistent in their assessment and recording of sexual offences that come to their attention • There was a clear increase in the number of historic sexual offences coming to the attention of the police following the launch of Operation Yewtree with historic offences recorded by police have increased by 134% over five years. • Younger adult victims appear under-represented in the police data • 1 in 8 single adult women living with children were victims of domestic abuse in the preceding 12 months to the survey compared with 6.4% of those living in a household with no children • For sexual assaults women who were separated and divorced (7.7%) were the most likely to be victims • There are clear indications that vulnerabilities such as long-term illness or disability and low-income impact on the likelihood of victimisation 35 Section 3: Is there evidence that victims are more willing to come forward to the police? Although police recorded crime has seen significant increases for both domestic abuse and sexual offences, as outlined the prevalence as measured by the CSEW has remained stable. This evidence thus far indicates that these increases are likely to have been heavily impacted by better compliance on behalf of police forces in recording crimes that come to their attention. It is not possible from the police recorded crime data to infer any increased confidence to report to police by those who experience these crimes. Such a measure is not captured within police data - which makes the question as to victim willingness a difficult one to explore. However, from their own assessments, ONS publications indicate that the increases seen in recorded domestic abuse are likely to be impacted by both increased compliance with recording standards and a willingness to report. In exploring this issue, we can draw upon CSEW data which asks victims about prevalence of reporting to police, as well as our own interviews with stakeholders. What evidence is there of increased willingness of individuals reporting to police? The strongest data in terms of levels of reporting sexual assaults and domestic abuse comes from the victims themselves in the CSEW (and previously the British Crime Survey – see table 2). Although, even though this is the strongest data, it is not without flaws. However, these questions are not routinely answered and so we only have snapshots across many years. For sexual offences, the proportions of victims that stated they reported to the police has remained stable since 1999/2000, and stable for domestic abuse from 2014/15. Essentially, there is no evidence within the CSEW of any increase in the proportion of victims reporting to police. As outlined, this data is the best proxy of victim willingness to report we have. In the case of domestic partner abuse, approximately 17% of respondents told the police about their experience over the last twelve months (to March 2018). This is a reduction from 21% when this was last asked as part of the survey in year ending March 2015. There has also been a reduction in the proportion of respondents telling anyone about their experiences – from 81% in year ending March 2015 to 72% in the twelve months to March 2018. For sexual assault, an average of 17% of respondents told the police over the time series available and there has been little change in respondents having told anyone. 1999/2000 (BCS) 2013/14 2014/15 2015/16 2016/17 2017/18 1999/2000 (BCS) 2013/14 2014/15 2015/16 2016/17 2017/18 CSEW -Victims of Sexual Assault 70% 58% 68% 18% 17% 17% CSEW - Victims of Domestic Partner Abuse 81% 72% 21% 17% Public Attitude Survey (London) – All victims 75% 74% 72% 71% 70% 69% 66% 65% Table 2 - Who respondents told following incident Survey Source Told Anyone Told Police 36 In addition, we know from police recorded crime, for example, how offences come to the attention of police remains largely unchanged over time. More than four in five (83%) recorded domestic abuse offences in the twelve months to March 2018 were reported via a phone call to the police, while one in ten recorded offences were reported at a police station. Whilst not covering sexual assault or domestic abuse, findings from the MOPAC Public Attitude Survey present a similar picture in terms of overall levels of reporting to the police. This survey covers over 12,000 Londoners per year, and as such is a larger sample of Londoners compared to the CSEW. As a useful proxy, PAS indicates around 65-70% of victims told the police they had experienced any form of crime – a figure that has been relatively stable over the past four years – albeit a small drop in the financial year 17/18. Slightly more respondents from the PAS report they told anyone of the crime – again a figure that has been slightly reducing over the previous four years. Our own research interviews asked our sample of service providers whether they have seen an increased willingness on the part of survivors to report to the police. Views were mixed overall, and instead of an increased willingness to report to police, respondents spoke about the perceived improvements in the knowledge and willingness of victims to seek out wider support. In this way, many of our respondents felt that demand on their own organisation had increased 31, in terms of the number of referrals to their organisation, increases in caseload over time and in the number of clients on their waiting lists (for those organisations that provided counselling for example). There were mixed views as to the reasons behind these increases, responses generally did not think it was being influenced by 'more crime' per se, but issues relating to the media (i.e., historic abuse by celebrities), better liaisons with wider services, or increases in the number of staff members employed by the support organisations themselves. Another strong theme to emerge from the interviews around the perceived increase in referrals was the improved location and accessibility of services, such as the location of IDVAs in hospitals, their co-location with the police, the use of out-reach workers (an example cited was an individual who worked with probation hubs, contacting female ex-offenders who have just left prison), and work with organisations that support female sex-workers. As touched upon earlier, one respondent at interview suggested the increased media coverage following the Savile case, for both historical and non-recent reporting, conveyed the message to survivors that the police wanted to talk to victims about the issue, and that there is help available. Furthermore, another example was provided where a victim had come forward regarding historic abuse after seeing media coverage – the provider stated that this had made the victim realise that they had been a victim of abuse whereas they had never thought of their relationship as abusive before. 31 Thirteen of fifteen respondents 37 What are the potential reasons for not reporting to police? There has been a wealth of academic research exploring the potential reasons for non- reporting to police of such sensitive crimes. In a large-scale review Voce & Boxall (2018) explored 21 quantitative victim self-report studies conducted in Australia, New Zealand, UK, USA and Canada since 1980 related to the reporting experiences of victims of domestic abuse reporting that ‘victim and offender characteristics are less reliable predictors of reporting than characteristics of the relationship and the incident itself’. That is, incidents that involve serious violence, an intoxicated offender or a child witness are more likely to be reported. Voce and Boxall (2018) also summarises many of the consistent barriers to reporting, such as: • Fear of not being believed by the police; • Desire to protect the offender; • Fear of negatively impacting the family (e.g. removal of children, dissolution of the relationship); • Economic dependence on the perpetrator; • Privacy concerns; • Fear of exposing their own illegal activities; and • Fear of retribution and the escalation of violence From our own review of literature various factors which may be important in understanding why survivors of sexual offences are less likely to report their experiences to the police were identified. Overall, research suggests that sexual violence between intimate partners is less likely to be reported to the police than stranger cases, particularly amongst male victims (Gartner and Macmillan, 1993). It is worth noting however, that some research has shown women who have experienced violent crime are more likely to report to police when victimised by their partner (Kaukinen, 2002). Psychological factors and the relationship with survivors reporting cases of sexual violence to the police is heavily discussed throughout the published literature. Notions of shame, blame and fear are consistently recognised as the most widely experienced barrier to reporting (Ceelen, 2016). In addition to the psychological barriers, choosing not to report sexual violence has been shown to sustain a survivor’s social identity, their relationships and educational prospects; thus, a desirable option as opposed to reporting to police (Khan et al, 2018). The extent to which a sexual offence case may be perceived as ‘real’ or ‘believable’ is also influential to a survivor’s decision to report their experience to the police (Brown et al, 2007). Rape that occurs alongside another crime has been suggested to increase the extent to which a survivor feels their allegation will be believed and consequently influences their likelihood of reporting (Addington and Rennison, 2008). Similarly, cases of sexual violence involving weapons or resulting in injury have been shown to increase police reporting, whereas those involving drugs or alcohol had the opposite effect (Fisher at el, 2003; Woltzky-Taylor et al, 2011). 38 In addition, we asked interviewees during our research to indicate reasons why they felt victims were reluctant to report victimisation, and to indicate any perceived gaps in reporting. Interviewees spoke of the stressors of the CJS - our interviewees raised issues such as a lengthy court process, cross-examination and a perceived lack of 'people focus' in the courts as well as the perceived quality of treatment from the police of such victims - as factors that could discourage victims from reporting. London’s Victims Commissioner has also conducted work with women and girls directly affected by domestic abuse who have experienced barriers in receiving support because of their immigration status32 leading to reticence to report the abuse to authorities in fear of being deported. Our interviewees also raised some wider groupings where they perceived more support was required - such as: • BAME victims may be less willing to come forward potentially influenced by their wider perceptions towards the police; • Males being less likely to come forward for fear of being viewed as weak; • How the potential fear of being ostracised or isolated from some communities may lead some not to report. The recently published analysis by ONS into sexual offences in England and Wales33 shows that although there are low levels of reporting to police, almost half of those victims that report did so because they wanted to prevent it happening to others or that it was the right thing to do. Conversely, the most frequent reasons for victims to tell anyone but not to tell the police were due to embarrassment, feeling that the police couldn’t help or that it would be humiliating to report. 32 https://www.london.gov.uk/press-releases/mayoral/protect-victims-of-crime-with-insecure-status 33 Sexual offences in England and Wales: year ending March 2017 ONS – February 2018 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/sexualoffencesinenglanda ndwales/yearendingmarch2017#reporting-of-sexual-assault-by-rape-or-penetration Section 3 Key Findings • Based on the most reliable evidence from the CSEW, there is no compelling evidence of an increase in victim confidence to report • For domestic abuse and sexual offences, the level of response to CSEW that indicates people have told police remains relatively consistent over time • Reasons for not reporting are varied and complex and often linked to the seriousness of the offence or the relationship with the offender • As shown in both interview responses and the response from surveys for victims who ‘told anyone’ - it appears victims are looking for support, rather than looking to report. However, this support appears to be more frequently sought from friends or family 39 Conclusions The ‘data conundrum’, highlighted by the consistent estimated prevalence provided via the subjective responses to the CSEW versus the steep and continuing increases in the objective measures of crime recorded by police, is not easily reconcilable. Estimating the prevalence in the London population also indicates that there is still a large gap between the level of recorded crime and those that people experience. There is evidence that changes to police activities, compliance with standards and changes to Home Office rules for counting crime are most likely to have made the largest impact on the levels of crime recorded. Indeed, it must be stated that the committed efforts of the MPS following the HMICFRS report of 2014 have resulted in independently assessed measurable success in improving crime data integrity. The latest report graded the MPS as ‘good’ in the overall judgment of the service and that the MPS “has made concerted efforts to improve crime-recording accuracy since our 2014 crime data integrity inspection report. Importantly, we found a commitment to accurate crime recording that is victim-focused and free from performance pressures of any kind.”. In the summary, HMICFRS highlighted that the MPS had implemented all recommendations from the previous report and that the force achieves good levels of recording accuracy for reported sexual offences. These inspection results seem to validate our main findings that the levels of recorded crime now better reflect those incidents coming to the attention of police, rather than an increase in the prevalence of such incidents. Furthermore, our own findings in relation to sexual offences recorded by the police, indicate consistent use of ‘no-criming’ in relevant situations and that the ‘downgrading’ of offences does not occur, indicating a consistent approach being applied by those investigating and recording crimes of this nature. However, it is difficult to extract from the recorded crime data as to what extent better compliance leads to improved confidence in victims of crime coming forward to report. ONS stated in their latest publication of November 2018 into domestic abuse across England and Wales, that the increase in recorded crime may be driven, in part, by an increased willingness for victims to come forward. It could be that the police response and the availability and awareness of other services may have impacted on the number of incidents coming to police attention. Although the estimates provided by CSEW responses shows no significant increase in respondents reporting to authorities, we also cannot dismiss the potential impact recent public campaigns such as the #MeToo movement may have had on victim and witness willingness to report incidents to police or others. It follows that an increased awareness amongst the public of issues of violence and harassment against women and girls would lead to less tolerance of domestic abuse and sexual offences going unreported. Also, as our research found, through increased publicity some survivors have come to realise that they may have been in abusive relationships, where they had not previously thought of these as such. We have also found that, in addition to the improved compliance with recording practices, other factors such as highly publicised investigations and convictions related to historic sexual abuse also impacted on recorded levels of crime. This saw a sharp increase in historic 40 offending coming to the attention of the police, and an increased proportion of offences per year being of a non-recent nature. Through this exploration we can say with more clarity who these crimes affect and where they occur. Although domestic abuse is experienced across all areas of London, our analysis shows that victims of domestic abuse who live in areas of London which are vulnerable to community stability issues experience more offences than those who live in areas that do not. These more vulnerable areas are often located in Inner London boroughs and areas with higher levels of deprivation. Although the vulnerable localities profile uses multiple indicators, this seems to chime with the findings from the characteristics of domestic abuse victims via the CSEW – particularly in terms of household characteristics. This could also be the case for those areas with low community stability issues – more often areas that are more affluent - with CSEW results showing that prevalence among those who earn more is reportedly lower. However, this may also indicate less willingness of those experiencing these offences in these areas to report to police. It is also of note that CSEW results indicate that those who live with long -term illness or disability report experiencing domestic abuse at a higher level than those who do not have these issues. Therefore, more needs to be done to understand the nature of domestic abuse in these situations and to ensure our responses to these vulnerable victims is appropriate. We know from our own analytics that more vulnerable locations experience more offences. Crimes are not equally spaced but cluster in specific areas - and as such many of these areas will not be sampled adequately by the CSEW, especially at the London level where the target interview range34 is between 3,800 and 3,900. There will also be sampling error, individuals not wanting to disclose victimisation, certain types of address are not included (i.e., residents of care homes are not interviewed) and so on. So, whilst the CSEW is the strongest data on victim reporting, it is not without limitation. Similarly, for sexual offending, although there is no correlation to areas with community stability issues, offences occur across London. However, unlike domestic abuse that occurs more within private premises, sexual offences often occur in public areas or parts of the capital where Londoners and others from further afield congregate for tourism, social and leisure purposes. We know from our own interviews and the wider academic literature that the reasons for reporting, or not, to the police or other authorities are complex and varied. Often, these reasons are personal to the victim and are rooted in emotional responses to the events such as embarrassment or shame. Moreover, the literature also suggests that often reasons for not reporting are linked to the seriousness of the offence or the relationship with the offender. Ultimately, regardless of better police capture of information or more confidence in reporting via data discussions and CSEW, one aspect that is undeniable is the increase on police demand over recent years. These are often serious offences and behind each of these police reports is an individual and in many cases these individuals may be vulnerable. The focus moving 34 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/methodologies/crimeandjusticemetho dology 41 forwards should be that the findings of this analysis will feed into MOPAC’s and our partners’ approach to tackling VAWG. Clearly, we need to focus our joint efforts on the areas of concern, to the people affected and to influence the willingness of survivors of domestic abuse and sexual offending to come forward – be that to report or to seek out support. 42 Appendix: MOPAC Interview Schedule (Introduction – background to research, research outputs, consent issues) 1. Could you tell us a bit about your background, and your role at [insert organisation]? Prompt if not spontaneously mentioned: - How long have you worked in this role? - What roles did you do before coming to this organisation? - Could you tell us a bit about what your day-to-day work involves? - In what capacity do you normally interact with survivors of domestic abuse/sexual violence? 2. Could you tell us about services provided by your organisation, and the types of survivors you work with? Prompt if not spontaneously mentioned: - What services does your organisation provide for survivors of domestic abuse/sexual violence? - Does your organisation provide services for all victims of domestic abuse/sexual violence, or does it work specifically with certain groups? - What types of people tend to seek help from your organisation, for example victims of certain crime types, or people from specific backgrounds or demographic groups? - At what point do survivors tend to come into contact with your organisation (for example, before or after they have reported to the police, at the time of an offence or after?) 3. Would you say there have been any changes to the number of survivors accessing your organisation’s services over the last few years [or time in role]? If necessary clarify: For example, have the number of survivors increased or decreased, or have you not seen any change? Prompt if necessary (if interviewee feels there has been a change): - When did you start seeing this increase/decrease? - What, if any, impact has this increase/decrease had upon your organisation? - What do you feel may be driving this increase/decrease? 4. Would you say that there have been any changes in survivors’ willingness to come forward to seek help for domestic abuse/sexual violence over the last few years [or time in role]? If necessary clarify: For example, do you feel that survivors have become more or less willing to seek help, or would you say that there has been no change? Prompt if necessary (if interviewee feels there has been a change): - Over what timeframe have you seen this change? - Have these changes been seen for any specific types of domestic abuse/sexual assault victims, or would you say they are across the board? - What impact do you think this increase/decrease in willingness to come forward has had upon survivors? - What impact do you think this change in willingness to come forward has had upon your organisation/the police? 43 5. What do you feel may be the reasons behind this change in victims’ willingness to seek help? Prompt if not spontaneously mentioned: - Do you feel there has been any impact of recent high-profile cases/media coverage/social media movements upon victims’ willingness to come forwards to seek help; this could be either positive or negative? (For example the Harvey Weinstein case or the #MeToo movement). - Do you feel there have been any changes in public confidence in the police or support organisations? - Do you feel there have been any changes in the levels of service provided by these agencies? - Do you think these changes have affected any specific groups of survivors more than others, for example those from certain backgrounds, or those who have experienced certain types of domestic abuse/sexual assault? 6. Thinking more widely, have you noticed any other changes in the types of survivors that access your organisation’s services over recent years [or time in role]? Prompt if not spontaneously mentioned: - Have you seen any changes in the types of crimes that victims are seeking help about? - Have you seen an increase in reported cases of harassment and malicious communications? - Have you seen any changes in victim demographics? - Have you seen any changes in victims’ previous experiences of domestic abuse/sexual assault, for example historical vs. recent offences? - Have you seen any changes in the types of pathways through which victims are referred? - Have there been any changes in the types of services offered by your organisation? 7. In general, how well do you feel the police respond to domestic abuse/sexual violence crimes? Prompt if not spontaneously mentioned: - How effective do you feel the police are at bringing offenders to justice? - How effective do you feel the police are at supporting vulnerable victims/safeguarding victims from harm? - How well do you feel the police work with other organisations (e.g. support organisations or the wider CPS) to support survivors of domestic abuse/sexual violence? - Are there any gaps or opportunities in the service provided for victims of domestic abuse/sexual violence? 8. What do you feel may be the key barriers that prevent someone who has experienced domestic abuse/sexual violence from seeking help? Prompt if not spontaneously mentioned: - Are there any differences in these barriers for reporting to the police compared with reporting to support organisations? 44 9. Do you feel there are any specific groups of survivors who may be less likely to seek help than others? Prompt if not spontaneously mentioned: - Are there any demographic groups that may be less likely to seek help from the police or support organisations about their experiences? - Are there any offence types where you feel survivors may be particularly unlikely to come forward for help? - What do you feel may be driving these differences in the likelihood for victims to come forward? 10. What improvements do you feel could be made to help provide a better service for victims of domestic abuse/sexual violence? Prompt if necessary: - For example, any improvements to services provided by the police, CJS or support organisations? (Thanks for participating in the research. Contact details of researchers) 1 The Lighthouse: 9-month interim evaluation report September 2019 Rachael Parker, Tim Read & Paul Dawson MOPAC Evidence & Insight 2 Contents Executive summary 3 Early findings to date 3 1. The Lighthouse evaluation 6 Methodology 6 2. Performance review 8 Performance data sources 8 Referrals to the Lighthouse 8 Initial assessments at the Lighthouse 9 Consent 9 Background of clients who consented to evaluation 11 Service allocation and addressing vulnerability 12 Offence details 13 Summary of performance review 14 3. Learning from early implementation 15 General perceptions of the Lighthouse 15 Consistency to the developed model 16 Ways of working 18 Referrals and demand 21 Partnership working 22 Challenges for the future 23 4. Discussion 25 Next steps 26 References 27 Appendices 28 Appendix 1: Focus group interview schedule, April 2019 28 Appendix 2: Professional survey, March 2019 29 Appendix 3: Case study of progression through the Lighthouse 31 Appendix 4: Number of referrals by borough 34 Appendix 5: Data from case management system on service user vulnerabilities 35 Appendix 6: Anecdotal information from staff on reasons for creating service 36 allocations on the case management system 3 Executive summary The Lighthouse, London’s Child House, opened at the end of October 2018 as part of a two- year pilot. Bringing together a range of organisations under one roof, the Lighthouse intends to be a child friendly, multidisciplinary service for victims1 of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden, it will replace the existing services to date and serve the five surrounding North Central London boroughs of Barnet, Camden, Enfield, Haringey and Islington. The Lighthouse intends to offer more enhanced features than previous services, with the foremost aim to be focused on the child. The Evidence and Insight (E&I) Unit are MOPAC’s in-house social research and analytical team and were commissioned to evaluate the Lighthouse. The E&I evaluation focuses on four distinct areas for analysis; a performance review; a process evaluation; impact evaluation and an economic evaluation. This report concentrates on the first two areas, looking at the first 9 months of operation of the Lighthouse.2 Early findings to date Process: Summary of performance review The data used for the performance review came from two main sources. The first was provided by the Lighthouse data officer and included aggregate data around referral month, borough and referral source, in addition to age categories and gender of the referrals. The second, and more comprehensive, data source was individual-level data produced from Excelicare, the case management system for the Lighthouse. This data was only for clients who specifically consented to have their data included in the MOPAC evaluation; and due to the amount of data provided (over 300 variables) this will enable the evaluation to present a more richly detailed picture of these clients and the services they receive3. The Lighthouse received 276 referrals between the end of October 2018 and June 2019, which is an average of 35 referrals per month. Most referrals originated from Camden, Enfield and Islington, who each referred 21% (n=58 or 59) of all referrals. Children’s Social Care made over half of all referrals (53%, n=146). There were no notable differences in the type of referrals received across the five boroughs. Out of the 276 referrals, 164 reached the stage of an initial assessment at the Lighthouse, which is where support begins for the clients. If these levels were matched for the remainder of the year, the Lighthouse would receive around 420 referrals, and would carry out 252 initial assessments within its first year. These two figures are somewhat lower than the estimated demand (i.e., 700 young people a year4 with support to 544 children and young people a year) but over twice the number of referrals in the CSA Hub in 2017 (118 referrals per year).5 It is unclear at present, why the number of referrals is lower than initially predicted. This may be because the original estimation was too high, that the original police data was not the most suitable to base predictions upon or could be linked to wider factors such as the communication across referrals agencies, or that it takes time for a service to be established and build awareness. The evaluation will explore this more going forward. 1 Referred to as victims throughout the remainder of the report 2 Details of the rationale for the establishment of the Lighthouse, and E&I’s overall evaluation approach are contained in E&I’s first evaluation report, published in April 2019 3 Several questions emerged when working with the data drawn from the case management system. Particularly, there were issues in the consistency of data entry among staff, as well as limitations to using the system itself. These issues will be worked upon moving forward with Lighthouse staff. However, the data herein is the best snapshot of learning there is at present regarding the Lighthouse. 4 Conroy et al. (2018) used police data to estimate the potential demand. 5 NHSE (2018) 4 Most children and young people (CYP) referred to the service were female (78%, n=216) and just over half (55%, n=151) were between 13-18 years of age (overall age range was between 0-18). Of the 276 referrals that reached an Initial Assessment (n=164), 92 consented to their data being used for the evaluation (therefore the rate of consent at the time of writing is 56%). Upon examining this data, it is evident that this is a highly vulnerable group of CYP with complex needs. For example, 27 were assessed at being at risk of further sexual abuse; over a third of service users have anxiety and/or depression (n=34), and 13 were reported as being at risk of suicide. The results of the Adverse Childhood Experience Questionnaire (ACE-Q) further highlighted that the service users (a sample of 54) have experienced a disproportionate amount of adversity in their lives so far; sexual abuse (n=42), parents divorced (n=29), domestic violence against parent (n=21) and emotional abuse (n=18) were among the most frequent experiences reported. Furthermore, the nature of the offences that brought the CYP to the Lighthouse show that most of the abuse was intra-familial (n=36) which reinforces the sheer lack of a safe or stable household for many of the service users. Process: Summary of learning from early implementation Findings related to the process of the Lighthouse were drawn across several sources, including: interviews with 7 Lighthouse practitioners, focus groups with 15 individuals (all but one were staff members), and written feedback from one other staff member, a survey with professionals who work within the 5 pilot boroughs which received 75 responses6, and finally feedback from service users and parents collected via paper questionnaire by Lighthouse staff. Service users and parents reported feeling very positive with the service they had received at the Lighthouse, that the staff engaged well, were easy to talk to and took their views seriously and sought to help effectively. The facilities themselves and the wider convenience of the location/appointment was also raised as a positive. Staff felt that the Lighthouse was achieving or working toward achieving its objectives. They recognised the benefits of having different agencies working together under the same roof – both in terms of benefits for the service user (access to different services on the one site, quicker access, and reduction of the need to repeat their story) and practitioners (easier access to other agencies, different areas of expertise available to them, and exposure to different cultures/understandings). Equally there was a recognition that having different agencies working together created some tensions (around terms and conditions of employment, working cultures, and operational norms) which has been identified before in partnership projects, and staff highlighted the plans being established to reconcile this in order to develop ‘the Lighthouse way’. The first nine months has seen some developments and changes in operational practice at the Lighthouse, principally around opening hours, service user’s circumstances following referral, and the initial assessment process. Some staff members also mentioned that the roles they were currently undertaking differed substantially from what they had anticipated at the outset of the project. This appeared to be the case particularly for the advocates, Social Care Liaison Officers and Police Liaison Officers. 6 The response rate was low. The survey was sent by E&I to 322 school contacts. It was also sent to 22 contacts within the NHS, social care and voluntary sector, who were also requested to send the survey onto other contacts within their area. Finally, it was also sent to police safeguarding contacts in North Area, North West, and Central North although it is unknown how many contacts were reached. It is not possible to accurately calculate overall response rate. 5 Certain elements of the Lighthouse model that were anticipated to form part of the service have been delayed and are yet to be implemented. Some of these are due to factors beyond the Lighthouse’s control (i.e., Section 28 and the Live Link) and others have been subject to minor delays. For example, it was intended that the service would be providing psychologist-led (rather than police-led) Achieving Best Evidence (ABE) interviews, whereas, the training for the psychologists has taken longer than anticipated and psychologist-led interviews have only recently begun. Lighthouse staff expressed some ongoing challenges pertaining to the Lighthouse’s infrastructure. The strongest theme to emerge here was the reported lack of sound-proofing throughout the Lighthouse building (i.e., within interview suites, medical rooms or the staff kitchen) which was an issue for staff and was highlighted as a risk to confidentiality. There were also some issues raised around the limitations of the case management system that had been introduced, with being described as ‘clunky’ and ‘cumbersome’ and reports of it being prone to crashing. Staff also raised the issue of capacity of the service, (particularly the mental health service provision) and whether existing service levels would be able to be maintained in the future (particularly if the number of self-referrals increased). They suggested several ways that this demand might be managed more effectively, such as ensuring appropriate referrals, more effective triage, and sign-posting individuals to alternative services outside the Lighthouse, and groupwork, when and where appropriate. This is the second in a series of MOPAC E&I Lighthouse evaluation reports to be released, enabling learning both internally as a catalyst for improvement, and externally to advance the evidence base. Future reports are expected in summer 2020, and a final evaluation report in 2021 which will aim to explore the effectiveness of the Lighthouse initiative against criminal justice, health and wellbeing outcomes. 6 1. The Lighthouse evaluation The Lighthouse, London’s Child House, opened in October 2018 as part of a two-year pilot. Bringing together a range of services under one roof, the Lighthouse intends to be a child friendly, multidisciplinary service for victims and survivors of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden it replaces the existing services7 and serves the five surrounding North Central London boroughs of Barnet; Camden; Enfield; Haringey; and Islington. The Lighthouse builds on the CSA Hubs, but offers enhanced support to children and young people (CYP) aged between 0 – 18 years old (or those aged over 18 to 25 years of age with learning delay or disability for whom a child or young person-oriented service appears more suitable)8, as well as non-offending parents/carers/family for up to two years. The service was commissioned to be open extended hours (initially Monday to Saturday 10:00 to 20:00 and by outreach Sunday 10:00 to 13:00) and is provided by University College London Hospitals NHS Foundations Trust (UCLH) in partnership with the Tavistock and Portman NHS Foundation Trust and the National Society for the Prevention of Cruelty to Children (NSPCC). It offers a joined-up approach where, if required, service users can get access to medical, social care, police, advocacy and therapeutic support ‘under one roof’ (for example, the service has two dedicated Metropolitan Police Service (MPS) liaison officers and two Social Care Liaison officers (SCLOs) working from the building). Although recognised as logistically challenging, this approach is paramount to providing a smooth and efficient service to the CYP and one of the unique elements of the model. Methodology The Evidence and Insight (E&I) Unit is MOPAC’s in-house social research and analytical team which has been commissioned to undertake an evaluation of the Lighthouse. The two-year evaluation will cover the processes of the Lighthouse (from design through implementation), monitor routine performance, as well as seek to robustly explore impact and cost benefit. As previously documented, the primary outcome for the Lighthouse broadly cover: • Enhanced referral pathways into and out of the Lighthouse • Enhance CYP, family and carer experience of support received post disclosure • Enhance CYP experience of the criminal justice process post disclosure • Enhance mental health and well-being outcomes for CYP • Enhance professional awareness, competence and confidence • Increased likelihood of charge or conviction for those cases within the Lighthouse • Enhance partnership working • Providing CSA victims care and support to reduce the long-term impact of victimisation The evaluation uses a mixed methodology approach – balancing qualitative context from staff, stakeholder or client feedback, particularly in the shorter-term, with the ‘harder’ performance figures indicating how the service is running on a day-today basis. It focuses on four distinct areas; performance monitoring; process; impact and economic analysis. The ability to successfully complete each element will depend on the quality and quantity of data and will be reviewed throughout the life of the research, as it is subject to change. Details of the evaluation plan are provided in E&I’s first evaluation report. 7 NB CYP Havens will continue to provide the acute/Forensic Medical Exam (FME) service. 8 Although no referrals were received for those aged between 18 & 25 during the period covered by this report. 7 Over the course of the two-year pilot there are four E&I evaluation reports planned, of which this is the second. Taking a broad action research perspective - findings from the evaluation are routinely fed back to the programme teams, the academic advisory group set up to advise the evaluation, to update partners at the official Partnership Oversight Board and other relevant meetings to ensure learning is continually shared within an active feedback loop. This report focuses upon the performance monitoring and process aspects of the evaluation, exploring the first six to nine months of the Lighthouse’s operation and draws from the following data: • Performance management data. This includes aggregate performance data provided by the Lighthouse as part of its reporting to the service commissioners. This provides an overview of service delivery between go-live (October 2018) and June 2019 (numbers of referrals, sources of referrals, number of assessments etc). In addition, data taken from Excelicare (Lighthouse’s case management system (CMS)) provides details of the service received by individuals at the Lighthouse, and about the individuals themselves. However, it should be noted that these data only relate to individuals who have consented to participate in the evaluation, a subset of those who have received the service overall. The number of individuals to consent is a potential evaluation risk and is explored in greater detail in the results. • Focus groups and interviews were held in April and May 2019 with Lighthouse staff to explore their perceptions of the early implementation of the Lighthouse (see Appendix 1 for focus group/interview schedule). Data were collected from 23 individuals (all but one were staff members); 15 people attended one of two focus groups, in-depth face to face interviews were undertaken with 7 practitioners, and written feedback to the interview questions was obtained from a practitioner with whom it proved impossible to arrange an interview. The staff who participated were from a range of occupations and included paediatricians, advocates, psychologists, admin staff, social care liaison officers and police liaison officers. • An online survey to stakeholders designed by E&I distributed in April and May 2019 to professional stakeholders (e.g., police, social care, medical, and education) who work within the five Lighthouse boroughs and who may have come across CYP who have experienced CSA/E (see Appendix 2 for survey questions). The survey in total includes 17 questions and covers themes such as confidence in identifying and addressing CSA and CSE, knowledge and awareness of the Lighthouse service, and if applicable experience of the Lighthouse service. There was a total of 75 respondents from several occupations, predominantly schools and police but also included some medical professionals.9 There are two further evaluation reports planned to track the lifecycle of the Lighthouse; another interim report planned for summer 2020 and a final evaluation report in 2021. The final report will aim to examine the impact of the Lighthouse in terms of criminal justice, health and welfare outcomes, comparing the Lighthouse to a control site. 9 The response rate was low. The survey was sent by E&I to 322 school contacts. It was also sent to 22 contacts within the NHS, social care and voluntary sector, who were also requested to send the survey onto other contacts in their area. Finally, it was also sent to police contacts in North Area, North West, and Central North however it is unknown how many contacts were reached. It is not possible to calculate overall response rate. 8 2. Performance review Performance data sources This section seeks to present a picture of the Lighthouse service in terms of its internal workings and processes, throughput, activities, and the demographics of its clients using data such as: - Numbers of referrals - Referrals by month - Number of initial assessments - Referring borough - Referral source - Service allocations - Client demographics - Client vulnerabilities and disabilities The evaluation team had two main data sources for this report. The first was provided by the Lighthouse data officer and which included aggregate data around referral month, referring borough and referral source, as well as the age categories and gender of the CYP. The second, and more comprehensive, data source was individual-level data produced from Excelicare. This data was for all clients who consented to have their data used in the MOPAC evaluation. At the point of writing this report, this was 92 individuals. This data was much more detailed and provided over 300 variables, although at this stage of the evaluation not all fields have been completed. Appendix 3 provides further detail of the journey and process of support a service user receives from the Lighthouse, through a fictional case study. Referrals to the Lighthouse Between the end of October 2018, when the Lighthouse launched, and the end of June 2019 there were a total of 276 referrals to the service, which works out as an average of around 35 referrals per month (see Figure 1). This number peaked (with 42 referrals) in December 2018; and otherwise remained generally stable across the first 9 months. At this level, the Lighthouse is on track to receive around 420 referrals within its first year, somewhat below the original estimation of 700 (Conroy et al, 2018)10. Camden, Enfield and Islington were the highest referring boroughs, each referring 21% (n=58-59) of all referrals (more details of referrals by borough are provided in Appendix 4). Children’s social care made over half of the referrals to the service (53%, n=146), whereas medical sources (GPs, A&Es and sexual health clinics) made up only 5% (n=15) of referrals. Referrals from the CSA Hub make up 10% (n=28) of the referrals, however the majority of these were in the first three months of opening, and there have only been two referrals from there since February.11 The majority of CYP referred were female (78%, n=216), and just over half (n=151) were in the higher age bracket of 13-18 years. overall, there were no notable differences across boroughs or organisation in terms of who they were referring (such as gender and age variation). 10 It is unclear why the number of referrals is lower than initially predicted. This may be because the original estimation was too high, that the original police data was not the most suitable to base predictions upon or could be linked to wider factors such as the communication across referrals agencies or awareness of the service. 11 This pattern of CSA referrals is due to the transference of CSA cases to the Lighthouse service upon opening. 9 Initial assessments at the Lighthouse Once the CYP have been referred to the service, the referral applications are assessed by one of the two Social Care Liaison Officers (SCLOs), who may then need to gather further information from the referrer. After discussing (for example, whether an initial assessment is appropriate for the CYP at this time) the referral at the daily allocation meetings, an action plan is made which includes assigning the case to the lead practitioner. The CYP is then invited in for their assessment by the initial assessment (IA) team (and these become allocations on the case management system). Between the end of October 2018 and the end of June 2019 the service carried out 164 IAs. After October (where 1 IA was conducted)12, the level of IAs remained consistent with an average of 21 per month. If this continues, the Lighthouse will carry out 252 IAs within its first year, which (as discussed before in terms of overall referrals) is somewhat lower than the 544 CYP that the service was commissioned to support in one year.13 Moving forward the evaluation will track throughput set against the predictions and seek to understand this distance in more depth. For the remaining 112 referrals that did not reach an IA between October and June, there were several reasons documented (although 50 did not have a reason recorded14) - to illustrate, 23 have since attended an IA; 11 are pending further information; 10 CYP and/or parents/carers did not want the service; 5 referrals did not meet the criteria; 5 received a consultation-only service; 3 are pending CYP availability; 1 CYP is not ready yet; 1 has ongoing court proceedings; 1 has ongoing consultation; 1 was referred elsewhere and 1 received Havens support. At each IA there may be several practitioners in the room,15 in the first nine months of service there were on average three professionals present at each IA. Most frequently this profession was a Paediatrician (n=79), followed by a Play Specialist (n=51). A category of ‘other’ was present for 44 IAs, and this includes individuals such as Social Workers, foster carers, or family members. At the IA stage, additional information is recorded about the CYP on the case management system by the Lighthouse staff (to supplement any information gathered at the referral stage), including information on medical history, vulnerabilities and disabilities and further assessments. Analysis of these details will be presented within the client background section coming up in this report. Consent It is also at the IA stage that the staff ask the CYP for consent to engage in their services, and in addition, this is also the point where consent is gained for their data to be incorporated in the evaluation. The issue of consent is crucial, without it, the evaluation will not be able to access personal data nor measure some of the key outcomes (particularly around the impact of the service) and as such low levels of consent would hinder the analytic ability of the evaluation, which could influence future decision making around the Lighthouse.16 Over the first 9 months of the evaluation a total of 92 CYP gave consent for the evaluation. Outside of the first month of October (which in terms of consent rate is an outlier given the low 12 The service opened to referrals on 23rd October, therefore it was unlikely to start seeing children until after at least a week in order to give them enough notice to attend. 13 As outlined earlier, this difference could be related to a variety of factors. 14 This is due to data quality issues which are being addressed by the Lighthouse and their Data Officer. 15 These may include a paediatrician, play therapist, advocate, sexual health nurse and clinical psychologist. 16 See Conroy et al. (2018) report for more detail on the Lighthouse evaluation and ethics. 10 number of clients at this point; n=1), through November to February the consent rate was an average of 24%. To promote consent levels, in February 2019, the evaluation team worked closely with Lighthouse staff on the issue (i.e., conducted a training session around consent and the evaluation and providing visual aids for the staff to use with clients to explain consent). After the session, and for the following months (i.e., February through June) consent levels increased (an average of 66% for these months; see Figure 1 for consent rate). Consent levels will continue to be monitored throughout the evaluation. Figure 1. Total number of referrals, IAs, number of services users who consented to the evaluation, and consent rate per month Once consent is given by the service user, this enables the evaluation team to use individual-level data to conduct more in-depth analysis, exploring throughput in more detail as well as presenting a broader picture around the background and needs of the Lighthouse clients. For example, drawing upon cases that consented17 we can identify that on average there were 31 days between a referral and an IA (the median was 23.5 days).18 Before exploring the individual-level data in more detail below, the evaluators conducted some preliminary analysis to compare the general characteristics of those that did consent to those that did not, to explore any potential biases (i.e. certain client types may have been more likely to not give consent, which if so would limit our analysis). These two groups were compared (at an aggregate level) on age category, gender, borough, and referral source. Overall, there were no significant differences between the two groups. Even with this comparability, at this stage limitations remain given the relatively low number of service users, and so there are limitations about the ability to generalise from this group to the overall Lighthouse cases. None-the-less, there is still value in exploring these characteristics to generate insights. Findings from this specific analysis on those to consent is outlined below. 17 This is based upon 89 cases for whom this data was available. Although 92 service users consented, the data captured for these individuals will vary due to data quality issues around the case management system. 18 For service users overall, the Lighthouse reported that for the period from May to July 2019 the average time from referral to IAs was 47 days (range 11-182 days) compared to 20 days the previous quarter. 0% 20% 40% 60% 80% 100% 120% 0 5 10 15 20 25 30 35 40 45 Oct 2018 Nov 2018 Dec 2018 Jan 2019 Feb 2019 Mar 2019 Apr 2019 May 2019 Jun 2019 Consent rate for evaluation Number of service users Referrals IAs Consented to evaluation Consent rate 11 Background of clients who consented to evaluation This section will provide a summary of the demographics, backgrounds, and needs for the service users who consented for their data to be used in the evaluation. Demographics Of the 92 service users, the majority were female (n=72); and their ages ranged between 2 to 18, with an average age of 11 years. The most common age was 16 years. Ethnicity data was recorded for 67 service users and there was an almost equal split between BAME and non-BAME clients (n= 32 and n= 35 respectively). From a total of 89 clients; most (n=73) were in full time education. Vulnerability Data presented in this section is taken from three sources: a self-report questionnaire conducted by Lighthouse staff (the Adverse Childhood Experiences Questionnaire (ACE-Q); practitioner assessments of vulnerability; and practitioner assessments of future risk. Looking across these in summation we see (as expected) a highly vulnerable client group19. The first source, The Adverse Childhood Experience Questionnaire (ACE-Q), is an internationally validated self-report tool encompassing 10-items across 10 areas which cover household dysfunction (parental separation/divorce, parental domestic violence, parental substance misuse and mental illness, and parent incarceration), child abuse (sexual and physical), and child neglect (emotional and physical). 20 The more events that a person experienced before the age of 18, the higher their ACE ‘score’ will be, and literature demonstrates that the higher the score (a maximum of 10) the greater the risk of health issues (i.e., mental or physical), substance misuse, victimisation and offending in adulthood.21 This emphasises the importance of providing holistic, integrated support to these young people to mitigate the risk of health and lifestyle problems in adulthood. As a baseline, a nationally representative survey of adults in the UK found that 46% of respondents reported at least 1 ACE, and 8% reported at least 4.22 Comparatively, in a study that looked at vulnerable young people with mental health problems in Scotland (who present serious harm to others), there was a much higher prevalence of ACEs; 93% (out of 130) had experienced at least 1 ACE, and 59% had experienced at least 4.23 At the Lighthouse, the ACE-Q was completed by 54 service users24, whose ages ranged from 3- 18. For these service users, the ACE scores ranged from 0-9, with an average of 3. The majority 19 The number of service users recorded within each of the three data sources varies widely (again, which is due to data quality issues that arose in the first few months of service). Therefore, at this stage the analysis of this data has been limited and we cannot cross-reference individual clients across these three data sources. As such, this section presents a summary of the three data sources separately to provide an overall picture. 20 The ACE questionnaire has been used internationally and the original ACE study used a 10-question tool (Felitti et al. 1998), however the ACE-Q has sometimes been adapted by other organisations or researchers and has either been shortened or lengthened in terms of the number of items (Bethell et al. 2017) 21 Felitti et al. (1998); Hughes et al. (2017); Zarse et al. (2019) 22 This was undertaken by Bellis et al. (2014) with 3885 18-69-year olds in the UK. There have also been many other studies, in various populations and nationalities, which have also shown that most adults (between 52%-75%) have experienced at least one ACE (Zarse et al, 2019) 23 Vaswani (2018). 24 ACE-Qs should be completed for all clients, however in the first few months of opening data completion was an issue as the service adjusted to using a new patient record system and ways of working. It is expected that this will improve as the service continues. 12 (n=48/54) of service users had a score of at least 1, and almost half (n=24) had a score of 4 or more which literature has considered to be ‘high risk’.25 This places the Lighthouse ACE scores above non-clinical national populations, but closer to other at-risk youth populations. The most common ACE for the Lighthouse service users was sexual abuse (n=42/54), followed by parents divorced/separated (n=29/54; see Figure 3 in Appendix 5 for the prevalence of all ACEs). The second data source here is taken from the documentation that Lighthouse staff record on the service users’ background including vulnerabilities, disabilities and medical history.26 Again, this data showed that a large proportion of service users have vulnerabilities (n=69 out of 89). These 69 service users presented a total of 163 vulnerabilities between them (at an average of 2.4). 40 service users have at least 1 vulnerability. Among the most frequent categories were anxiety and/or depression (n=34/89), followed by history of domestic violence (DV, n=16/89) and CSE (n=15/89). There were also 13 service users who had a history of self-harm, and 13 who were reported as a suicide risk. A full breakdown of service user vulnerabilities is presented in Table 3 within Appendix 5. Client disabilities, where applicable, are also recorded on the system. Out of the sample of 89 service users, 23 have a known disability, and 9 have more than one. For these 23 service users, there were a total of 36 disabilities between them. Learning disability categories were most commonly recorded such as autism, communication speech and language therapy, and mild, or moderate learning difficulties, and 16 service users were recorded having at least one of those areas. The third source of data from the CMS concerning vulnerability are risk assessments which are completed by a primary care worker at the Lighthouse, and are updated as and when a service users’ risk changes.27 There’s a variety of valuable information captured to further the knowledge on vulnerabilities, including whether a CYP is currently known to children’s social services (n=44/61), or whether they have previously been known to children’s social care or early intervention support services (n=26/61). The assessments also show that 6 of 61 service users are subject to care order or child arrangement orders. A key part of the risk assessment is also for the worker to assess whether there is further risk to the CYP or others. There were risk assessments recorded for only 41 service users.28 As such, it is apparent that many of the service users are still at risk, particularly of further abuse; there is at least ‘some’ risk or concern of this for 27 service users (see Table 4 in Appendix 5). Two CYP were assessed at being a high risk of suicide. Only two were assessed for concern of being a risk to others. Service allocation and addressing vulnerability At the IA, the team will discuss options and seek to support. Therefore, based on the needs and wants of the CYP and their families, they may be allocated to different aspects of the service. The case management system records the types of Lighthouse worker that the clients are allocated to. However, due to the methods that cases are allocated to staff on the case 25 Glowa et al. (2016) 26 In most cases, the information at IA is recorded by the Doctor or Health and Wellbeing practitioner. Data is also provided on medical histories, although there is a lot of overlap between the categories recorded there, and the categories recorded on vulnerability areas. For this reason, only one dataset has been reported here. 27 It is unknown how regularly the risk assessments are revised at present. 28 As mentioned previously, this appears to be due to data quality issues in the early stages as the service adjusted to a new case management system and ways of working. The numbers of risk assessments completed in the first 3 months of 2019 was over 3 times the number completed in the last 3 months of 2018. 13 management system (CMS; see Appendix 6 for overview of these methods), there are some implications for what conclusions can be drawn from this dataset for the evaluation at this stage. Table 1 below presents the output from the CMS in terms of how many CYP (again, only those who have provided consent for the evaluation) have been allocated to a Lighthouse worker/service, the caveat is that it could be for several reasons which are currently unknown given the data entry. The top appears to be Paediatrician (n=77), followed by advocacy (n=59) and CAMHS (n=42). Given the data quality at this stage, the evaluation is not able to assess any links between the needs of the service users and the Lighthouse services received. These issues will look to be resolved for the next stage of the evaluation, and E&I will work with the Lighthouse to find an alternative way of accurately measuring the volume of services received. Table 1. Number of CYPs allocated to services on the case management system Service Number of CYP allocated Paediatrician 77 Advocacy 59 CAMHS (Children and Adolescent Mental Health Services) 52 Play specialist 42 Letting the Future In (LTFI) 29 35 Sexual health nurse 11 Protect & Respect (P&R) 30 9 Social Care Liaison Officer 7 Police Liaison Officer 6 Duty case worker 5 Total 303 Offence details The Lighthouse’s Police Liaison Officers record data on Excelicare around criminal justice details, including characteristics of the offence, and key dates in the criminal justice processes (i.e., police actions and decisions, CPS actions). At this stage, it is too early to examine criminal justice processes and timeliness through the system, however some analysis has been undertaken to understand the nature and types of offences that the service users were victims of. This analysis focusses upon the 92 cases where service users provided consent for the evaluation. The most frequent offence was intra-familial sexual abuse (n=36/92), followed by peer-on-peer sexual abuse (n=26/92). 31 Around half (n=49/92) of offences occurred within the five Lighthouse boroughs, 7 were within another London borough, and 8 occurred outside London (including 2 other countries).32 29 LTFI is an NSPCC service within the Lighthouse which supports children aged 4-17 to recover from the impact of the sexual abuse: https://learning.nspcc.org.uk/services-children-families/letting-the-future-in/ 30 P&R is an NSPCC service within the Lighthouse offered to CYP aged 11 to 19 who are either experiencing exploitation, or require support to learn about healthy relationships: https://learning.nspcc.org.uk/services-children-families/protect-and- respect/ 31 No distinction was made between rape cases or other child sexual abuse cases. 32 The remaining 28 are unknown. 14 In 73 cases the CYP made a disclosure of the offence, most commonly to someone in their family (n=31/73). Other individuals disclosed to were someone in their school (n=10/73), 7 were police, 3 were friends/peers, 3 were social services, 3 were health services, and 16 were either recorded as ‘other’ or unknown. For a further 10 cases the offence came to light due to suspicion from either a family member (n=6), health service (n=3), or social services (n=1). For the remaining nine cases it is not recorded how the offence came to notice. 26 out of the 92 CYP were recorded as being repeat victims, and for 9 of the cases it was noted that there were other people within the family who are at risk from abuse. Summary of performance review Over the first 9 months, the Lighthouse has received 276 referrals and delivered 165 IAs (which is a 60% conversion rate). Although both figures are below the predictions at the onset of the service, the number of referrals represents over twice the number of referrals (118) per year from the CSA hub in 2017. The difference could be for a variety of reasons such as the validity of the predictions, linked to an aspect of early implementation, or referral awareness. Most referrals came from Camden, Enfield and Islington, and from Children’s Social Care. The theme of vulnerability was clearly seen across a range of data sources (psychometric, staff capture and risk factors) and they have experienced a disproportionate amount of adversity in their lives so far. This clearly demonstrates the need for a service that seeks to integrate multiple strands of support. Effective partnership working, regular training and adequate support for the Lighthouse staff are key in ensuring this. These issues will be expanded upon within Section 3. The nature of the offences that brought the CYP to the Lighthouse show that most of the abuse was intra-familial. This reinforces the lack of safe and stable household that some service users are in. Several wider issues also emerged – such as the case management system itself and needing to learn more about how data is entered onto the system, so details such as an accurate understanding of those accessing services can be gained. This is not unique to the Lighthouse as CMSs are not designed as research tools. The issue of consent is also worthy of note – positively, this improved over the nine months with support from the researchers and lighthouse staff working together. Nonetheless, it should be continuously monitored. 15 3. Learning from early implementation This section explores learning from a range of sources, notably focus groups carried out with Lighthouse staff, a survey with professionals, and feedback from service users at the Lighthouse (parents and CYPs), with the aim of presenting learning related to the set-up and continued implementation of the Lighthouse. As the evaluation progresses, the aim will be to track these issues over the course of the Lighthouse. In the current fieldwork - six themes emerged; these were: • General perceptions of the Lighthouse; • Consistency to the developed model; • Ways of working; • Referrals and demand; • Partnership working; and • Challenges ahead. General perceptions of the Lighthouse Members of staff have been collecting feedback from parents and CYPs about their experiences at the Lighthouse. Between February and June data was collected from 66 people - 20 parents, 27 young people aged between 13 and 18, 11 young people aged between 8 and 12, and 8 children aged under 8.33 The questions asked of the service users, and their format, varied depending on the age of the respondent, so it has largely only been possible to combine the responses from the parent and 13-18 groups (47 respondents) but overall, respondents reported feeling very positive about the service they had received at the Lighthouse, the facilities that were available there, and the convenience of the location/appointment.34 In terms of the support provided by the Lighthouse, respondents felt that they had been listened to at the Lighthouse (46/47 reported this statement to be ‘certainly true’), that it was easy to talk to the Lighthouse staff (43 ‘certainly true’, 4 ‘partly true’), and that their views and worries had been taken seriously (46/47 ‘certainly true’). They also felt that the staff at the Lighthouse knew how to help with the problem (38 ‘certainly true’, 8 ‘partly true’), that they had been given enough explanation about the help available (44 ‘certainly true’ 3 ‘partly true’), and that the staff at the Lighthouse were working together to help with the problem (43 ‘certainly true’, 3 ‘partly true’, 1 ‘untrue’). Service users aged 8-12 were asked to indicate which of a number of elements of the Lighthouse’s service they liked (choosing from ‘Drinks/snacks’, ‘Game/toys/activities’, ‘People’, ‘Rooms’, ‘Reception’, or ‘Everything’). Of the 11 CYP to respond, ‘reception’ was mentioned by 6 respondents (5 of which were under the ‘everything’ banner), ‘drinks/snacks’ were selected by 7 respondents, ‘games/toys/activities’ and ‘rooms’ by 10 people, and ‘people’ by all 11 respondents. As with the first report, the feedback from the professionals’ survey showed a positive response to the opening of the Lighthouse. 35 Many described it as being ‘good’ or ‘fantastic’ whilst 33 All four groups of service users were asked how many appointments/visits they had had at the Lighthouse. Of the 61 who replied, 29 had attended once, 20 had attended 2-4 times, and 12 had attended 5 or more times 34 The respondents overwhelmingly felt that the facilities at the Lighthouse were comfortable (46/47 ‘certainly true’), that their appointments had been at a convenient time (36 ‘certainly true’, 6 ‘partly true’), and that it was ‘quite easy’ to get to the premises (40 ‘certainly true’, 6 ‘partly true’). Overall, when asked if the service received at the Lighthouse was good 45/47 felt the statement was ‘certainly true’ 35 Feedback was received from 19 respondents who had said they were aware of the service. 16 commenting that the service is very much needed within the community. Similarly, when staff members were asked whether they felt that the Lighthouse was meeting its objectives they either felt it was or was working towards meeting them. Lighthouse staff reported that awareness of the service was good, but that there was still scope for improvement. Not only were there differing levels of awareness among boroughs, but across many organisations as well. The findings from the professionals’ survey showed that more than half of respondents were ‘aware’ of the Lighthouse service (n=25/47). The most common reason for being made aware of the service (n=11) was through safeguarding training or a Lighthouse staff member attending a meeting or forum of teachers and social workers. Of those aware of the service, about a third said it was well known within their respective organisations (n=8/25). Eight also said that they had made a referral to the Lighthouse, either via email or phone or via social worker; most of them found the process easy. Consistency to the developed model Certain elements of the Lighthouse model that it was anticipated would form part of the service have been delayed and are yet to be implemented. As outlined in the first evaluation report, Section 28 (s.28) of the Youth Justice and Criminal Evidence Act 1999 sets out a range of special measures which should be available to help vulnerable and intimidated witnesses give their best evidence at a criminal trial, including pre-recording cross-examination, and it was anticipated that the pre-recorded cross-examination would take place at the Child House rather than a child having to give evidence at court. While this remains the long-term ambition, the delayed implementation of s.28 nationally due to technological issues has prevented this happening in the Lighthouse. Similarly, the use of the Lighthouse as a Live Link location, allowing children to give evidence remotely without attending court has still not yet been implemented, although discussions are on-going with the judiciary/HMCTS. In addition, it was intended that the service would be providing psychologist-led (rather than police-led) Achieving Best Evidence (ABE) interviews. In the event, the training for the psychologists has taken longer than anticipated and psychologist-led interviews have only recently begun (psychologists were trained in December 2018, and from January 2019 had a period of six months of supervised interviewing as part of the quality assurance process). In addition, over the first nine months several changes have been made to operational practice at the Lighthouse. For example, the initial intention had been to open the Lighthouse on Saturdays. However, this proved impossible between October and December 2018 because security staff were not available at weekends. Once security had been hired, the Lighthouse opened on Saturdays between January and March 2019. However, in April 2019 the decision was taken not to open the Lighthouse on Saturdays because of low levels of usage and feedback “I think it is extraordinary, I think if you think about where we were a year ago which was a piece of paper with our bid, and then we got it and in a very short space of time 6 months the building was transformed...So the point about it is that you are going from, in a year where you have gone from nothing to an all running service which is absolutely extraordinary”. 17 received from service users about their preferred hours of operation36. However, the weekend is still available for ad hoc activities (i.e., running a group if requested by parents), and members of staff will still attend case conferences off-site at the weekend. Similarly, there have been changes made to the initial assessment process in terms of staff attendance (which are discussed in greater detail in the ‘Referrals and Demand’ section later in this report). In addition to these, a weekly whole team case review of all open cases began in June 2019 (following the need for this being identified in the staff survey in March 2019). In May 2019 the Lighthouse ran a parents’ groupwork course. The course was based on the ‘circle of hope’ model, with the contents agreed in advance with the CPS (contract with families to not talk about their own abuse). Ten parents attended, and there was 100% attendance for the whole course. It was asserted that the offer to parents overall had been more substantial than originally anticipated due to the complex needs they presented. The majority of IAs resulted in parent work alongside the intervention to the child, which could include one-on-one work for one or both parents, as well as access to the 10-week parent psycho-education group, work which was not reflected in Excelicare. Some staff members also mentioned that the roles they were currently undertaking differed from what they had anticipated at the outset of the project. This appeared to be the case particularly for the advocates, Social Care Liaison Officers (SCLOs) and Police Liaison Officers (PLOs). Staff suggested that they had anticipated that each child would be allocated an advocate who would stay with the child throughout their time at the Lighthouse, but that this did not always happen pre-IA, with implications particularly for the SCLO role to take a lead in coordinating the referral process and liaison with local social care. The advocates themselves had expected that there would be more information sharing in the Excelicare system, and user access was revised six months into the pilot. A similar situation arose for the PLOs who were not doing the role they anticipated in terms of data access/provision, but whose role had expanded to encompass several other aspects, most prominently the training of psychologists for psychologist-led ABE interviews. Whilst this training was originally identified as part of the PLO role, upon initiation of the service it became a more significant part of their role, and as such it was anecdotally reported to have taken approximately 120 hours of the PLO’s time.37 Specific reasons for the unexpected growth in the PLO role are unknown however the service is brand new, and not everything can be predicted; this is important learning for the service model. The need for better feedback and communication from the Senior Leadership Team (for example via regular team meetings) was one of the six improvement workstreams that emerged in late 2018 from the first team away day38, and led to a team briefing starting in January 2019. A staff survey in March 2019 identified further learning about ways of working together, culture and values. Positively, as a result, there had been a number of attempts on the part of management to try and improve relationships within the Lighthouse.39 Nevertheless, the issue still arose at the point that E&I’s interviews and focus-groups took place suggesting this work needed to continue. 36When asked in the feedback when they would prefer their appointments at the Lighthouse to be, of the 50 respondents who replied (the question was asked of parents, 13-18-year olds and 8-12 year olds), 25 indicated it would be ‘during the day’, 20 ‘after school’, 4 on Saturday, and one at another time. 37 Training activities carried out by the PLO included formal training, observations, role-playing training, and supervised ABEs. 38 The other workstreams were the appointment system, building issues, lack of supervision, Initial assessment and Excelicare). The communication workstream had identified the following as areas for improvement; lack of regular meetings for; speciality team updates and discussion – including feedback to and from the Operational Meeting for team leads, reflective space in addition to the Weekly Case review, allocation meetings not working. 39 These included breakfast with the Service Manager between October and December 2018, boxing in the park after work, yoga, swimming before work, team breakfasts, eating lunch together, a team afternoon (shared lunch, training and activity afterwards) and 2 team building days (in December 2018 and July 2019). 18 There were mixed messages across staff about the extent to which there were clear written processes/policies at the Lighthouse, particularly in the very early stages. Some staff felt there was a lack of clarity around procedures, and described the uncertainty that resulted, however, this was disputed by others who pointed to the written Lighthouse guidelines that were in existence. In all likelihood, this is more likely to be a factor of how effectively Lighthouse guidelines are communicated and then embedded into working practices. Initial assessment The IA process was also one of six workstreams for improvement identified in late 2018, and it was clear that this was an area of work where practice had evolved during the 9 months since the Lighthouse had gone live.40 It was suggested by Lighthouse staff that a number of changes had been made to the IA process, including better preparation for the CYP about what to expect, changes to where the IA was held initially, and the choice about which staff attended the IA being given to the CYP/ parents/ carers.41 At its inception, the model introduced had assumed that representatives of all the services present in the Lighthouse would attend IA meetings. Staff highlighted that the processes for the IA had evolved over time. While the core offer from the Lighthouse was for children and young people to have a ‘menu’ of services to choose from, the logic of having staff present from each service was deliberated, in terms of the impact this had on resources and staff availability, and in terms of the child’s experience, whether it was ‘too many people’ to meet at once, although it was acknowledged that this was difficult to predict case by case. Ways of working The Lighthouse has now been operational for nine months (six months at the point the focus groups and interviews were undertaken with the staff) giving practitioners the chance to reflect on factors that impact on the day to day working of the initiative. It should be stressed that the comments that follow appeared in a context where staff were keen to emphasise the strengths and achievements of the Lighthouse. Indeed, as one member of staff put it, ‘for many of the 40 The six workstream areas identified were: the appointment system, building issues, lack of supervision, Initial assessment, team meetings and Excelicare. Issues identified by staff in relation to the IA had been; lack of clarity about who was responsible for what task, whether assessments should take place at visits or referral, gaining agreement for service, who should take a lead in the assessment, advice or assessment tasks, who was responsible for documenting what on Excelicare, who was accountable overall, and held risk in mind, and who should meet the child on arrival. 41 The changes in full were; better preparation in advance for the CYP, the provision of snacks/drinks and free fruit in the IA, breaks in the IA, choice being given to the CYP and family about who was in the room for the IA, starting the sexual health element of the appointment, starting the IA in the therapy room rather than the medical room, better documentation on Excelicare, the introduction of a post-IA review meeting, and debrief in the allocation meeting. “I think at the beginning there was an attempt to have like a ‘perfect’ offer to families and as time has gone on, and you’re aware that you have got to be really careful about resources and that you can’t have 5 people deciding on what the offer is going to be right at the beginning, you might need 2 people to meet the family first, and then other people come in as they’re needed and adapting things, for example, in terms of the initial meetings with the families, then things have needed to change” 19 team, despite the challenges, it is an absolute privilege to be working at the Lighthouse’ describing her peers as ‘outstanding professionals in their field’ with a unique level of expertise ‘dedicated to getting the offer to children right’. Nevertheless, staff raised some issues relating to the equipment and infrastructure of the Lighthouse. These are not new as they had been identified in the first evaluation report, and ‘building issues’ had been one of the work streams for improvement identified in late 2018 suggesting more needs to be done on the topic.42 Inadequate sound-proofing in parts of the building was the concern identified most frequently by staff, particularly in the ABE suite, the mezzanine/staff kitchen area on the second floor, and the medical examination and therapy rooms on the 4th floor. Subsequent to the focus-groups/interviews remedial sound-proofing has been provided for the ABE suite, involving the insulation of and carpeting for the rooms above the suite. However, at the time of the interviews/focus groups staff described how they worked around these soundproofing problems, such as walking around the office without shoes during ABE interviews (which are held on the floor below the main office area) or talking quietly in the kitchen area (staff were aware that service users coming into reception could hear talking and laughing from the staff-kitchen). However, other staff members stressed how important it was for them, in what was a stressful environment, to have access to an area where they could relax, and there was concern that the absence of such an area might have an effect on staff’s wellbeing and ability to carry out their roles as effectively as they would like. Staff also mentioned problems relating to telephony (i.e., poor mobile phone signal in the building, and a lack of land-lines) and there was some criticism of Excelicare, the Lighthouse’s case management system, with people describing the system as ‘cumbersome’, ‘time-consuming and clunky’ and prone to crashing. Respondents reported difficulties finding data on the system, frustration that the same data had to be input more than once, and generally felt that it compared unfavourably to other systems they had previously used.43 It was hoped that some of the existing flaws in the system would be addressed by future planned changes to the system in subsequent versions of the CMS (for example, the absence of a suitable appointment system in Excelicare, which was felt to be a particular limitation), and it was suggested that staff were getting used to 42 The ‘building issues’ specifically mentioned were lack of shelving in cupboards, need for ‘engaged’ door signs to stop interruptions of therapy sessions, open kitchen area causing issues with confidentiality, lack of soundproofing in ABE suite, CCTV cameras to be removed and feed viewed in reception, need for lockable drugs and DVD cupboards, need for blinds in ABE suite). 43Excelicare was one of the six improvement workstreams identified in December 2018. Factors mentioned in the workstream document were; the system crashing, staff being unable to share case notes, and difficulties in recording telephone contacts/initial assessments that did not become assessments, visits, pre-assessment work, medical follow ups on the system. 'the biggest issue with the estates is the soundproofing. So, it’s shocking. I mean the ABE suite, obviously, is really bad. But we’re finding out more and more that if someone is in one of the therapeutic rooms and you are outside waiting you can hear everything that is being said… If you’ve got a young person that is having a confidential conversation and their mum has been asked to wait outside the room, but mum can hear everything that they are saying, that’s not confidential. So, there’s just no soundproofing.’ 20 the system.44 It was also suggested that some of the problems with Excelicare had been caused by the tight timescale for implementation, and a perceived lack of involvement of Lighthouse staff in its design, although practitioners had been involved in the process.45 As expected, to a degree, problems with the system, and their resolution, had only become apparent following go- live and the operational use of the system. Staff were, however, keen to praise the role played by the part-time data officer (recruited in April 2019) in assisting with issues around the use of Excelicare, providing training, and encouraging completion of consent forms and data compliance generally. Respondents also identified additional facilities they felt were needed at the Lighthouse, or where existing facilities should be reconfigured (i.e., the knocking-through of one of the rooms adjacent to the medical examination room to enlarge the latter, and the need for an additional examination room), although again, staff stressed that they did their best to work around existing arrangements. Issues with aspects of the medical infrastructure were also raised (colposcope, curtains, bins). Again, notwithstanding the fact that before and during the mobilisation stage of the project professionals had been involved in the design of the premises, some respondents still perceived there to have been a lack of Lighthouse staff involvement 46 Another frustration expressed was the length of time taken for issues within the building to be remedied, and it was suggested that these combined had a detrimental effect on staff morale. The staff spoke frankly on these issues and expressed a keenness for the various challenges around building design and electronic systems to be captured and recorded as crucial learning, particularly if the service were to be replicated elsewhere. These are key aspects of the service that could be overlooked while designing and developing a service such as this but are integral to the staff being able to carry out their jobs effectively. The staff’s flexibility and willingness in finding ways to work around these building and soundproofing issues (such as walking without shoes on, or not booking two adjacent therapeutic rooms at the same time) should also be noted. The importance of reception at the Lighthouse as the first point of contact for service users and parents/carers was stressed by a number of respondents, and the need to recruit administrative staff who had experience of working in a safeguarding context, above and beyond clerical skills (although again, the detrimental impact on the service user’s impression of the reception area by the lack of sound-proofing was stressed)47. 44The (lack of) an appointment system was another of the improvement workstreams identified, which was described in the workstream document as ‘leading to issues with visibility of staff availability, appointment booking and reporting’. While a workaround had been introduced (an Excel based ‘appointment scheduler’ for all staff to complete and the sharing of Outlook diaries) this was described as ‘frustrating for staff, difficult to use and therefore is not being used consistently’. 45 Involvement with practitioners was carried out in a condensed period and there was less time for operational testing to identify problems with the system, and their resolution, prior to go-live. These challenges were recognised early on and are the reason why a development budget for the CMS has been provided to be utilised for the second stage of user acceptance testing (UAT2). 46 It should be noted that, because the design process took place before staff had been recruited, the professionals concerned were not necessarily those now working in the Lighthouse. 47 Lack of administrative support was also identified by medical staff, particularly around the sending out of letters and booking appointments. Again, it was suggested that admin staff had been recruited at the wrong level, and with the wrong sorts of skills, it was suggested the need was for people with experience of working in a medical environment. “we have traumatised people coming into the building so the response that you have for people, the way you greet them, the way you answer the phone, the way you think about what might be going on for the person, all that stuff is quite high-end don’t you think?” 21 Supervision, or rather the lack of it, had been another of the areas of concern identified by Lighthouse staff at their away day in December 2018 and had subsequently been adopted as a workstream for improvement.48 Some respondents felt they were well supported in their current roles both internally by Lighthouse peers and managers, and externally to the Lighthouse (i.e., respondents mentioned support from their parent/host organisations, and the part played by support from outside their direct line management). There was also positive feedback on the quality of the whole team training which is provided monthly. Things are going in the right direction, however there were some staff who felt the level of support could be improved upon. Factors mentioned were the absence of the Senior Leadership team, the absence of reflective space, and the need to provide support beyond operational meetings. Referrals and demand Referrals As has been seen, in the period under study there were 276 referrals to the Lighthouse, at an average of 35 per month. Practitioners broadly felt that levels of referrals coming in had been as they expected. Again, reflecting the performance data, staff reported that most referrals had been received from social services, with some from education and GPs. Staff were aware that there had been few self-referrals, and were keen to address this, to “reach the true unmet need”. Some staff also asserted that the type of referrals being made to Lighthouse were different to those in other services they had worked in before (CSA hubs for example). The Lighthouse cases were described as more ‘raw’ and ‘complex’, in that the children and young people had deeper mental health needs with incidents of serious self-harm and suicidality; this reinforces the findings outlined within the performance data. In addition, it was asserted that the service increasingly recognised the need to collect data about the referral prior to the Initial Assessment taking place, something that placed more emphasis on the role of the social care liaison officers (SCLOs). As the performance data indicated, respondents identified that there had been differences in the numbers of referrals from the five pilot boroughs. Camden and Islington were consistently reported as making the most referrals (which was the case when the interviews and focus groups were completed), with Barnet and Haringey reported to be making very few referrals. As suggested earlier, staff believed there was variation in the awareness of the service across boroughs and organisations, and it was further suggested by staff that distance might be the explanation for the lower levels of referrals from these two boroughs. Capacity In terms of staffing and recruitment, during the period covered by this report there was only one unfilled post at the Lighthouse (the senior advocate role - although part-time cover for the role was in place). However, one of the two PLOs left the service at the beginning of May 2019. The Lighthouse were able to obtain 2 weeks cover for the post in June, and part-time cover (one week on and one week off) from mid-July onwards. Notwithstanding the reduction in PLO staff, the remaining officer was still required to provide up to 1.5 days a week cover for the CYP Havens as previously agreed. In addition, the departure of the PLO came at a time where the other PLO 48The factors mentioned in the workstream document were staff concern about lack of supervision and lack of spaces to feel they can contain case management and risks. Lack of clinical supervision in LTFI or Advocacy teams and staff felt that was a gap and a risk. Concerns also raised about proposed frequency of 35 days, whereas 2 weekly felt more suitable. 22 had already organised four-weeks annual leave. In addition, during the period one of the SCLOs was injured in an accident at work and off sick for 3-4 weeks. There has been a vacancy in senior administration since May. It was suggested that certain elements of the service (CAMHS in particular) was already working at capacity, although there was some discussion about what being at full ‘capacity’ meant, as different roles were contracted for different caseload numbers. Notwithstanding the fact that levels of staffing were largely as anticipated, and as outlined by performance data the numbers of referrals and IAs were below expected; staff expressed concerns about the ability of the Lighthouse to sustain the current level of service in the future. There is a strong possibility that staff already feeling worked to maximum capacity is related to the complexity of the cases that they are seeing which need more attention and time than previously anticipated. Staff were aware that most of the referrals to the Lighthouse had come from within the system, particularly from social services, and were concerned that if the level of awareness of the service increased in the community and a wider audience was reached, then there was potential for the demand to increase (especially on top of an increase in self-referrals). Suggestions were made as to how any increase in demand should be managed through day to day work. For example, better onward signposting or having fewer staff attend the initial assessment. It was acknowledged that a balance needed to be struck between providing a good quality service that retained its core aims and values but ensuring that the pace of work is achievable and managed appropriately. Partnership working The literature around child advocacy centres and Barnahus suggests that one of the benefits of the approach is the improved partnership working it brings (Herbert & Bromfield, 2016, 2017) together with improved awareness and confidence in dealing with CSA on the part of professionals (Landberg & Svedin, 2013). This has also been documented across many other disciplines (i.e., offender management). Feedback form Lighthouse respondents echoed many of these advantages’ such as partnership working and information sharing. Respondents were especially positive to the range of services and partners in terms of benefits to the service user. It was described as “quite wonderful …when it works well the benefits that has on the child and family in the room and the fact we can coordinate everything under one roof”. Staff were positive that the partnership working had resulted in earlier communication between Lighthouse colleagues across different disciplines. It was suggested this was particularly useful when there was a need to gather further information from colleagues, rather than “having to wait on the end of the phone or wait for an email in regard to getting a bigger broader picture of what could be going on” as had been the case in the past. Staff also appreciated the opportunity to benefit from the expertise and different viewpoints of agencies other than their own; enabling “People talk about, you know, opening the floodgates, and you open a new service and people come. I think as I said, to begin with we scooped up what we should have always had. I think that if we promote it, we’ll be blown away by the number of children that come forward, I think that’s a real issue.” 23 them to build a broader and more detailed understanding of how other agencies functioned: “it's good for staff learning as well, I’ve learned stuff about medicine that I wouldn't have known, and the criminal justice system". More specifically, the benefits of having individuals who understood how other agencies worked was particularly stressed. For example, the ability to create easier pathways into services that beforehand would not have been possible was very well appreciated. For example: “trying to get [a] pathway into CAMHS when you’re not a CAMHS person is, well it’s impossible… whereas if you’ve got CAMHS practitioners talking to other CAMHS practitioners then I think there’s a real hope that that can happen”. Members of staff were asked to identify if there were any agencies who were not currently located in the Lighthouse who they felt should be. The potential additions mentioned were education liaison staff, a CPS case worker, for the Lighthouse to undertake the forensic examinations, and to have a presence of local social workers in the Lighthouse, although it was cautioned that there might be issues around space. Similar to the evidenced benefits of partnership working, there can also be challenges in bringing individuals from a range of organisation together – for example, differences in language, terminology or organisational cultures that often result in working tensions. For example, respondents described staff had different terms and conditions and pay levels for often doing ostensibly similar jobs or tensions between the medical model and social work perspective. Staff were keen to discuss the process of determining the ‘Lighthouse way of doing things’, to reconcile the different cultures and bring all their approaches together and find a way of working for the service as a whole, described as a significant, and on-going, learning process. Challenges for the future When asked to identify the main challenges to be faced by the Lighthouse in the future the most popular responses were the uncertainly about future funding, the ability of the service to manage any increase in demand (i.e. if self-referrals increase) especially in relation to mental health support. This does raise an important issue of how the Lighthouse will adapt to any future change. Suggestions of developing some form of triage were made to manage capacity and the requirement to sign-post individuals more effectively to outside services where appropriate, or the use of a ‘hub-and-spoke’ model where certain aspects of the service would be provided centrally, but other elements offered in the local boroughs. “'People are really highly qualified in this organisation there is so much experience and everybody has a clear idea of what it is, it is just bringing it together as a cohesive – ‘This is the Lighthouse and this is the way that we do it’, and it is not CAMHS, it is not paediatrics and it is not social care, it’s unique because people are not going to want to hear how we have modelled a social care system here, or a health system here, they want to hear about something that is different that they can take from” 24 “you need to get much more efficient about not accepting cases that actually are not right for here. You need to be clearer about, if we are saying cases are not right for here, what’s the offer in terms of networking and plugging people into other services, and we have done that and we’ve had great success in plugging people into places like the Portman, maybe NCATS but actually saying, we will take the responsibility now you have made the referral, we’re going to help progress that” 25 4. Discussion The focus of this report has been to highlight findings and issues relating to the process of the first nine months of the Lighthouse’s implementation. Data for this report was gathered from several sources, including the Lighthouse’s case management system, interviews and focus groups with Lighthouse staff, a survey with professionals within the community, and service user questionnaires. The service has had steady levels of referrals within its first few months which have averaged as 35 referrals per month, and the number of IAs carried out has averaged at 21 per month. These numbers suggest the referral rate is lower than was originally predicted – however, there may be several reasons for this, ranging from the original predictions being incorrect, to awareness levels of the service still improving, to its full potential yet to be reached. In their interviews and focus groups, staff discussed how the current levels of demand and referrals ‘felt’ for them on the ground, and they were of the perception that referral levels were as expected, albeit cases were more raw and complex than their previous experience. This suggests the issue here, is more likely than not, to be the limitations in the original predictions. Indeed, staff acknowledged that there had been very few self-referrals to the service so far and expressed a keenness to increase this to reach an ‘unmet need’, although with the knowledge that this would have further implications on increasing demand and capacity. Staff are clearly concerned about issues of sustainability, and the capacity of the service (especially mental health) and whether existing service levels will be able to be maintained in the future. Nevertheless, the demand and capacity will need to be addressed and managed. Amongst the methods suggested by staff to manage this demand more effectively were ensuring appropriate referrals, more effective triage, and sign-posting individuals to alternative services outside the Lighthouse where appropriate. It is a recommendation that the service put plans in place to mitigate problems around capacity and demand, staff themselves highlighted some possibilities to begin this discussion (i.e., more effective triage, better signposting to alternative services) which could themselves lead to changes in how the Lighthouse works – something that will be tracked in future reports. E&I obtained individual-level data for service users who reached an IA stage and who consented for their data to be used within the evaluation; this number was 92 service users. The amount of data fields received from the CMS (over 300 variables) provides details on the backgrounds of the clients, their needs, some medical history, services they are allocated to, and details of the criminal justice cases. At this stage, E&I cannot work with all variables as some of the data is incomplete (either due to data quality issues, or because certain case details haven’t yet been entered onto parts of the CMS). An additional caveat for the data is that the consent rate (calculated using the total number of IAs) is at 56% and there is room for improvement here. The consent issue will be crucial for the evaluation’s ability to robustly evaluate the impact of the service on health and wellbeing and criminal justice outcomes and will be monitored on an ongoing basis. Looking across a variety of data captured (e.g., ACE-Q scores, risk assessments and medical histories) it is clear they are a very vulnerable group, far more so than non-clinical populations. Again, these findings are echoed in the interviews and focus groups with staff who also highlighted the high level of need across the group and how the cases are more ‘raw’ than experienced before. This also reinforces the importance of having a service that integrates multiple strands of support to appropriately meet each need of the CYP. Effective partnership working, regular training and adequate support for the Lighthouse staff are key in this. 26 Staff felt that the Lighthouse was achieving or working toward achieving its objectives. There was a recognition by Lighthouse staff of the benefits of having different agencies working together under the same roof – both in terms of benefits for the service user (access to different services on the one site, quicker access, and reduction of the need to repeat their story) and practitioners (easier access to other agencies, different areas of expertise available to them, and exposure to different cultures/understandings). Equally there was a recognition that having these different agencies working together created tensions (around terms and conditions of employment, working cultures, and operational norms) which the organisation was seeking to reconcile, an ongoing process described as developing ‘the Lighthouse way’. There have been several aspects of anticipated service delivery at the Lighthouse that have not yet been delivered (psychologist-led ABEs, and the introduction of the Live Link facility, for example). In addition, certain operational practices at the Lighthouse have evolved during the first 9 months of operation – the opening hours of the service, the role played by the SCLOs (particularly in relation to service users’ early contact with the Lighthouse), and practice around the initial assessment (largely due to the recognition that there was a balance to be struck between the stated aims of the Lighthouse, logistical issues (the resource implications of all services being present), and the impact on the CYP/parents/carer). As in the first evaluation report, staff raised issues about elements of the Lighthouse’s infrastructure/estate. Their major being the reported lack of sound-proofing throughout the building, particularly in the ABE suite, the medical rooms on the 4th floor, and the mezzanine floor/staff kitchen on the first floor. The absence of soundproofing was felt to be of concern because of the potential risk it posed to patient confidentiality, and because of the impact it has on staff’s operational practice and morale. There were concerns too about the poor mobile phone signal in the building, and the limitations of the case management system (CMS) that had been introduced. Several data quality issues arose from the CMS, through this which generated further work and discussion with the Lighthouse manager and staff to help E&I understand the context around how data is entered onto the system. This further highlighted the limitations of the system for the staff, and its limitations as a research tool. A particular issue is the inability to accurately calculate the number of services accessed by CYP. The evaluators will seek to address this for future reports. Next steps The current document is the second published report within the MOPAC Evidence and Insight two-year evaluation of the Lighthouse to have access to data on those who received a Lighthouse service. However, as mentioned previously, at this point the available data is on a relatively small number of service users (and a relatively small percentage of those who have been referred overall) which has limited the analysis possible. The next evaluation report is planned for the summer of 2020, at which point it is anticipated that details of many more service users will be available for analysis. The next evaluation report will also contain findings from the service users’ perceptions of the Lighthouse. 27 References Bellis, M.A., Hughes, K., Leckenby, N., Hardcastle K.A., Perkins, C., & Lowey, H. (2014) Measuring mortality and the burden of adult disease associated with adverse childhood experiences in England: a national survey. Bethell, C.D., Carle, A., Hudziak, J., Gombojav, N., Powers, K., Wade, R., Braveman, P. (2017) Methods to assess adverse childhood experiences of children and families: toward approaches to promote child well-being in policy and practice. Conroy, L., Hobson, Z., Parker, R., & Read, T. (2018) The Lighthouse: London’s Child House Initial Evaluation Report. MOPAC Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A. M., Edwards, V., Koss, M.P., Marks, J. S. (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. Glowa, P.T., Olson, A.L., Johnson, D.J. (2016) Screening for Adverse Childhood Experiences in a Family Medicine Setting: A Feasibility Study. Herbert, J L and Bromfield L (2016) Evidence for the Efficacy of the Child Advocacy Center Model: A Systematic Review Trauma Violence and Abuse 2016 Vol 17 (3) 341-357. Herbert, J L and Bromfield L (2017) Better Together? A Review of Evidence for Multi-Disciplinary Teams Responding to Physical and Sexual Child Abuse Trauma Violence and Abuse 2017 1-15. Hughes, K., Bellis, M.A., Hardcastle, K.A., Sethi, D; Butchart, A., Mikton, C., Jones, L., Dunne, M.P. (2017) The effect of multiple adverse childhood experiences on health: a systematic review and meta-analysis. Landberg, A & Svedin C G. (2013) A Quality Review of 23 Swedish Barnahus. Save The Children NHS England (2018) London Child Sexual Abuse Services Learning Report. London. Vaswani, N. (2018) Adverse Childhood Experiences in children at high risk of harm to others. A gendered perspective. Centre for Youth & Criminal Justice. Zarse, E.M., Mallory, R.N., Yoder, R., Hulvershorn, L., Chambers, J.E. & Chambers, R.A. (2019) The adverse childhood experiences questionnaire: Two decades of research on childhood trauma as a primary cause of adult mental illness, addiction and medical diseases. 28 Appendices Appendix 1: Focus group interview schedule, April 2019 1. Now that the Lighthouse has been open for a few months, do you think it is currently meeting its aims and objectives? a. (Prompt: Improved partnership working between police and social services; improved therapeutic outcomes for children and their families; improvements in children’s and families’ experiences of the criminal justice process; and improvements in the quality of investigations) 2. What do you think of the referral pathways? a. Is the level of referrals what you expected? b. Are the referral pathways working well? (Any areas for improvement?) c. Do you think there’s enough external awareness/knowledge of the service? if not, with whom, and how could this be improved 3. One of the claimed benefits for the Lighthouse is improved partnership working arising from co- location. We’d like to ask you some questions about partnership working at the Lighthouse a. What are the benefits of co-location? b. Any challenges that have arisen from co-location? c. Anybody not co-located who should be? d. How would you say the way you work here differs from the way you worked in previous roles/organisations? 4. We’d like to ask you some questions about the physical infrastructure, IT systems and telephony at the Lighthouse a. Do the systems/infrastructure enable you to do your job? i. If not, how could they be improved? b. Have you been provided with enough training to use these systems effectively? i. If not, what additional training is needed? 5. How confident do you feel carrying out your role? a. How do you feel about the level of support you’ve received in your role? (Prompt – is support provided formally/informally? Any training/CPD needed?) 6. While you’ve all joined the service at different points, have these first few months/weeks/days been what you expected? a. (Prompt: if not, why not? Any surprises?) b. Any changes to service delivery that you would make? (Prompt, hours of opening). 7. What do you think the main challenges for the Lighthouse will be; a) Over the next 6 months? b) In the longer term? 8. Any other comments? 29 Appendix 2: Professional survey, March 2019 Section 1: About You 1. Please tell us your job role [free text] 2. Which area do you work in? [list boroughs] 3. What type of organisation do you work for? a. Health – A&E b. Health – Sexual health c. Health – Mental health d. Health – Maternity Units e. Health – GP Surgery f. Health – Other g. School – Secondary education h. School – Primary education i. Community group j. Victim support agency k. Youth club l. Police m. Other (Please specify) Section 2: Experience of dealing with cases of CSA/E 1. Have you ever been involved personally with cases of CSA/CSE? (Yes / No / Don’t know) 2. How would you rate your confidence in: (Scale – 1 = Not very confident to 7 = Very confident) a. Overall knowledge in regard to CSA/CSE b. Identifying CSA in a CYP c. Identifying CSE in a CYP d. Knowing who to report CSA/CSE cases e. Knowing how to report CSA/CSE cases f. Knowing how to support a CYP who may have experienced CSA/CSE g. Knowing where to find further support for a CYP who may have experienced CSA/CSE 3. How do you think your organisation addresses CSA and CSE? (Scale – 1 = Not very well to 7 = Very well) If you have answered ‘6 = Well’ or ‘7 = Very well’ can you tell us why? If you have answered ‘1 = Not Very Well or ‘2 = Fairly well’ can you tell us why? 4. What services are you aware of that are currently available in your area to support children and young people experiencing CSA/CSE? 5. To what extent do you agree or disagree: (Scale – 1 = Strong Disagree to 7 = Strongly Agree) h. I have received training about CSA/CSE i. If ‘YES’ - What training have you received and how long ago? i. I would appreciate more training on CSA/CSE j. There are suitable services to support CYP that experience CSA/CSE in our area k. The organisation I work for has a clear policy about cases of CSA/CSE l. CSA/CSE is a problem in our area m. The police should do more around CSA/CSE in our area n. Social services should do more around CSA/CSE in our area 30 o. Health should do more around CSA/CSE in our area p. The pathways to accessing CSA services in our area are clear to use and follow q. The CSA pathways are effective and work well r. I know what the CSA pathways in our area are for referring children and young people to specialist services s. Partnership working around CSA/CSE could be improved 6. What do you think could be improved in the pathway for children who have been sexually abused/exploited? [Tick all that apply] - Training of practitioners working within child protection/safeguarding roles - Providing a child friendly environment - The facility for the child to only tell their account once - A clear referrals pathway for treatment & support - Multi-agency working under one roof - Holistic health assessment offered to all children and young people - Early emotional support immediately after disclosure or abuse - Longer term therapeutic support for the child - Long term advocacy support up to and including the court trial - Providing services so the child does not need to attend the police station - Improved support for the child to give their criminal justice account (from disclosure to cross examination at trial) -Child Psychologists leading the main police interview - Providing services so the child does not need to attend court - Support for the non-offending family/carer - All services in one place for the child -Longer opening hours, including evenings and weekends - Other please state: Section 3: The Lighthouse 1. Are you aware of The Lighthouse, a service providing support for CYP in the boroughs of Barnet, Camden, Enfield, Islington or Haringey who have experienced CSA/E? 2. If yes to q1 - please detail below how you were made aware of the Lighthouse. E.g. media communications, direct email, word of mouth? [free text] 3. If yes to q 1 - What are your thoughts on the opening of the Lighthouse? 4. If yes to q 1 - What do you believe are the main aims of the Lighthouse? 5. If yes to q 1 – how well known would you say the Lighthouse is within your organisation/where you work? [scale 1=not known at all, 7=Very well-known] 6. If yes to q1 - have you made any referrals to the Lighthouse? 7. If yes, how did you make the referral? Phone Email Website 8. If yes to q6, how did you find your interaction with the Lighthouse/how easy was it to make a referral – scale of 1-7? 9. Has your interaction with the Lighthouse increased your knowledge/confidence of dealing with cases of CSA/E? 31 Appendix 3: Case study of progression through the Lighthouse. The case study below has been developed by Lighthouse staff to illustrate the progression of a ‘typical’ service user through the service and is used in presentations to external partners. It is a composite example and the details of the case, and name, are fictitious. Teah is a 11-year-old girl, eldest of three siblings. Her family have been supported previously by Children’s social care when there were concerns regarding DV, emotional abuse and neglect. There has been no involvement for the last 18 months. Social Care Liaison Officer • Reported to school that she had poor relationship with stepfather and previously reported Incidents of verbal chastisement and name calling by parents. School reported a change in presentation since the beginning of the year with Teah presenting as more anxious and withdrawn and isolated from peers. • This week after a ‘stay safe’ workshop in school Teah disclosed to her class teacher that she problems sleeping and waking up at night. • Teah discloses being woken by her step father who ‘did things to her’. She became very upset during the discussion and the teacher became anxious about exploring this with her. Teah not asked about the timescales of last contact with her stepfather ➢ The school make an immediate referral to Children’s social care. • The Lighthouse received a copy of the 87A and SCLO and Paediatrician dialled into strategy meeting. • Further information shared regarding children’s and family circumstances, social care, and health in order to establish forensic/medical issues. • Police information revealed historical concerns regarding Teah’s mother being victim of physical assault by Teah’s father. Also unrelated criminal convictions of current stepfather and historical police merlin relating to DV incident with stepfather’s birth children and partner. • Referral to progress to a joint Section 47 investigation. Allocated social worker and OIC to visit the school before end of school day to speak to Teah. • Lighthouse advice on establishing timelines as if contact with suspect had happened within 7 days the Haven would be contacted to establish if forensics may be obtained which would require attendance at the Haven. Lighthouse ABE facilities discussed. ➢ Initial investigation visit with OIC and ASW to school. • Teah minimally disclosed that her stepfather gets into her bed at night. She shares bedroom with a younger sibling Kiely aged 8. Discloses made to touch his genitals and he had touched her on the ‘chest and between her legs’. She was very anxious about her mother finding out and what would happen to her stepfather. • The last reported incident two days ago. Mother attends school and with Leah is taken to the Haven by the OIC and social worker. • The younger siblings Kiely aged 8 and Paul aged 7 are taken to Grandma’s and spoken to by other officers. Kiely confirmed to SW that stepdad got into Teah’s bed but was reluctant to say anything else. Police Liaison Officer ➢ Haven Medical • Finding show an obvious red scar in the fossa navicularis which is between the hymen and posterior fourchette. Medical follow-up required to review findings further details and Teah and mother advised about the Lighthouse. • Teah and her mother agree to stay overnight with maternal grandmother. • Stepfather is arrested and bailed away from the home. 32 • Lighthouse referral is made with Teah & mother’s consent • Allocation meeting and Teah allocated a primary case holder and IA team ➢ ABE at the Lighthouse • Psychologist Led ABE at the lighthouse is requested and organised by OIC and in partnership with Lighthouse PLO Teah & mother supported on day by Lighthouse Advocate • Child discloses ongoing episodes of digital penetration and being forced to masturbate suspect. • Occurs almost weekly for a period of three months. • 8-year-old sister Kiely present in the room when the abuse happens ➢ Witness Kiely’s ABE. • During social care assessment discloses that on a number of occasions she had woken and found stepfather in Teah’s bed. • Mum reports to SW finding Kiely looking at pornography on stepfather’s phone two months previously. School also report incident of other children complaining of being followed into a toilet cubicle by Kiely. • As it is a child sibling, they are suitable for psychologist led ABE too Clinical Psychologist ➢ Initial health and wellbeing assessment • Teah attended for Initial health and wellbeing assessment including medical ‘health’ review. • Before Teah arrived, the team planned the assessment and who will be involved. The team consisted of the Advocate who was present for the ABE, Play specialist, Consultant Paediatrician, clinical nurse for sexual health and Clinical Psychologist. The IA team agreed to meet with the SW for half an hour first to agree with advocate who has already met family who would be the best people to be part of the assessment. • The IA team then met with mum and Teah initially together, before separating off. Leah with play specialist in prep for the health review, and Clinical Psychologist with mum and doctor initially. (IA lasted full morning). • Mother disclosed to us that she is feeling very distressed and ‘triggered’ by her previously undisclosed her own experience of childhood child sexual abuse and current chronic depression and anxiety which has been further exacerbated by the disclosure. Last episode of Post-natal depression after third child. Never engaged with mental health services apart from some PND support organised through health visitor. • We discussed what mum might benefit from (options) and also gave some psycho ed materials about disclosure, how to support your child, possible emotional impact on parents. We then discussed mum’s own needs as an adult separate from her daughter: rape crisis friends and family service/ Tavistock trauma adult service. • Psychologist offered mum separate follow up and time to think about what she might want/ need. • Mum explained that Teah has been suffering with anxiety and low mood more recently, picking at scabs on her arm. However, she is very engaged with her school-based art therapist – who now knows about the disclosure. • Assessment of Teah: ‘bag of feelings’ – she talked about home, school, hopes and lessons. She described her sleeplessness, social isolation and ongoing worries and ‘tummy aches’ at school and home. • Our screen for anxiety and low mood (Rcads parent and child measures) indicated clinical levels of anxiety and low mood, but not significant mental health risk to Teah. • The IA team came together at the end of the morning with mum and Leah and SW to clarify our understanding, agree what to write to family (and copy to GP and network with their consent), and check agreed next steps. We all agreed (including Teah) that the therapy at school work feels important and continues and offered consultation to the therapist at primary school - and 33 school pastoral lead meeting if that would be helpful in thinking about the impact of trauma on learning. • Mum said she felt that she would like some more ongoing work with her to understand and support Leah. We explained that LTFI parent work offer and that this might be a bridge to future parent course access. Mum attended the Light House psychoeducational course for parents ‘circle of hope’ run by advocacy and CAMHS jointly. • We continued to discuss and revisit the pre-trial therapy guidance and consent issues and we also discussed these with the school clinician. We continue to link regularly with the school clinician. • Ongoing work with siblings including Psychology support for ASW in engaging in direct work with siblings • Advocacy agreed to continue to engage with school and social worker and OIC. • Adult based trauma service was able to start work with mum six months after this referral and this is ongoing 34 Appendix 4: Number of referrals by borough, Oct 2018 - June2019 Figure 2. Number of referrals by borough Table 2. Referrals per borough, per month Oct 2018 Nov 2018 Dec 2018 Jan 2019 Feb 2019 Mar 2019 Apr 2019 May 2019 Jun 2019 Camden 7 9 9 9 3 6 7 4 5 Enfield 2 4 10 6 6 6 7 9 8 Islington 3 10 7 5 7 7 5 8 6 Haringey 0 10 10 6 2 8 6 5 3 Barnet 3 7 4 3 5 3 5 9 1 Other London Borough 0 1 2 0 0 2 5 0 1 59 58 58 50 40 11 0 10 20 30 40 50 60 70 Camden Enfield Islington Haringey Barnet Other London Borough 35 Appendix 5: Data from the case management system on service user vulnerabilities Figure 3. Breakdown of ACE scores for 54 Lighthouse service users Table 3. Number of service users with each vulnerability present Vulnerabilities documented No. of CYP (n=89) Anxiety/Depression 34 Other 25 History of DV 16 CSE 15 History of self-harm 13 Suicide risk 13 School/education problems 13 Concerns over safety 8 Eating Disorder 6 Risk of further harm 5 Sexualised behaviour 5 Local Authority care order 4 Drugs/Alcohol 3 Missing from home 3 Repeat attender 0 Table 4. Risk assessment results for 41 service users Risk Ongoing/potential immediate risk Some risks/concerns No identified risk Further abuse 5 22 14 Self-harm 7 9 25 Suicide 2 12 27 Risk to others 0 2 37 42 29 21 18 16 15 15 10 9 2 0 5 10 15 20 25 30 35 40 45Sexual abuseParents divorcedDV against mother or fatherEmotional abuseParent mental illnessEmotional neglectParents alcohol/drugs issuesPhysical abusePhysical neglectParent in prison Number of service users 36 Appendix 6: Anecdotal information from staff on reasons for creating service allocations on the case management system Anecdotal evidence was gathered from the Lighthouse staff around data entry practices with the case management system (CMS), particularly with allocating service users to various Lighthouse services. On the CMS, a Lighthouse worker could be ‘allocated’ to a case for several reasons outlined below. 1. As stated already, during one of the daily allocation meetings, several workers may be allocated to cases on the CMS as the ‘IA (initial assessment) team’. This would occur prior to the CYP being officially worked with by the service, however it is not possible to differentiate between when a worker has been allocated a case in an IA capacity, or if they have formally received the CYP as a client to their individual service. 2. An additional reason for a system allocation is when the child and/or family visit for a ‘show around’ of the Lighthouse by the staff, which would occur prior to the CYP officially being seen by the service, and the staff involved at this early stage may still receive the ‘case’ on the system. 3. Additionally, if a parent receives a service from the Lighthouse (e.g. LTFI), then the worker will be allocated to the case of parent’s child on the system, and it’s not possible to differentiate between a child and parent allocation. 4. Finally, if a child is receiving one of the Lighthouse services and their worker within that service changes, that could also add an additional allocation to that service. In light of the above, the service are: 1) Closing cases as soon as an IA is over if the service is then not needed long term 3) Reviewing this in phase two of the Excelicare development during 2020 4) The service have since changed practice and updated the allocations to address this Alcohol Abstinence Monitoring Requirement A process review of the proof of concept pilot Melissa Pepper and Dr. Paul Dawson MOPAC Evidence and Insight Unit February 2016 2 Continuous Alcohol Monitoring tag – image courtesy of Alcohol Monitoring Systems Inc. 3 Alcohol Abstinence Monitoring Requirement A process review of the proof of concept pilot – summary As part of his 2012 manifesto pledge to introduce ‘compulsory sobriety for drunken offenders’, the Mayor of London successfully lobbied for legislation to allow for the introduction of the Alcohol Abstinence Monitoring Requirement (AAMR). The new sentencing power, introduced as part of the Legal Aid, Sentencing and Punishment of Offenders (LASPO) Act 2012 allows courts to impose a requirement that an offender abstain from alcohol for a fixed time period of up to 120 days and be regularly tested, via a transdermal alcohol monitoring device in the form of a ‘tag’ fitted around the ankle, as part of a Community or Suspended Sentence Order. From July 2014, the Mayor’s Office for Policing And Crime (MOPAC) conducted a 12 month proof of concept pilot in four boroughs (Croydon, Lambeth, Southwark and Sutton) which comprise the South London Local Justice Area. The aims of the pilot were:  To test how widely courts use the AAMR, and the technical processes within the criminal justice system.  To evidence compliance rates with the AAMR.  To evidence the effectiveness of ‘transdermal tags’ in monitoring alcohol abstinence. Utilising a range of methods including stakeholder and offender surveys, interviews with stakeholders and MOPAC officers, and analysis of performance monitoring data, this process review sets out learning from the 12 month (31 July 2014 – 30 July 2015) AAMR proof of concept pilot and helps to build the evidence base to inform discussions around further roll out of the AAMR across London and beyond. Basics around the AAMR and those sentenced to wear the alcohol tag Over the 12 month pilot period, 113 AAMRs were imposed with an average length of 75 days. AAMRs were given for a range of crime types most commonly in relation to violence or drink driving related offences. Almost three quarters (73%, n=82) of AAMRs resulted from Croydon Magistrates’ or Crown Court. There were over 6,500 monitored days in the pilot period during which over 298,000 alcohol readings were taken (at an average of over 2,600 readings per AAMR or approximately 45 per monitored day). In theory, the technology should take around 48 readings per individual per day (depending on time of tag fitting and removal) thus indicating that the technology underpinning the AAMR is working as intended. Evidence & Insight February 2016 4 The AAMR had a compliance rate of 92% over the pilot period, based on the number of cases returned to court and convicted of breaching their AAMR (n=9/113). Of these nine cases, five had their AAMR revoked and therefore failed to complete. The remaining four subsequently went on to complete their AAMR following their return to court. Current compliance with the AAMR appears higher than for some other orders, however direct ‘like- for-like’ comparisons should be treated with caution due to varying offence types, offender characteristics, processes of dealing with breach, and lengths of orders. As expected, the AAMR cohort did not present an extensive criminal background with an average of eight guilty sanctions, six guilty court occasions, and an average Offender Group Reconviction Scale version 3 (OGRS3 2 year) score of 35% (placing them at a low risk of reconviction). In terms of offending histories, the AAMR offenders broadly align more to the general offending population in the UK, particularly those who receive community sentences. Understanding and implementing the AAMR The AAMR was designed and implemented well from the outset, something that had a positive knock on effect throughout the course of the programme. Whether it be the strong governance structure, clear documented tools and information, training, effective partnership involvement throughout design and implementation, or the dedicated MOPAC team (including a project manager with ‘in the field’ experience) – the positive AAMR implementation cannot be over stated. All AAMR practitioners and offenders held a firm understanding of the AAMR's aims and ways of working. However, some NPS/CRC interviewees felt that more information on the AAMR could have been provided to a range of groups to the benefit of the programme. To illustrate, to defence solicitors (as it was perceived they were often unaware of the requirement), the public (to improve knowledge or as a preventative measure) or to allow offenders an opportunity to see the alcohol tag and monitoring equipment in court, in addition to the written information they receive. Using the AAMR The AAMR was largely welcomed by respondents as ‘another tool in the box’ of community sentences, offering an innovative and tailored response to alcohol related offending, and filling a gap in sentencing for alcohol related offences committed by non-dependent offenders. There were some reservations around narrow pilot eligibility criteria for offenders to receive the AAMR, and the requirement for total abstinence in certain cases. However, interviewees felt that a period of abstinence on the AAMR had the potential to give offenders a ‘pause’ in drinking, time to reflect on alcohol consumption and its impact on offending behaviour, relationships and work, and an opportunity to break the cycle of routine drinking. Some NPS and CRC interviewees gave examples of how they had tailored products around the AAMR to support offenders further and use the opportunity as a 5 ‘teachable moment’, including letters sent at the end of the AAMR and literature/advice around alcohol consumption going forward. Whilst the AAMR was considered to be a punitive response for the purposes of the pilot, these possible rehabilitative elements were highlighted by some NPS/CRC and judiciary interviewees as a welcome unintended consequence. Offenders surveyed were largely unhappy about the appearance and ‘wearability’ of the tag, however overall were positive they could complete the order. Concluding thoughts Whilst it is too soon at present to robustly evaluate the impact of the AAMR on offending behaviour or costs, this process evaluation generates learning on the pilot through the views and experiences of stakeholders involved in design and implementation, and the offenders themselves who were sentenced to wear the alcohol tag. Insights from the pilot year indicate the importance of effective design and implementation. However, consideration should be given to the sustainability of this level of project management should the scheme be expanded. Wider roll out of the AAMR would provide a larger sample size and opportunities to explore the impact on offending behaviour, costs and wider possible benefits including health, community and economic outcomes. In light of plans in the 2015 Conservative Party Manifesto to make sobriety orders more widely available, and the extension of the AAMR pilot for a further six months to January 2016, this timely report offers useful insights to inform any expansion of the scheme and can be used as a blueprint for future evaluation efforts in this area. 6 Alcohol Abstinence Monitoring Requirement A process review of the proof of concept pilot Introduction As part of his 2012 manifesto pledge to introduce ‘compulsory sobriety for drunken offenders’, the Mayor of London successfully lobbied for legislation to allow for the introduction of the Alcohol Abstinence Monitoring Requirement (AAMR). Included as part of the Legal Aid, Sentencing and Punishment of Offenders (LASPO) Act 20121, the AAMR is a new sentencing power which allows courts to impose a requirement as part of a Community or Suspended Sentence Order2 that an offender abstain from alcohol for a fixed time period of up to 120 days and be regularly tested, via a transdermal alcohol monitoring device in the form of a ‘tag’ fitted around the ankle which detects consumption of alcohol through sweat (for the purpose of the pilot the tags do not monitor offender location or movement). The technological innovation has a focus on tackling alcohol related offending - and in this way the drive to introduce the AAMR in London was particularly timely. Much has been written about the heavy contribution alcohol makes within violent crime, wider offending and public disorder in the UK, with London disproportionally impacted. The total cost of alcohol-related harm to society is estimated to be £21 billion, with alcohol recognised as a major cause of attendance at Accident and Emergency and hospital admissions (Public Health England, 2014a). The 2013/14 Crime Survey for England and Wales (CSEW) estimated that over half (53%, n=704,000) of the 1.3 million violent incidents against adults in England and Wales included an offender perceived to be under the influence of alcohol. Whilst the volume of incidents has fallen - something that sits comfortably within the overall decrease in crime England and Wales has seen since the mid-1990s - the proportion of violent incidents in which the offender has been perceived to be under the influence of alcohol has remained remarkably stable over the previous ten years3 indicating a longstanding resistant association between alcohol and violence. The CSEW also provides further insights, indicating that alcohol related violent incidents were more likely to occur between strangers, at weekends, during the evening/night, and within a public space, with victims also more likely to receive greater injuries (ONS, 2015). In terms of police data within England and Wales, after a period of decline in violence with injury (a decrease of 27% in 1 Section 76 of the LASPO Act 2012 sets out a number of conditions around the AAMR including that the offender is not dependent on alcohol, that consumption of alcohol is an element of the offence or contributed to the commission of the offence for which the order is to be imposed, and that monitoring by electronic means or by other means of testing are in place. 2 Referred to collectively as a Community Based Order. 3 In the CSEW 2004/05 the proportion of violent incidents where the offender was perceived by the victim to be under the influence of alcohol was also 53%. Evidence & Insight November 2015 7 financial year 2014/15 compared to financial year 2004/05), more recently this type of offending has increased (16% in financial year 2014/15 compared to the same period the previous year) with 40 of the 44 forces within England and Wales recording a rise in violence with injury4. Focussing upon London, violence with injury has risen by 19% (financial year 2014/15 compared to the same period the previous year). Furthermore, internal Mayor’s Office for Policing And Crime (MOPAC) analysis indicates the majority (76%) of the increase within London can be attributed to non-domestic abuse violence with injury, with Friday and Saturday evenings/nights being peak times in key geographic areas - something that clearly suggests an association with the night-time economy (rolling 12 months to January 2015). Indeed, London experiences disproportionate levels of alcohol related crime, with the highest rate per 1,000 population (9) compared to other English regions (ranging from 4 to 5) (Public Health England, 2014b). Wider data also contributes to the picture - a fifth (20%) of Londoners think that people being drunk or rowdy in public places is a problem (MOPAC Public Attitude Survey (PAS), quarter 1 2015/16), a trend that has remained largely stable over the previous year5. Outside of the focus on violence, alcohol is shown to contribute to a range of crime types (e.g., see McSweeney, 2015) including criminal damage and road casualties. Indeed, despite substantial year on year decreases with current figures the lowest on record, the total number of casualties of all severities in drink drive accidents in Great Britain in 2013 was 8,270, of which 1,340 were killed or seriously injured. Around 14 per cent of all deaths in reported road traffic accidents in 2013 involved at least one driver over the drink drive limit (Department for Transport, 2015a). Furthermore, according to the CSEW, around 6.2 per cent of drivers in 2014/15 said they had driven whilst over the legal alcohol limit at least once in the last 12 months, broadly unchanged over recent years (Department for Transport, 2015b). The AAMR proof of concept pilot started on 31 July 2014 with a high profile launch by the Mayor of London at Croydon Magistrates’ Court attracting considerable regional, national and international press coverage. The pilot ran for 12 months6 in the boroughs of Croydon, Lambeth, Southwark and Sutton (which comprise the South London Local Justice Area (LJA)) and aimed to target between 100 and 150 offenders. The aims of the pilot were: 4 Police recording practices impact significantly on crime figures, and it is thought that incidents of violence are more open to subjective judgements about recording and thus more prone to changes in police practice. An inspection conducted by Her Majesty’s Inspectorate of Constabulary (HMIC) on crime data integrity published in August 2014 highlighted issues regarding the classification of crimes across all forces in England and Wales, and the Office for National Statistics state that the renewed focus on standards has particularly affected violence related crime recording. 5 The PAS explores the views of residents across London around crime, ASB and policing issues via face to face interviews with over 12,800 respondents per year. In quarter 1 2014/15 18% of Londoners thought that people being drunk or rowdy in public places was a problem. 6 The pilot has been extended for a further 6 months to January 2016 to allow further time to consider the future use of the AAMR. While performance data will continue to be monitored, the process evaluation and all information contained in this report cover the initial 12 month pilot period only. 8  To test how widely courts use the AAMR, and the technical processes within the criminal justice system.  To evidence compliance rates with the AAMR.  To evidence the effectiveness of ‘transdermal tags’ in monitoring alcohol abstinence. For the purposes of eligibility to receive the AAMR, offenders had to commit an offence for which consumption of alcohol was a contributing factor, reside within one of the four pilot boroughs, and not be dependent on alcohol7. Although not limited by crime type, MOPAC recommended that offences linked to domestic violence were excluded from the AAMR8. The Deputy Mayor for Policing and Crime (DMPC) agreed a budget of up to £260,000 for the proof of concept pilot to cover monitoring equipment and overall project delivery costs. The AAMR is an evidence based innovation inspired by a similar approach from South Dakota, USA which reported reductions in re-arrest of Driving Under the Influence (DUI) offenders (see Loudenburg et al, 2010; Kilmer and Humphreys, 2013), but operated in accordance with UK legislation9. The specific innovation is the first compulsory sobriety scheme of its kind in Europe10 and forms a key part of the MOPAC response to tackling and reducing the volume of alcohol related crime within London. Evidence based policy making - indicative insights The MOPAC Evidence and Insight team - a team of social scientists based within MOPAC - were commissioned to conduct research on the AAMR innovation to generate learning11. The research aims were to:  Describe and assess the set up and implementation of the pilot.  Monitor the basic performance data behind the AAMR.  Assess the technical performance of transdermal devices. 7 The NPS use the Alcohol Use Disorder Identification Test (AUDIT) tool to assess whether the offender is suitable for an AAMR i.e. drinking at non-dependent levels (scoring below 20 on the AUDIT tool). Factors such as the offender’s living situation and personal circumstances should also be taken into consideration prior to proposing the AAMR. 8 During development of the pilot, there were concerns over domestic abuse cases being made subject to an AAMR, before it had been fully tested. This was in relation to potential consequences, such as the abstinence of alcohol creating additional risks for the victim and diverting attention away from specific interventions that are designed to tackle offending behaviour. This position is only applicable for the period of the pilot. Full details of the AAMR, eligibility and suitability criteria, and how it works in practice are available at http://www.london.gov.uk/sites/default/files/AAMR%20toolkit%20FINAL.pdf. 9 One of the key differences between the use of sobriety technology in the USA and UK is the manner in which it is enforced. The US system allows for immediate detention following breach whereas the primary legislation which governs breach action in England and Wales is the Criminal Justice Act 2003 which outlines that an offender is usually returned to court for breach action after a first breach notice has been served and the offender has failed to comply for a second time. 10 Transdermal technology and criminal justice responses with sobriety conditions have been used (or are planned to be used) elsewhere in the UK (e.g. Northamptonshire, Dover, Cheshire, Glasgow, the Home Office Conditional Cautions with Sobriety Requirements pilot), however the AAMR is the first to use the technology in a compulsory, punitive setting. 11 Research outputs (i.e. the interim and final evaluation reports) were peer reviewed by external independent academics. The MOPAC Evidence and insight team were not involved in developing or implementing the AAMR in any way. 9  Assess (as far as possible) the effect of the pilot on offenders, crime, costs and the relevant criminal justice agencies. Building upon an interim review published in March 2015, this report outlines the implementation of the AAMR over the initial 12 month pilot period through the views and experiences of stakeholders involved in pilot delivery, and offenders sentenced to wear the alcohol tag itself, and presents learning to inform any future roll out of the technology. Since the start of the pilot, wider conversations around the use of sobriety orders and alcohol monitoring technology have continued, most notably featuring in the 2015 Conservative Party Manifesto (Conservative Party, 2015: 59). The pilot itself has also been extended for a further six months to January 2016 (although the process evaluation covers the initial 12 month pilot period only), placing this timely report in an essential position to inform any expansion of the scheme. Methods Given the length of the AAMR pilot (12 months), and expected throughput of offenders (100 to 150), it was not possible to robustly evaluate (e.g., randomise or generate a comparison group) the impact of the AAMR on offending behaviour, costs or working practices of stakeholders12. The most appropriate research was a process evaluation to generate learning and develop insights that may influence how future schemes or expansions are implemented. Echoing Dawson and Williams (2009) reflections on the challenges of conducting policing and criminal justice evaluations, this study selected the most feasible robust design approach while stressing the caveats of what the research can and cannot say. A range of methods were used to triangulate learning and address the main research objectives of the pilot (see appendix one for a full evaluation timeline, and survey and interview details). This includes:  Training/awareness raising feedback survey: Fifty five stakeholders (National Probation Services (NPS), Community Rehabilitation Companies (CRC), and judiciary) completed a brief paper survey designed to capture early AAMR understanding and perceptions as part of four initial training/awareness raising events hosted by MOPAC at the outset of the pilot.  Stakeholder surveys: Three online surveys exploring understanding and experiences of the AAMR were conducted with stakeholders across the pilot period. The AAMR project manager emailed a survey link to approximately 55 and 75 12 The Ministry of Justice define a proven re-offence as any offence committed in a one year follow-up period and receiving a court conviction, caution, reprimand or warning in the one year follow-up or a further six month waiting period (Ministry of Justice, 2012). 10 stakeholders at the start, mid-point and end of the pilot13. A reminder email was sent approximately two weeks later and verbal reminders were given in Local Implementation Group and Programme Board meetings. In an attempt to improve response rates (particularly over the summer leave period) a slightly more experimental approach was taken when disseminating the final survey with a personalised invitation to take part including quotations from the AAMR offender cohort to motivate interest. There were 58 responses to surveys in total, with 18 respondents completing more than one survey throughout the pilot period. It was not possible to observe changes over time, therefore survey responses have been collated and overall themes drawn out.  MOPAC and stakeholder interviews: 35 semi-structured interviews exploring views, understanding and experiences of the AAMR in greater depth were conducted with 26 MOPAC officers and stakeholders across the pilot period14. It is highly likely that there was some overlap between survey respondents and interviewees.  Offender surveys: Surveys exploring understanding and experiences of the AAMR, first impressions of the tag, and perceptions of what life might be/was like while wearing the tag were conducted with 44 (out of a possible 113, or 39%) offenders at the time of fitting their tag and 27 (out of a possible 94, or 29%) during tag removal. Surveys were designed by the researchers and given to the offender for self-completion by the EMS officer fitting/removing the tag. Although not without limitations, this was the most practical approach available for obtaining innovative and insightful data on offender views. Completing the survey was not compulsory and some individuals chose not to take part15.  Performance monitoring data: A range of performance data was gathered, including recorded crime, PAS and other emergency services to set the backdrop to the work, on offenders who received the AAMR (e.g., borough of offence and residence, average length of the requirement), and technical data on the tag itself. Police National Computer (PNC) data was also explored to gain insights into the criminal background (or not) of the AAMR offenders. 13 Questions differed slightly in each survey to reflect the stage of the pilot. Previous respondents were not required to answer all questions in later surveys. There was no obligation to take part in surveys therefore respondents were self-selecting. Copies of all surveys are available on request. The number of survey recipients varied at each stage of the pilot due to some staff changes/additions. 14 Potential interviewees were identified with the AAMR project manager and contacted via email by the researchers. There was no obligation to take part in interviews therefore participants were self-selecting. Some interviewees took part in more than one interview at a different stage of the pilot (see appendix one for further details). The majority of interviews were face to face however, due to availability, three were conducted on the telephone. Detailed notes were taken in each interview and analysed to draw out key themes. 15 Offender surveys are continuing throughout the pilot extension period, however will now be sent by EMS staff via text message directly to offenders for self-completion at time of tag fitting and removal. Texting offenders such details is an innovative communication method that may be amenable for future research. 11 Given the size of the research cohort (in stakeholder survey and interviews, and offender surveys), some caution should be used when considering results. Fieldwork data set out in this report only reflect the views of those who took part in surveys and interviews. Results The report organises learning gleaned into the following themes:  Places: including contextual data about the pilot boroughs and a brief analysis of recorded police, PAS and other emergency service data.  People: presenting headline performance data on the actual AAMR cohort over the twelve months of the pilot including technical aspects, compliance levels, and criminal history.  Process: exploring the roll out of the pilot, how it is being used, and its influence, through the views and experiences of stakeholders and offenders. Places This section briefly outlines the four AAMR pilot boroughs that comprise the South London LJA, providing some context around alcohol related crime and disorder and how the pilot boroughs compare to other parts of London. This gives some insights into where may be suitable for any expansion of the pilot. Appendix two presents a range of alcohol related crime indicators and ranks each by borough. The indicators are:  Metropolitan Police Service (MPS) recorded crime where a feature code has been added to indicate that a suspect has been drinking alcohol.  Drink driving arrests.  Alcohol related crime per 1,000 population.  Incidents of night time violence and disorder recorded by ambulance, British Transport Police (BTP) and Transport for London (TfL) bus drivers.  Londoners’ perceptions of people being drunk or rowdy in public places from the PAS. In terms of the pilot boroughs, Southwark is placed in the top ten on all indicators (three in the top five) with the exception of drink driving arrests where it is ranked twelfth. Lambeth ranks in the top ten boroughs on three indicators, most notably second highest in London for alcohol related crime per 1,000 population and incidents of night time violence and disorder recorded by ambulance, BTP and bus drivers. Turning to the remaining pilot 12 boroughs, Sutton (placed at a low level across the indicators except drink driving where it ranks fourth) and Croydon (placed low across indicators except night-time violence where it ranks fifth) would appear to have a less evidenced alcohol issue compared to the other pilot sites, however this indicates how different areas can use the AAMR to target their own local alcohol related problem. Looking elsewhere in the data, Westminster, Camden, Hackney and Newham all rank consistently high across the indicators (e.g., ranked in the top 10 in at least 4 out of 5 indicators) suggesting alcohol related need in other boroughs should the scheme be expanded. The evaluation attempted to explore the ‘pool’ of cases that were both eligible and suitable to receive an AAMR in order to better understand demand and potentially missed opportunities to impose the requirement. Appendix three presents data on Total Notifiable Offences (TNOs)16 and arrests in AAMR pilot boroughs where an alcohol feature code was present17. Although offering some interesting contextual information, data caveats limited the usefulness of this analysis therefore the AAMR project manager conducted a manual review of all Community Based Orders imposed with qualifying offences within the South London LJA between 1 February and 31 March 201518. This indicated that of the 170 offenders that received a Community Based Order for an AAMR qualifying offence within the South London LJA, around a fifth (21%, n=35/170) were deemed eligible for an AAMR within the confines of the pilot (i.e., alcohol was a contributing factor of the offence, the offender resided within one of the pilot boroughs and was not alcohol dependent, the Order did not contain an Alcohol Treatment Requirement (ATR), and offences were not domestic violence related). Two thirds (n=23/35) of eligible cases went on to receive Community Based Orders with an AAMR. It was not possible to ascertain whether lifestyle factors (e.g., alcohol dependency, medical conditions or source of electricity at residence) may have precluded the remaining twelve offenders from receiving an AAMR, however opportunities may have been missed in these cases. The pool of eligible and suitable cases would increase if pilot restrictions (especially geographical boundaries) were removed in any future expansion. 16 Total Notifiable Offences (TNO) is a count of all offences which are statutorily notifiable to the Home Office. 17 ‘MF’, suspect/accused had been drinking prior to committing offence or ‘GA’, alcohol consumed at scene by suspect/accused. Feature codes are not mandatory and therefore it is likely that this data, in part, reflects individual officer recording practices and may considerably underestimate the scale of alcohol related offending. 18 For the purposes of the manual review a qualifying offence was defined as driving with excess alcohol, assaults (e.g., common assault, assault by beating, actual bodily harm or assault on a police constable), criminal damage, public order offences, and other offences such as resisting or obstructing a police constable. 13 People This section presents performance data for the cohort of offenders sentenced to an AAMR within the twelve months of the pilot. Basics around the AAMR In total, 113 AAMRs were imposed over the twelve months of the pilot (see appendix four)19. Ninety four were completed or terminated within the twelve month pilot, increasing to 107 as at 15 October 2015 (including six that were removed by the court in the interest of justice e.g., alcohol dependency or issues with the offender’s place of residence). Overall, the AAMR had a compliance rate of 92% over the twelve months of the pilot, based on the number of cases (n=9) returned to court and convicted of breaching their AAMR as a proportion of all cases imposed20. Of these, five had their AAMR revoked and therefore failed to complete21. The remaining four subsequently went on to complete their AAMR following their return to court. Five of the nine cases convicted of breach were Community Based Orders with a standalone requirement of an AAMR. The remaining four cases were Orders with multiple requirements, one of which was an AAMR22. An alternative method of considering compliance is via positive completions (i.e., those which expire 19 113 AAMRs were imposed on 111 unique individuals. Two offenders were placed on the AAMR for two separate offences. As of 24 November 2015, 135 AAMRs had been imposed. 20 Measuring compliance with Community Orders is complex with no consistent definition (Cattell et al, 2014a). For the purposes of the pilot, compliance with the AAMR requirement (as opposed to the whole Community Based Order which may contain more than one requirement) was measured by the number of offenders returned to court and convicted of breach (e.g., consumption of alcohol, tampering with the AAMR monitoring equipment, or a refusal to allow monitoring to take place), rather than a single failure to comply for which legislation directs an NPS/CRC Responsible Officer to issue a breach notice letter/warning of breach action. Other ways of looking at compliance include successful completions of orders (i.e., positive completions - those which expire normally without being revoked for breach/failure to comply or for a further offence, or which are terminated early by the court for good progress) and initiation of enforcement action (i.e., breach notice letter) by the probation service. See appendix 5 for all compliance data. 21 An AAMR was revoked where the court deemed the breach (including consumption of alcohol (2 cases), consumption of alcohol and tampering with the monitoring equipment (1 case) or refusing to be tagged (2 cases)) of the requirement so serious that continuation of the requirement was not considered appropriate. 22 In all four cases there were only two requirements – AAMR and Unpaid Work. Key learning Data indicates that all of the pilot boroughs present levels of alcohol related need, in particular Southwark and Lambeth. Other London boroughs (e.g. Westminster, Camden, Hackney and Newham) rank consistently high across data indicators suggesting alcohol related need in other parts of London should the scheme be expanded. A two month ‘snapshot’ review of all Community Based Orders imposed within the South London LJA indicates that two thirds (n=23/35) of eligible cases went on to receive Community Based Orders with an AAMR. Opportunities to use the AAMR may have been missed in around a third of eligible cases (n=12/35, however reasons for not imposing an AAMR may have been valid, for example, if the offender was alcohol dependent). 14 normally without being revoked for breach/failure to comply or for a further offence, or which are terminated early by the court for good progress). This gives a completion rate of 95% (based on the number of positive AAMR completions (n=96) as a proportion of all AAMRs completed/terminated (n=101, this figure excludes the six AAMRs that were removed in the interest of justice). See appendix five for further details. Of the 101 AAMRs completed/terminated (excluding the six that were removed in the interest of justice), almost three-quarters (74% or 75 cases), remained totally compliant (i.e., the tag did not record any confirmed drinking or tamper events) throughout the duration of their AAMR. The remaining 26 cases failed to comply at least once (i.e., recorded a drinking event and/or tamper event). In accordance with wider legislation applicable to all community sentences (schedule 8 and 12 of the Criminal Justice Act 2003), an offender is usually returned to court for breach action after a first breach notice has been served and the offender has failed to comply for a second time. Appendix six sets out the enforcement timeline to which all Community Based Orders are subject, which states that cases should be listed before a magistrate court within twenty days of an offender’s second failure to comply. The majority (n=6/9) of AAMR breach prosecutions were conducted within the required time parameters (20 days of the offender’s second failure to comply), with an average of 16 days. The remaining three cases fell slightly outside of the twenty day marker (with 21, 23 and 24 days respectively) due to delays in the enforcement process. Current compliance with the AAMR (92%) is higher than other orders, however it should be noted that direct ‘like-for-like’ comparisons between compliance rates of different orders should be treated with caution due to varying offence types, offender characteristics, processes of dealing with breach, and lengths of orders23. Furthermore, the current study is based on a small number of offenders over a short time period. However, to contextualise, further analysis by the NPS in 2014 estimated a compliance rate of Community Based Orders managed by the NPS and CRC of approximately 61% (based on the projected number of cases and proportion of enforcement referrals (39%)). Just over three quarters (79%) of offenders in the Offender Management Community Cohort Study ended their Community Orders in a ‘positive manner’ with the majority of these (70%) expiring normally and the remainder (9%) completed early for good progress (Cattell et al, 2014a). Indicators of offender compliance in terms of the proportion of orders and licences successfully completed (including recalls) (i.e., those that expire normally without being revoked for breach/failure to comply or a further offence, or which are terminated early by the court for good progress) was 84% in London in financial year 2014/15, higher than the national rate (79%), with some variation by requirement type (e.g., the completion rate for Community Payback in London was 82% but slightly lower for Alcohol Treatment Requirements (80%) 23 The LASPO Act 2012 allows courts to impose a requirement that an offender abstain from alcohol for a maximum of 120 days. In comparison, the maximum period of a curfew is 12 months, while an Unpaid Work requirement can be imposed for a maximum of 300 hours. 15 and Drug Rehabilitation Requirements/Drug Testing and Treatment Orders (67%)) (Ministry of Justice, 2015a). AAMR lengths ranged between 28 and 120 days with an average of 75 days. Over half (61%, n=69) of AAMRs were part of a Community Based Order containing multiple requirements (e.g., Supervision, Unpaid Work, Curfew etc.) with the remainder (n=44) sentenced as a standalone requirement. The majority of AAMRs (83%, n=94) were given as part of a Community Order (the remaining 19 were part of a Suspended Sentence Order). In terms of the technology, there were over 6,500 monitored days in the pilot period during which over 298,000 alcohol readings were taken (an average of over 2,600 readings per order or approximately 45 per monitored day). In theory, the technology should take around 48 readings per individual per day (depending on time of tag fitting and removal) therefore this indicates that the technology underpinning the AAMR is working as intended. Previous research (e.g. Dougherty et al, 2012; Leffingwell et al, 2013) has also indicated correlations between transdermal alcohol readings and other forms of alcohol measurement including breath tests, self report and observations in a laboratory setting24. Alcohol tags should be fitted within 24 hours of sentencing (on the same day if notification is received from the court before 4pm) at a time slot agreed with the offender. In the majority (89%, n=101) of cases, the court notified EMS of the order either on the day of (n=89) or day following (n=12) the sentence. Of these, most (82%, n=83) were tagged either on the day (n=45) or within one day (n=38) of notification25. The 113 AAMRs were ordered in relation to 128 offences. Almost two thirds of these (63%, n=80) were violence (n=41) or drink drive (n=39) related offences. The range of crime types for which AAMRs were ordered listed in appendix seven indicates that sentencers chose to use the requirement across a variety of offences. In-depth analysis of case notes and NPS national management information data (from the nDelius system) conducted by the project manager indicated that around a quarter (24%, n=27) of all AAMR cases were linked to the night-time economy (e.g., committed after 8pm and involving some sort of ‘commercial’ aspect such as a bar, pub, late night food retailer, cab driver etc.). Croydon was the most ‘active’ of the four boroughs throughout the pilot with almost three quarters (73%, n=82) of AAMRs resulting from Croydon Magistrates’ (n=77) or Crown Court (n=5) (see appendix eight). 24 It is not possible to know whether all instances of alcohol consumption were detected in the pilot, however, the tag provides continuous 24/7 alcohol monitoring, uses electrochemical fuel cell technology that is also used in commercially available breath testing devices, and by testing wearer’s sweat every 30 minutes, can detect if someone has consumed a small, medium or large amount of alcohol. The tag stores and records test results every 30 minutes which are referred to as transdermal alcohol concentration (TAC) readings. TAC readings can range from .000 (no alcohol detected) to .487 (x 6 UK drink drive limit) and are mapped to reports for probation that indicate compliance or non-compliance. Specific confirmation criteria are applied to TAC readings/alerts when they are above .020 for three consecutive readings or an hour and a half. The criteria that is used to determine if a subject has consumed alcohol provides for safeguards to prevent false positives that may be associated with ambient alcohol (e.g. cosmetics, work environments, alcohol based products etc.). The tag will also flag up tamper/removal attempts and mechanical issues such as low batteries or other maintenance related issues. During the pilot there was one tag that needed to be replaced and five multiconnect units (the modem that transmits the data) that required replacement due to connectivity issues. The tag stores up to 60 consecutive days of test results while activated, therefore data was secure in these cases. AMS is currently redesigning the base station to include more robust connectivity options. 25 The most common reason for not fitting a tag within 24 hours of notification was due to a ‘no access visit’ i.e., the offender was not at the property when they said they would be, or the tag fitter was unable to gain access to the property. 16 Discussions between the AAMR project manager and staff at Camberwell Green Magistrates Court in addition to points raised in research interviews indicated that reasons for the disparity between the volume of AAMRs ordered at this court compared to Croydon may include geographical restrictions of the pilot, staff turnover, the impact of Transforming Rehabilitation, the suitability of offenders (e.g., a judiciary interviewee felt that those sentenced at Camberwell Green often had chaotic or complex lifestyles that precluded them from receiving an AAMR), and motivation of probation and judiciary staff to use the AAMR. Demographics and criminal background In terms of basic demographics of the offenders who received the AAMR, the majority were male (88%, n=98/111) and white26 (66%, n=73/111), with an average age of 33 years (ranging between 18 and 63 years. Over half (59%, n = 66/111) were aged between 18 and 34 years). Appendix nine presents headline PNC data on 102 offenders sentenced to the AAMR in the twelve month pilot period27. In terms of formal criminal history – the AAMR group present with a cumulative total of 1337 arrests, with an average of 13 arrests each (ranging from 1 to 88); a total of 771 guilty sanctions28, at an average of 8 each (ranging from 1 to 58) and a total of 612 guilty court occasions, at an average of 6 occasions each29 (ranging from 1 to 55). This includes the offence for which they received the AAMR. The majority of the group hold between one and ten (80%, n=82) and 11 and 20 (13%, n=13) guilty sanctions, although there is a lengthy ‘tail’ when exploring overall sanctions (see appendix ten) that indicates a minority of AAMR offenders do present with more prolific levels of crime (e.g., 3 offenders have between 21 and 30, and 4 have more than 30 guilty sanctions). Indeed, one fifth of the AAMR cohort (21%) demonstrated a level of criminal versatility – that is they hold guilty sanctions in four or more different offence types. As a group, their average age of first arrest was 24 years of age, with slightly older average age of first sanction and guilty occasion at court (25 and 26 years respectively). One third (34%, n=35) received the AAMR for their first guilty court occasion. The remainder had at least one other guilty court occasion with almost a third (30%, n=31) having five or more in their history (see appendix eleven). The Offender Group Reconviction Scale version 3 (OGRS3) scores for the AAMR cohort were calculated at the point of receiving the AAMR. OGRS uses static factors, such as age at sentence, gender, offence committed and criminal history to predict the likelihood of 26 Including ‘White: British/English/Welsh/Scottish/Northern Irish’, ‘White European’, ‘White Irish’ and ‘White Other’. 27 PNC records for all 111 unique individuals sentenced to the AAMR in the twelve month pilot period were requested. Data was returned for 102 offenders. PNC analysis in this report is based on 102 records only. 28 This is the number of guilty sanctions (including convictions, cautions, warnings and reprimands) including both court and non-court sanctions. 29 The number of unique court occasions where the outcome was a guilty verdict. Several sanctions could be sentenced within the same court occasion. 17 proven reoffending within a given time (e.g., either one or two years after starting their Community Order. This research reports the two year score). Offenders with a higher OGRS score are at greater risk of reoffending and more likely to breach their Community Orders. For example, 11% of offenders included in the Offender Management Community Cohort Study with a very low risk of reoffending (based on OGRS scores) breached, compared to over a third (34%) of those with a very high risk of reoffending (Cattell et al, 2014a). As a group, the average OGRS3 score was 35% (ranging from 6 to 82) placing the AAMR cohort as a whole at low risk of reconviction. Three quarters of the group would be defined as low (38%, n=39) or very low risk (39%, n=40) although a minority would be identified as medium (18%, n=18) or high (5%, n=5) risk of reconviction (see appendix twelve). To place this in a wider context - as expected, in terms of offending history the AAMR cohort present far less than high demand offending populations and broadly align more to the general offending population in the UK, particularly those who receive community sentences (Farrington, 2005; Dawson and Cuppleditch, 2007, Ministry of Justice 2015b, 2010; Blakeborough and Richardson, 2012) Key learning 113 AAMRs were imposed over the twelve month pilot period, the majority of which were given as part of a Community Order, with an average length of 75 days. Offenders receiving the AAMR were largely male and white, with an average age of 33 years. AAMRs were most commonly given in relation to violence or drink driving related offences. 94 AAMRs were completed/terminated in the pilot period, increasing to 107 as at 15th October 2015. The AAMR had a compliance rate of 92% over the twelve month pilot period, based on the number of cases (n=9) returned to court and convicted of breaching their AAMR as a proportion of all cases imposed. There were over 6,500 monitored days in the pilot period during which over 298,000 alcohol readings were taken (at an average of over 2,600 readings per AAMR or approximately 45 per monitored day), indicating that the technology underpinning the AAMR is working as intended. As expected, the AAMR cohort does not have an extensive criminal background, with an average of 8 sanctions and 6 guilty court occasions each, and an average OGRS score of 35%, although there is a minority that present more prolific histories. As a cohort, in terms of offending histories, the AAMR offenders present far less than high demand offending populations and broadly align more to the general offending population in the UK, particularly those who receive community sentences. DEMOGRAPHICS KEY POINT 18 The AAMR process Drawing from the methods outlined earlier (training/awareness raising feedback surveys, stakeholder surveys and interviews, and offender surveys), this section discusses the AAMR pilot process under four themes: setting up and getting going, delivering the AAMR, influence of the AAMR, and sustainability: insights from the pilot year. Given the size of the research cohort, some caution should be used when considering results30. Setting up and getting going Stakeholders and offenders generally have a good understanding of the AAMR, however there are opportunities for wider publicity… All respondents to the stakeholder surveys indicated that they understood the aims and objectives of the AAMR (n=40/40 respondents rated their understanding as a ‘5’, ‘6’, or ‘7’ with ‘7’ being the maximum - ‘very well’) The majority stated that they understood the eligibility and exclusion criteria for offenders to receive the AAMR (n=38/40), that they had been provided with enough information to use the AAMR in their role (n=36/40), and knew where to get more information about the AAMR (n=33/40). This was proportionately higher than views given in the initial training/awareness survey (conducted throughout June and July 2014 - n=47/54, n=42/54, n=40/48 and n=38/52 respectively). Training/awareness raising sessions were generally well received, however around a quarter of respondents to the training/awareness raising feedback survey expressed dissatisfaction with some feeling that the sessions were delivered too quickly (n=13/53). A small number of respondents to the stakeholder surveys and NPS/CRC interviewees also highlighted some reliance on self-learning and information from colleagues, that the slow initial uptake of AAMRs may have led to staff ‘skill fade’ following training, the importance of face to face sessions due to limited time available to read training materials, and that not all NPS/CRC officers are aware of the AAMR. Despite this, on the whole, respondents to the stakeholder surveys were largely satisfied with the training, awareness raising and communication they had received about the AAMR (n=31/40, a further 5 were neither satisfied nor dissatisfied, 2 had not received any, while the remaining 2 were fairly or very dissatisfied). The majority of survey responses (n=45/58) also indicated that stakeholders were very or fairly satisfied with the communication they had received from MOPAC around the AAMR. Interviewees also largely spoke positively about the training, awareness raising and communications received from MOPAC in terms of it being clear, straightforward and 30 Due to the low response rate to individual surveys, responses to all three surveys have been collated (n=58) and overall themes drawn out. Eighteen respondents completed more than one survey throughout the pilot period, therefore response numbers sometimes reflect the number of surveys completed, rather than the number of individual respondents. Survey respondents who had responded to previous surveys were not required to answer all questions in later surveys therefore response numbers differ depending on question. 19 leaving them well informed to carry out their role in the AAMR pilot. NPS/CRC partners who were unable to attend training sessions or became involved later in the pilot due to staff changes or sick leave valued additional onsite one to one training from the project manager, indicating the need for ongoing training opportunities. There were some mixed views from judiciary respondents around AAMR training, awareness raising and communications. While those who responded to the survey were generally satisfied and some interviewees commented that there had been considerably more information on the AAMR compared to other community sentencing options (with one suggesting that ‘a couple of sides of A4’ of information and guidance would suffice), others felt that there hadn’t been enough training, awareness raising and communications and that more mention should be made of the AAMR in sentencing guidelines/the sentencing judiciary kit. Some NPS/CRC interviewees also felt that more information on the AAMR should have been provided to defence solicitors, who are often unaware of the requirement and eligibility and exclusion criteria. Furthermore, wider publicity in general (to the public as well as professionals) on the AAMR would have been useful, perhaps even serving as a preventative measure to would-be recipients of the tag. Indeed throughout the surveys, respondents consistently ranked prevention of alcohol related crime as the most important goal for the AAMR31. All (n=44) offenders who completed a survey during their tag fitting understood why they had received the order, how the equipment works, what they must do to comply with the AAMR, thought that the information they had received was useful, and stated that they knew how to get more information. During their tag removal the majority of offenders (n=18/27) agreed that it was easy to contact somebody to get more information about the AAMR if they needed to (n=6/27 disagreed with this). Some offenders gave freetext comments in their surveys that the staff fitting or removing their tag were friendly and informative. This generally positive understanding of the AAMR was likely the result of a determined effort by MOPAC to design, implement and communicate effectively from the outset of the pilot and throughout. To illustrate, the implementation was supported by training/awareness raising sessions (held both centrally at City Hall and in pilot boroughs), on site one to one training, a suite of specially designed products (e.g., a toolkit, leaflet, posters), the recruitment of a dedicated project manager and regular, bespoke 31 Survey respondents were asked to rank what they considered to be the most important goals of the AAMR from a set list of 11 options. The full list (in order of importance according to survey respondents were): To prevent people committing alcohol related crime; To improve public safety; To improve public confidence in the ability of the criminal justice system to tackle alcohol related crime; To reduce the cost of alcohol related crime to statutory services (e.g., police, health); To prevent people committing crime in general; To change attitudes about the use of alcohol and acceptability of behaviour; To prevent people drinking excessive amounts of alcohol; To punish offenders; To support people to reduce the amount of alcohol they drink; To speed up the process of dealing with alcohol related offenders in the criminal justice system; To support people to tackle problems in their lives. The AAMR project manager…has been a very useful and accessible source of information, clarification and advice (NPS survey respondent) 20 communications (e.g., ‘seasonal’ messages to remind stakeholders of the AAMR at Christmas) – all of which received broad support from respondents. For example, the AAMR toolkit32 (which set out the aims and objectives of the pilot, ways of working and roles/responsibilities of partner agencies) was well received by stakeholders surveyed who felt it was useful (n=35/39), however some MOPAC staff and stakeholders reflected it would have been more beneficial to have launched it earlier (i.e., it was launched after the initial training although before official pilot start). The majority of stakeholders who stated they had seen the AAMR information leaflet and poster (see appendix thirteen) found them useful33. Regular communications from the project manager and a quick response to questions was important to interviewees and this was often raised as a successful part of the pilot. Some interviewees commented that the pilot may not have run as smoothly as it did without a knowledgeable and responsive project manager who could quickly deal with any issues as they arose34. In respect of the offenders, the majority who responded to a survey at tag removal reported positively on information they received including a guidance document (n= 27/27), questions and answer sheet (n=20/27) and signposting advice (n= 19/27). The findings above around understanding and implementation of the AAMR are encouraging. Previous research suggests that clear training, guidance documents and monitoring instructions are a key part of ensuring a programme is delivered with integrity, enabling it to have the best opportunity to be effective (Dawson and Stanko, 2013). Indeed, insights from the Home Office Sobriety Conditional Cautions scheme highlighted a general lack of stakeholder understanding of the process which contributed to substantial implementation issues experienced by the majority of pilot sites (Home Office, 2013). It is fair to state that MOPAC recognised this, with the interviewed MOPAC staff themselves highlighting the importance of clearly mapping out processes and roles in the set up and delivery of the pilot. Strong engagement and partnership working… 32 Available at https://www.london.gov.uk/sites/default/files/AAMR%20toolkit%20FINAL.pdf 33 N=28/30 respondents who had seen the poster found it useful (however an additional 14 respondents had not seen it); n=35/37 respondents who had seen the leaflet found it useful (with an additional 7 respondents stating they had not seen it). 34 For example, issues that the project manager identified and responded to included: further work with HMCTS to ensure that the AAMR was fully explained to the offender in court (so they knew what was required of them in terms of tag fitting and complying with the AAMR), streamlining communication between the court and EMS tag fitters, supporting offender managers to interpret alcohol monitoring graphs, overseeing the enforcement process in the early stages of the pilot to ensure that enforcement occasions were not missed, and ensuring that offender managers monitored the AAMR mailbox (where communication from EMS was sent) daily. Key learning Both practitioners and offenders surveyed and interviewed held a firm understanding of the AAMR's aims and ways of working - likely attributable to the range of bespoke products MOPAC delivered to ensure integrity of the approach, and regular communication between practitioners and the AAMR project manager. However, publicity of the AAMR could have been more widespread to ensure that legal professionals (in particular, defence solicitors) and the general public were aware of the new technology. Response from project manager when information requests come through is very professional and swift (CRC survey respondent) 21 Given the AAMR is a multi-agency approach, the importance of team work in pilot design and development, securing buy-in from partners (particularly NPS and CRC colleagues who were instrumental in the delivery of the AAMR) and ensuring they took ownership of the process from the start, was recognised by MOPAC staff as critical – with one commenting ‘there would have been no pilot’ without this. Stakeholders interviewed and surveyed generally felt that engagement had worked well, with some survey responses even indicating that the AAMR pilot had enabled them to develop relationships with new partners or improve those with existing partners (n=30/58 and n=32/58 respectively). Although the MPS engaged with the pilot, officers took on less of a role than anticipated at the outset35, however as an organisation that is considerably affected by alcohol related crime it would be useful to explore their role, and that of other partners (e.g., local authorities) further, in the event of any expansion of the AAMR. An MPS licensing officer who responded to the stakeholder survey felt that the AAMR scheme could benefit from police involvement as officers could provide further information which “may help join up the dots and assist the courts and probation service decision to give an AAMR….It could be that the individual is coming to notice in other areas of policing and it would be useful to have a full picture when making assessments on crime prevention strategies”. Some respondents to the stakeholder survey highlighted other partners who may benefit from being involved including alcohol assessors (who could consider the AAMR when an individual is found unsuitable to receive treatment for dependency), Community Payback managers, and health and substance misuse professionals. The working relationship between MOPAC and the equipment providers (AMS/EMS) appeared to be strong from the outset and largely worked well between EMS and delivery stakeholders, some of whom commented in interviews on a speedy response from EMS following queries. However, a small number of NPS/CRC interviewees highlighted occasions early on in the pilot where reports from EMS went to the wrong mailbox and that transmission of information from the court to EMS experienced some initial ‘teething problems’. Learning from these incidents seemed to be taken on board and 35 Initially it was anticipated that the MPS would ensure the influence of alcohol in an offence was highlighted via a flag or marker on the MG5 (case summary) form. Logistically it proved difficult to capture this information, as MG5s are not stored centrally (therefore making it impossible to dip sample to explore content) and flags not used consistently. Attempts were made to understand what prompted an AAMR assessment (e.g., through information on the MG5, via a Pre-Sentence Report (PSR) interview etc.) via a questionnaire to be completed by the NPS court officer, however this largely proved unsuccessful. As such, it was difficult to ascertain how consistently alcohol was flagged on the MG5 or the influence this did or did not have on the identification of eligible cases to receive an AAMR. Good communication between agencies and partners has improved the process. There is willingness for the project to succeed between partner agencies (MPS survey respondent) The planning and implementation that went into the front end of the project paid off in the end as the service delivery went really well (AMS survey respondent) 22 You need people with enthusiasm for the AAMR and a good overall manager to "get people on board with it" (CRC survey respondent) efforts made on the part of MOPAC and EMS to ensure that processes were more streamlined. This was supported by the experiences of the offenders surveyed, most of whom (n=42/44) stated that the process of receiving the alcohol tag (from sentence to having it fitted) was straightforward. Failing to communicate and engage with partners early to ensure they understand the aims of the programme and their role within it has been shown to contribute to implementation failure (Maguire, 2004). Indeed, early and ongoing engagement and buy in with partners and good working relationships has been identified as essential factors in the development and running of other schemes (e.g., see Blakeborough and Richardson, 2012). The AAMR governance structure, including monthly Local Implementation Groups focusing on operational delivery, quarterly Programme Boards setting the strategic direction, and internal project meetings, is likely to have played a key role in engaging stakeholders from the outset and throughout, however a small number commented that this seemed to slow down slightly towards the latter stages of the pilot. Equipment provider (AMS/EMS) interviewees also emphasised the importance of their own governance structure, which allowed them to regularly review and learn from pilot implementation. Furthermore, an important aspect in this strong communication (and wider delivery) was the establishment of a dedicated project team within MOPAC, which included a full-time AAMR project manager seconded from the NPS, who appeared to play a pivotal role in encouraging buy- in from partners, communicating, and supporting the delivery of the AAMR ‘on the ground’. With a strong background in electronic monitoring and enforcement and a range of appropriate tools, the project manager had a firm understanding of working patterns, and well established contacts within HMCTS and the NPS/CRC. Indeed, both MOPAC staff and stakeholders (within freetext comments from the survey and interviews) consistently highlighted the importance of these roles - one particular benefit being the ‘in the field’ practical experience of the project manager. On a wider point, with the AAMR being rolled out during significant changes to offender management as part of the Transforming Rehabilitation agenda, effective engagement with partners in this challenging climate could be viewed as a noteworthy success of the pilot. Delivering the AAMR Key learning Engagement between MOPAC, the equipment provider and delivery stakeholders worked well throughout the pilot. A strong governance structure including regular implementation and programme oversight meetings from the outset and throughout is likely to have played a key role in this. The establishment of a dedicated MOPAC team including the appointment of a project manager with practical ‘in the field’ experience and established relationships ‘on the ground’ appears to have been a pivotal aspect of the effective roll out and implementation of the AAMR. 23 Another punishment requirement for low serious offenders so takes some of the burden off UPW (NPS survey respondent) Identifying and addressing challenges It is expected that pilot schemes change during their lifespan as key learning emerges (Dawson and Stanko, 2013) and the AAMR pilot was no exception. Notable challenges identified throughout the pilot period included geographical restrictions of the pilot areas and the immediacy of contact with the offender after a failure to comply. MOPAC and partners made a variety of amendments throughout the pilot in response to these challenges including: allowing the AAMR to be used in Crown Courts in the South London LJA, as a punitive requirement after breach, and with offenders who commit offences outside of the four pilot boroughs (but within London, reside in one of the pilot boroughs and are sentenced in the South London LJA). In addition, responsibility for initiating first contact with an offender when non-compliance is detected moved to EMS (rather than NPS/CRC) reducing the likelihood of delays due to the Monday to Friday working patterns of NPS/CRC officers (EMS operates a 24 hours a day, 7 days a week service). Identifying challenges and refining the AAMR model in response again demonstrates the importance of the central AAMR team, in particular the dedicated project manager who had well-established relationships with stakeholders. Positive opinions about using the AAMR…. NPS/CRC and judiciary stakeholders surveyed and interviewed were largely positive about the AAMR technology, perhaps supported by previous experience of other forms of electronic monitoring equipment (i.e., curfew tags). NPS/CRC and judiciary interviewees largely welcomed the AAMR, referring to it as 'another tool in the toolbox', ‘armoury’ or ‘arsenal’ and ‘another string to the bow’ which offered an alternative community sentencing option (particularly to Unpaid Work (UPW)) with a punitive element, of which there are reasonably few. They highlighted the benefits of the AAMR as a tailored, targeted and innovative response that accurately monitored alcohol intake and allows offenders to go about their daily life (e.g., employment, care responsibilities etc.) with minimal disruption. Some NPS/CRC interviewees felt that the AAMR contributed to filling a gap in sentencing for alcohol related offences committed by non-dependent offenders, a cohort who were sometimes ‘lumped together’ with dependent drinkers (who may present different criminogenic and lifestyle needs) on an Alcohol Treatment Requirement (ATR). Indeed, almost two thirds of survey responses (n=38/58) indicated that the AAMR was a useful additional tool. Survey responses also indicated support for a wider roll out of the AAMR across London (n=52/58) and nationally (n=51/58), echoed by interviewees, however some stakeholders highlighted in other parts of surveys and interviews that there needed to be further work to fully understand the impact of the AAMR on costs and offending behaviour. 24 Over half (n=33/58) of stakeholder survey responses stated that the AAMR would be more successful when delivered in combination with other requirements, however, a member of the judiciary interviewed warned against ‘crowding’ too many in to one sentence, which may ‘set an offender up to fail’36. In a similar vein, a senior probation officer felt that issuing another punitive requirement such as UPW alongside an AAMR, in some cases, seemed excessive. However, a number of NPS/CRC interviewees spoke about possible benefits of the AAMR alongside a supervision requirement for certain offenders. They felt that the purpose of the requirement and readings from monitoring graphs could be a good focus of conversation around levels of alcohol consumption, and impact on behaviour and general lifestyle. In all AAMR cases, offenders receive Identification and Brief Advice (IBA), and signposting to support services if required, from their Responsible Officer within the NPS/CRC. IBAs are shown to lead to one in eight people reducing their alcohol consumption to within a level which is recognised as safe or low risk, and evidence suggests that they are more effective for harmful and hazardous drinkers than dependent drinkers (Andrews, 2010; Blakeborough and Richardson, 2012; see also Raistrick et al, 2006)37. When asked in later surveys whether views about the AAMR had changed since the start of the pilot, no respondents said they had got worse with the majority stating ‘got better’ (n=24/39) or no change (n=15/39, as they had not dealt with cases or were awaiting results of research). Those who stated it had got better related this to compliance, the technology and monitoring, feedback from staff and service users, partnership, engagement and communications, and having another option to manage alcohol misusing offenders. Some reservations about eligibility criteria and requirement of total abstinence … Some judiciary and NPS/CRC interviewees commented that the AAMR was ‘missed’ as a sentencing option on some occasions, highlighting the need for continued communication and reminders both centrally and ‘on the ground’. A judiciary interviewee felt that sentencers sometimes get into ‘comfort zones’ and may need to be prompted to consider other options. A magistrate survey respondent highlighted the importance of probation colleagues stating that they should support the judiciary by “…considering AAMRs when writing reports so that sentencers can consider whether an AAMR is appropriate and suitable as part of a sentence”. Some NPS/CRC and judiciary interviewees were surprised that the AAMR hadn’t 36 In their study of key predictors of compliance with community supervision in London, Gyateng et al (2010) also noted that the likelihood of breach increased significantly with the number of requirements imposed on an offender. 37 An IBA typically involves identification using a validated screening tool to identify ‘risky’ drinking, and brief advice aimed at encouraging a risky drinker to reduce their consumption to lower risk levels. I may possibly have dealt with one or two cases where AAMR may have been a possible disposal but it does not feature as a significant element in my sentencing armoury in practice as the conditions seem to suit only a very limited range of offences (Judiciary survey respondent) 25 been used more throughout the pilot period and queried if the ‘right type’ of cases were coming through the courts, whether pilot boundaries were overly restrictive, and the impact of frequently updated measures, acts and requirements in general which can sometimes be overwhelming for delivery stakeholders. A perceived narrow eligibility criteria for the AAMR pilot and suitability of the offender assessment tool (AUDIT) was also raised in the freetext comments of some survey responses. Two NPS/CRC interviewees speculated that the judiciary sometimes feel that requiring a person to abstain from alcohol is too punitive and that a more proportionate response would be to use the technology to enforce a reduction (rather than abstinence) of alcohol intake38. An interview with a district judge corroborated these views; however another judiciary interviewee disagreed, stating that a shorter period of enforced abstinence was preferable to a longer spell of reduced alcohol intake. Responses to breach in the USA39 were discussed by judiciary and MOPAC interviewees with some commenting that the pilot had indicated that the response to breach in the UK context was adequate, however another (from the judiciary) felt that the speed of breach process for all Community Based Orders (not exclusively the AAMR) was problematic. Influence of the AAMR As outlined in the methodology section, the research was not able to robustly evaluate the impact of the AAMR on offender behaviour, stakeholder workload or costs. However, it is possible to present some staff and offender insights around the levels of influence the AAMR may have. 38 The LASPO Act 2012 also allows the court to specify that an offender cannot drink more than a specified amount of alcohol, thus allowing for the possibility of minimal drinking rather than abstinence. 39 One of the key differences between the use of sobriety technology in the USA and UK is the manner in which it is enforced. The US system allows for immediate detention following breach whereas the primary legislation which governs breach action in England and Wales is the Criminal Justice Act 2003 which outlines that an offender is usually returned to court for breach action after a first breach notice has been served and the offender has failed to comply for a second time. Key learning The AAMR has been largely welcomed by respondents as ‘another tool in the box’ of community sentences, offering an innovative and tailored response to alcohol related offending, filling a gap in sentencing for alcohol related offences committed by non- dependent offenders. However, there are some reservations around narrow pilot eligibility criteria and the requirement for total abstinence in certain cases. Continued communication is important to ensure that delivery stakeholders continue to recommend and use the AAMR as a sentencing option. 26 On offender lifestyle and behaviour… The majority (n=41/44) of offenders surveyed were very or fairly confident that they would successfully complete the AAMR (i.e., not drink alcohol) at the point of tag fitting. Offenders who completed a survey at the time of their fitting were largely unhappy about the appearance and ‘wearability’ of the alcohol tag. The majority felt that the tag was bulky (n=39/44) while over half (n=27/44) stated that the tag was not comfortable to wear. Indeed, almost two thirds (n=28/44) disagreed with the statement ‘the alcohol tag is better than I thought it would be’. On the whole, views were largely similar at the time of tag removal: almost all (n=26/27) stated that the tag was bulky and two-thirds (n=17/27) that the tag was not comfortable to wear. A similar proportion (n=18/27) disagreed with the statement ‘the alcohol tag is better than I thought it would be’. In freetext fields on the survey, some offenders referred to the size of the tag, that it disturbed their sleep and limited their clothing choices, and concerns about not being able to bath while wearing it40. Some NPS/CRC and judiciary interviewees and survey respondents also saw this as a negative of the AAMR, raising health and safety concerns for offenders with an active or very visible job or lifestyle wearing the tag. Interviewees felt that more information about the AAMR should be provided to offenders at court so they are aware of the tag size and what is required of them. Despite this, most (n=31/44) offenders surveyed reported that they were generally not worried about wearing the alcohol tag at the point of fitting, although there was some concern around what their friends and family (n=28/44) or strangers (n=23/44) would think of the tag. On the whole, views were similar when the tag was removed: the majority (n=21/27) were not worried about wearing the tag, however many were concerned about the views of friends and family (n=19/27) and strangers (n=18/27). Research on the ethics of electronic monitoring of offenders presents different views around stigmatisation effects of wearing a tag, including that it can both hinder development because the offender feels socially excluded and disadvantaged, but also that it may foster a sense of repentance or have a deterrent effect (see Billow, 2014). Offenders were asked for their views on the effect of the AAMR on different parts of their life including family, relationship with partner, children and friends, work, education, health, and attitudes to the police, at time of tag fitting and reflecting back after removal (appendix fourteen)41. Small numbers make results difficult to interpret, however at tag 40 The AAMR tag must not be submerged in water therefore offenders cannot bath while wearing it, however can shower. AMS are currently in the process of testing a new tag strap and buckle designed to be more comfortable for the wearer. 41 A number of ‘life areas’ set out in the survey – in particular ‘relationship with children’, ‘education’ and ‘ability to comply with other orders/programmes’ – were not applicable for some survey respondents. It’s big and ugly. I’m also not happy about the bathing arrangements (Offender survey respondent) …The tag is very big and uncomfortable to wear in bed (Offender survey respondent) 27 fitting few (between 2 and 9) respondents felt that the AAMR would make any of the ‘aspects of life statements’ outlined in the survey worse, with the largest proportion usually stating ‘no change’ or ‘make better’ depending on the statement. Most notably at tag fitting, over half (n=25/44) felt the AAMR would make their health better (n=17/44 stated ‘no change’. Health benefits were also the most frequently mentioned theme in the limited number of freetext comments captured in offender surveys) while almost half (n=20/44) stated that the AAMR would make family life in general better (n=16/44 stated ‘no change’). Respondents were more likely to report ‘no change’ to relationships with friends (n=26/44), work (n=20/44), attitudes to the police (n=27/44) and probation/offender manager (n=25/44). At tag removal, the largest proportion of survey respondents mostly commented that the AAMR had ‘no change’ on the different parts of their life outlined. Similar to the survey conducted at tag fitting, those who responded to the tag removal survey were most likely to report life areas that were ‘made better’ by being on the AAMR as ‘family life in general’ (n=10/27) and ‘health’ (n=13/27). At tag removal, survey respondents were asked to rate their experience of being on the AAMR from ‘1’ being ‘very negative’ to ‘7’ being very positive. Almost half (n=13/27) of respondents rated their experience at the more positive end of the scale (between ‘5’ and ‘7’). Six respondents felt that their experience was ‘very negative’ (ranking ‘1’). Commenting more generally about the potential effect of the AAMR, most offenders surveyed felt that it would be useful in terms of stopping people committing crime (n=29/44) and helping people to drink less alcohol in the long term (n=29/44) at the point of tag fitting. Responses to these questions were similar when the tag was removed (n=18/27 in terms of stopping people committing crime and n=19/27 in terms of drinking less alcohol). Stakeholders had mixed views about the usefulness of the AAMR as a way to tackle offending (n=21/52 in the training/awareness survey disagreed that the AAMR was a useful way to tackle alcohol related offending), and the majority (n=31/58) of responses to the stakeholder survey were unsure whether the AAMR would stop people committing crime in the long term. However, almost two thirds (n=36/58) agreed that being on the AAMR would help people to drink less alcohol in the long term and play a more positive role in society (n=35/58), while almost three quarters (n=42/58) thought that it would help people in other areas of their lives (e.g., work, family, health). Although reporting no impact on re- arrest rates, Blakeborough and Richardson’s (2012) evaluation of the Home Office Alcohol Arrest Referral pilot, found statistically significant reductions in alcohol consumption It can be a precursor to helping people address their alcohol misuse…which they may have not considered before (CRC survey respondent) I have stopped binge drinking for the long term and I have seen how good life can be without drinking (Offender survey respondent) 28 It has given me the opportunity to work with people with alcohol issues in a different way - it provides a period where they are not drinking and gives them the opportunity to see the difference this can make - this gives me something to build on (CRC survey respondent) between the time of offenders receiving brief alcohol interventions in a criminal justice setting and the follow up period (for those who were able to be contacted). Although the authors were not able to determine whether changes were a result of the pilot, impressions from offenders were that the intervention may have prompted reflections on drinking behaviour and identified motivational levers. As part of the pilot, the AAMR is considered to be a punitive response to alcohol related offending, however both NPS/CRC and judiciary interviewees highlighted possible rehabilitative elements as a welcome unintended consequence. Although unsuitable for alcohol dependent offenders and, for the purposes of the pilot, unlikely to be used extensively with ‘hardened career criminals’ thus limiting the likely impact on offending behaviour, interviewees felt that a period of abstinence on the AAMR had the potential to give offenders a ‘pause’ in drinking, time to reflect on alcohol consumption and its impact on offending behaviour, relationships and work, and an opportunity to break the cycle of routine drinking. Even if not sustainable in the longer term, some felt that the duration of the AAMR at least may give respite to families, communities, local police and health providers, and allow offenders to focus on other areas of their life. Some NPS and CRC interviewees gave examples of how they had tailored products around the AAMR to support offenders further and use the opportunity as a ‘teachable moment’, including letters sent at the end of the AAMR and literature/advice around alcohol consumption going forward. Positive relationships between an offender and Offender Manager and tailoring discussions to individual needs may reduce the likelihood of an offender breaching their Community Order (Cattell et al, 2014b), however some research suggests more mixed results about the impact of relationships (see McSweeney et al, 2013). On costs and stakeholder workload… Although attempts were made in surveys and interviews to collect data around time taken to assess, prepare, induct, enforce and manage AAMRs in comparison to other requirements, respondents often found this difficult to quantify due to the varying nature of cases and workloads. It is a useful tool that allows people a time of reflection regarding their drinking and to see changes that this brings about (CRC survey respondent) 29 There was limited evidence to draw conclusions about the effect of the AAMR on stakeholder workload, with different parts of the fieldwork presenting mixed views. Early indications from conversations with a small number (n=3) of NPS/CRC staff presented at interim report stage in appendix fifteen suggested that while time required to assess, prepare, induct and enforce the AAMR were largely comparable to other requirements, day to day management of the AAMR required considerably fewer NPS/CRC staff hours (around 3 hours per case) compared to other requirements including UPW, supervision and curfews (between 15 and 25 hours). However, on reflection, one of the stakeholders who took part in these early conversations said in an interview at mid-point of the pilot that it was perhaps taking slightly longer. This was due to managing the AAMR email box and in some cases where NPS/CRC officers introduced their own processes such as writing letters to offenders at the end of their AAMR to advise around changes in alcohol tolerance levels and drinking patterns in the future. It was clear from the stakeholder survey that the AAMR had not reduced workloads (n=30/58 disagreed with the statement ‘the AAMR has reduced my workload’) and most responses disagreed (n=23/58) that offenders were dealt with any more quickly on the AAMR compared to other requirements. A small number of freetext comments in stakeholder surveys suggested that AAMR assessments and inductions took slightly longer. However, interviews with NPS and CRC stakeholders indicated that, although the AAMR did introduce new work (some elements of which may initially take slightly longer while getting to grips with new language, processes and technology etc.), this was not particularly onerous or prohibitive, and in most cases largely similar to other requirements (although it was acknowledged that this could change as caseloads increase). The AAMR had no impact on the workload of judiciary interviewees as it required similar resources to other community sentences. One judiciary interviewee speculated that the AAMR had the potential to reduce workloads as repeat alcohol offenders may be less likely to return to court, at least for the duration of their order. NPS/CRC staff require ‘time and space’ to ensure a case is eligible and an offender suitable to receive an AAMR. This was an issue frequently iterated by an interviewee from the NPS who felt that, to date, NPS/CRC staff had been allowed this with no pressure from the judiciary to conduct ‘quick time’ assessments, and was keen that this should continue if the AAMR is rolled out more widely. A full cost benefit analysis of the AAMR could be explored when more cases are available and a longer time period has elapsed to consider implications for reoffending and wider impact on health and other partners. 30 Sustainability: insights from the pilot year In both surveys and interviews, stakeholders shared their reflections on the pilot year and ‘lessons learned’ that they felt should be considered in any plans to expand the AAMR (although some highlighted the need for more in-depth evaluation to fully understand the impact on costs and offending behaviour). Key insights included:  A clear operating model: Clarity around scope and expected delivery was important in the AAMR scheme and the MOPAC team worked hard to ensure this was maintained throughout the pilot. Any future schemes should establish clear, documented processes and channels of communication, outline roles for key partners, and provide training and guidance material for delivery stakeholders and sentencers. This should be supported by solid governance arrangements that encourage continuous learning and improvement. A MOPAC interviewee emphasised the value of 'keeping it simple', avoiding unnecessary or complicated processes.  Engagement with partners: MOPAC staff interviewed emphasised the importance of strong partnership working from the outset and throughout the Key learning Offenders surveyed were largely unhappy about the appearance and ‘wearability’ of the tag, however overall were positive they could complete the order. Some felt that there might be health benefits from being on the AAMR. Interviewees felt that a period of abstinence on the AAMR had the potential to give offenders a ‘pause’ in drinking, time to reflect on alcohol consumption and its impact on offending behaviour, relationships and work, and an opportunity to break the cycle of routine drinking. Some NPS and CRC interviewees gave examples of how they had tailored products around the AAMR to support offenders further and use the opportunity as a ‘teachable moment’, including letters sent at the end of the AAMR and literature/advice around alcohol consumption going forward. There was limited evidence to draw conclusions about the effect of the AAMR on stakeholder workload, with different parts of the fieldwork indicating mixed views with some stakeholders commenting that assessments and inductions may take slightly longer (e.g., while getting to grips with new language, processes and technology etc.). However, this was not felt to be particularly onerous or prohibitive and in most cases largely similar to other requirements. 31 AAMR pilot, highlighting the need for engagement at the 'right' organisational level (i.e., enthusiastic decision makers who can 'champion' the work), securing buy in (rather than just consensus), and involving partners in project design (not just expecting them to deliver). Good relationships with equipment providers are also pivotal. This was strong throughout the MOPAC pilot with clear and open communication which assisted with continuous review and service improvement as the scheme progressed. A survey respondent from the MPS suggested that there could be a greater role for partnership work with the police going forward, both in terms of identifying offenders who may be suitable to receive an AAMR and sharing information.  Dedicated staff: The role of the AAMR project manager was frequently mentioned by stakeholders throughout the evaluation as a positive feature, with some suggesting that the pilot may not have run as smoothly without his continued communication and quick response to questions. From a MOPAC perspective, the 'on the ground' experience and well established contacts of the project manager seconded from the NPS were invaluable. Stakeholders interviewed and surveyed also highlighted the importance of a dedicated person in each partner agency who can promote the AAMR, problem solve and answer questions. The value of a single point of contact (SPOC) was also recognised in Home Office findings around the Sobriety Conditional Cautions scheme (Home Office, 2013). Issues around sustainability of these resources and ‘scalability’ of the pilot should be considered if the scheme is expanded.  Delivering the AAMR: Good quality assessment and induction, along with clear guidance around eligibility and suitability, and effective communication between partners, (particularly the courts, tag fitters and NPS/CRC) were identified by stakeholders surveyed and interviewed as essential factors in the AAMR pilot. Continued communication both centrally and ‘on the ground’ (e.g., between NPS/CRC and sentencers) is important as the AAMR ‘beds in’ to ensure that it is not missed as a sentencing option.  A ‘teachable moment’ with offenders: Some NPS/CRC officers interviewed spoke about opportunities for the AAMR to be used as a ‘teachable moment’ and outlined processes they had developed including advice to offenders (face to face or via letter) around changes in tolerance levels following a period abstinence and relationships with alcohol going forward. Billow (2014) argues that, when combined with other crime prevention measures including education, electronic monitoring has the potential to impact on rehabilitation. Although a punitive measure for the purposes of the pilot, the technology may present wider opportunities around addressing offending and other negative behaviour. 32  More information for offenders at court: A number of NPS/CRC interviewees felt that more information about the AAMR should be made available to offenders in court, including examples of the alcohol tag and monitoring equipment, so individuals have a better understanding of what will happen during tag fitting.  Opportunities for wider use of the technology: As the scheme progressed, some partners reflected on the pilot boundaries, suggesting that there were potentially more innovative uses for the AAMR beyond current restrictions. This included domestic violence cases (as part of a package of measures to address offending behaviour42), wider responses to drink driving offences, and to support dependent drinkers as part of their treatment programme (although not via abstinence). Kilmer and Humphreys (2013) also recognised the opportunities for exploring wider uses of sobriety schemes, perhaps in line with treatment for dependent drinkers. A judiciary interviewee was interested in further work to explore the length of time on a tag required in order to see positive changes in behaviour. The South Dakota Model found greater reductions in future offending compared to control groups for participants with at least 90 consecutive days of alcohol testing (although there were still lower rates of future offences compared to matched controls with participants with at least 30 days) (Loudenburg et al, 2010). It would be interesting to explore this further in a UK context. 42 Kilmer and Humphreys (2013) reported a 9% decrease in domestic violence arrests as part of the 24/7 Sobriety Program in South Dakota. 33 Concluding thoughts This report helps to build the evidence base to inform discussions around the AAMR work to date, and also to inform any roll out of the AAMR across London and beyond. Whilst it is too soon at present to robustly evaluate the impact of the AAMR on offending behaviour or costs, this process evaluation generates learning on the proof of concept pilot through the views and experiences of stakeholders involved in design and implementation, and the offenders themselves who were sentenced to wear the alcohol tag. The AAMR has been generally welcomed by delivery stakeholders as an additional community sentence option that offers an innovative and tailored response to alcohol related offending. Learning generated from the pilot year presents a positive message in terms of offender compliance, the technology working as expected, and programme implementation, with a strong understanding of the aims of the pilot and how the AAMR works in practice amongst both offenders receiving the sentence and stakeholders involved in its delivery. However, the fieldwork identified offender and stakeholder concerns about the appearance and ‘wearability’ of the alcohol tag, while some stakeholders highlighted the importance of ongoing communication to ensure that sentencers continue to use the AAMR as a sentencing option. Clear communication and consistent engagement with stakeholders from the outset of the pilot and throughout, and a project manager with ‘on the ground’ experience and well established contacts has meant that MOPAC has been able to identify challenges and amend the model accordingly throughout the pilot. There is considerable evidence that highlights the value of effective implementation and the sheer challenge when implementing innovation (Dawson and Stanko, 2013). In this context, the positive findings around AAMR implementation cannot be over stated. However, consideration should be given to the sustainability of this level of project management should the scheme be expanded. This report contributes to a currently limited British research evidence base around interventions to address alcohol related offending in a criminal justice context (McSweeney, 2015; McSweeney et al, 2009). Wider roll out of the AAMR would provide a larger sample size and opportunities to explore innovative approaches to selecting those individuals to receive the sentence (e.g., randomisation). Future studies could generate a valid counterfactual to enable confident conclusions, and explore the impact of the AAMR on offending behaviour, costs and wider possible benefits including health, community and economic outcomes. In light of plans in the 2015 Conservative Party Manifesto to make sobriety orders more widely available, and the extension of the AAMR pilot for a further six months to January 2016, this timely report offers useful insights to inform any expansion of the scheme. Awareness about the use of sobriety orders and the technology that underpin them in a UK context is still developing. This process evaluation of the AAMR proof of concept pilot is a basis on which to build knowledge, and can be used as a blueprint for future evaluation efforts in this area. 34 Appendices Appendix 1: Evaluation timeline, survey and interview details Survey details Survey type No. of respondents Respondent details Start of pilot (September 2014) 19 14 x National Probation Service (NPS) and Community Rehabilitation Company (CRC) 5 x Alcohol Monitoring Services Ltd (AMS - manufacturer of the transdermal alcohol monitoring equipment) and Electronic Monitoring Services (EMS – responsible for fitting, removing and maintaining the equipment) Mid-point of pilot (March 2015) 24 9 x Judge/Magistrate 9 x NPS and CRC 2 x AMS and EMS 2 x Her Majesty’s Courts and Tribunal Service (HMCTS - legal advisors) Jun14 Jul14 Aug14 Sep14 Oct14 Nov14 Dec14 Jan15 Feb15 Mar15 Apr15 May15 Jun15 Jul15 Training/awareness raising session feedback surveys Initial stakeholder survey; interviews with key MOPAC staff Interviews with key delivery stakeholders Mid-point stakeholder survey; interviews with key MOPAC staff and delivery stakeholders End of pilot stakeholder survey; interviews with key MOPAC staff and delivery stakeholders Performance monitoring data collection; offender surveys; desk based analysis (e.g., review of key literature, PNC and recorded crime analysis) ‘Proof of concept’ pilot launch (31 July 2014) Interim evaluation report (Feb 2015) Final evaluation report (Autumn 2015) 35 1 x Metropolitan Police Service (MPS) 1 x Local Authority Nine respondents stated that they had completed a previous survey. Surveys were anonymous; as such it was not possible to track responses. End of pilot (August 2015) 15 11 x NPS and CRC 2 x AMS and EMS 1 x MPS 1 x Local Authority Nine respondents stated that they had completed a previous survey. Surveys were anonymous; as such it was not possible to track responses. Interview details Interview type No. of interviews Interviewee details Start of pilot (September 2014 – January 2015) 7 3 x MOPAC officers 2 x Judge/Magistrate 2 x NPS and CRC Mid-point of pilot (April - May 2015) 9 4 x NPS and CRC 3 x MOPAC officers (also interviewed at start of pilot) 2 x Judge/Magistrate (one had been interviewed at start of pilot) End of pilot (July - September 2015) 19 10 x NPS and CRC (two had been interviewed at start of pilot) 3 x AMS and EMS 3 x Judge/Magistrate (one had been interviewed at start of pilot) 2 x MOPAC officers (also interviewed at start and mid- point of pilot) 1 x HMCTS 36 Volume Rank Volume Rank Crude rate per 1,000 population Rank Volume Rank % of respondents answering 'problem' Rank Barking and Dagenham 786 11 321 2 10.53 7 470 20 35 2 Barnet 997 8 207 13 7.0 25 416 23 10 28 Bexley 654 18 174 20 5.8 31 292 28 7 32 Brent 703 15 213 11 9.3 16 680 10 14 21 Bromley 704 14 351 1 6.7 26 422 22 10 28 Camden 802 9 183 18 10.28 9 812 7 24 8 Croydon 649 20 190 17 9.2 17 821 5 18 14 Ealing 1049 5 290 5 9.9 14 575 14 28 6 Enfield 540 24 157 24 8.1 22 503 17 14 21 Greenwich 742 13 251 7 9.2 17 594 13 17 17 Hackney 794 10 113 28 10.63 4 753 8 30 5 Hammersmith and Fulham 676 16 166 22 10.16 10 349 25 19 12 Haringey 593 22 205 14 10.0 12 726 9 17 17 Harrow 293 31 140 25 6.6 28 264 29 10 28 Havering 615 21 168 21 7.5 23 381 24 18 14 Hillingdon 1139 3 195 15 8.6 19 507 15 13 25 Hounslow 1249 2 192 16 9.8 15 501 18 25 7 Islington 1006 7 93 29 10.92 3 632 11 20 11 Kensington and Chelsea 538 25 21 30 8.5 21 308 27 19 12 Kingston upon Thames 433 28 319 3 6.1 30 261 30 14 21 Lambeth 403 29 161 23 11.02 2 1016 2 23 9 Lewisham 574 23 182 19 10.1 11 605 12 14 21 Merton 476 27 0 31 6.7 26 314 26 15 20 Newham 1019 6 127 27 10.59 4 962 4 40 1 Redbridge 650 19 217 10 8.6 19 464 21 18 14 Richmond upon Thames 260 32 0 32 5.6 32 196 32 9 31 Southwark 1128 4 208 12 10.55 4 1002 3 23 9 Sutton 538 25 304 4 6.5 29 254 31 13 25 Tower Hamlets 747 12 132 26 10.53 7 816 6 34 3 Waltham Forest 675 17 220 9 10.0 12 491 19 16 19 Wandsworth 353 30 262 6 7.3 24 506 16 13 25 Westminster 1553 1 244 8 14.42 1 1348 1 31 4 Borough Incidents of night time violence and disorder recorded by Ambulance, British Transport Police and TFL bus drivers (January - December 2014) Respondents to MPS PAS who think that people being drunk or rowdy in public places is a problem? (Rolling 12 months to quarter 2 2014/15) Public Health England Local Alcohol Profiles for England alcohol related recorded crime (2012/13) Total Notifiable Offences (TNOs) with feature code MF ('suspect/accused has been drinking prior to committing offence') and GA ('alcohol consumed at scene by suspect/accused') recorded by the Metropolitan Police Service (Rolling 12 months to January 2015) Drink driving arrests (positive breath test, refusing to provide a breath test) recorded by the Metropolitan Police Service (Rolling 12 months to January 2015) Notes on appendix 2  The MF and GA feature codes are not mandatory therefore this data, in part, may reflect recording practices/use of the feature code by officers (which can differ by borough), rather than an accurate picture of alcohol related offending.  Recorded drink driving arrests may be skewed by police activity/operations in different boroughs.  Public Health England alcohol related recorded crime (based on the Home Office’s former ‘key offence’ categories), all ages, persons, crude rate per 1,000 population. Office for National Statistics 2011 mid-year populations. Attributable fractions for alcohol for each crime category were applied where available, based on survey data on arrestees who tested positive for alcohol by the UK Prime Minister’s Strategy Unit.  Night time violence and disorder recorded by Ambulance, British Transport Police (BTP) and Transport for London (TfL) bus drivers includes incidents between 7pm and 7am covering/merging the following data sets: BTP recorded incidents at a station coded as violence or disorder; Ambulance call outs to assault; TFL Bus driver reported violence and disorder; London Fire Brigade deliberate fires (comparatively small numbers). Appendix 2: Alcohol related crime indicators by borough 37 Appendix 3: TNOs and arrests in AAMR pilot boroughs, 1 August 2014 to 31 July 2015 TNO 105,284 TNO (feat. MF/GA) 2,697 Total Arrests (feat.MF/GA) - Residing only 1 787 Total Arrests (feat.MF/GA) - Residing and Committed 2 406 Drink Driving Arrests 3 732 Notes on appendix 3  MF feature code = Suspect/accused had been drinking prior to committing offence.  GA feature code = Alcohol consumed at scene by suspect/accused.  Feature codes are not mandatory therefore this data, in part, may reflect recording practices/use of the feature code by officers (which can differ by borough), rather than an accurate picture of alcohol related offending.  1= Where arrested individual resides in one of the four pilot boroughs and committed an offence within the MPS. This number excludes domestic abuse related offences.  2 = Where an arrested individual resides and committed an offence in one of the four pilot boroughs.  3 = Positive breath test/ refusal of breath test. 38 Appendix 4: Headline performance data on AAMRs imposed, 31 July 2014 to 30 July 2015 Headline performance data on AAMRs imposed, 31 July 2014 – 30 July 2015 Total number of AAMRs 113 Number of AAMRs completed/terminated 94 (this increased to 107 as at 15/10/15) Compliance 92% Based on the number of cases (9) returned to court and convicted of breaching the AAMR as a proportion of all AAMRs imposed to date (113). See appendix 5 for further details Arresting borough (In 2 cases the arresting borough was unknown) Croydon 57 Lambeth 19 Southwark 14 Sutton 21 Borough of residence Croydon 56 Lambeth 15 Southwark 15 Sutton 27 Sentencing court Croydon 82 Camberwell Green 31 Community Based Order with a standalone requirement of an AAMR 44 Community Based Order with multiple requirements one of which is an AAMR 69 Community Orders 94 Suspended Sentence Orders 19 Average length of AAMR 75 days Range of length of AAMR Upper 120 days Lower 28 days Reason for ending AAMR (data to 15/10/15) Completed 96 Revoked following breach 5 Removed in the interest of justice (see appendix 5) 6 Total monitored days 6,584 Total readings taken 298,004 39 Appendix 5: AAMR compliance (data to 15 October 2015) Total number of AAMRs imposed between 31st July 2014 & 30th July 2015: 113 Completed or terminated AAMRs: 107 Live cases: 6 (As of 15/10/2015) Offenders were 100% compliant: 75 (i.e., no confirmed drinking or tamper events) Awaiting breach proceedings or conclusion of the Court case: 4 (3 warrants remain outstanding) Offenders failed to comply at least once: * 26 Offenders went on to complete their AAMRs: 21 AAMR/Order revoked by Court following breach action and failed to complete their AAMRs. 5 Offenders were returned to Court for breach action and subsequently went on to complete their AAMRs: 4 Offenders received a breach notice letter (formal warning) and subsequently completed their AAMR: 17 AAMR removed from the Order by the Court in the interest of justice** 6 Cases where breach action is currently under review: 2 Please note: * In accordance with legislation, an offender would usually be returned to court for breach action after the 2nd alleged failure to comply and after a 1st breach notice has been served. Examples of failures to comply include consuming alcohol or tampering with the tag. ** ‘Interest of justice’ includes cases where the AAMR has become inappropriate due to factors such as: alcohol dependency, work related health and safety concerns, or issues with the offender’s place of residence. Where such circumstances have arisen, the cases have been returned to Court for applications to amend the Order. 40 Appendix 6: Enforcement timeline for Community Based Orders *This is a historical indicator but is often still observed by the NPS in order to ensure expedience with breach proceedings. Offender’s 2nd failure to comply with the Community Order or Suspended Sentence Order Day 0 Day 10 (Target) Day 5 Last day for offender to supply an acceptable reason and supporting evidence By day 20 Case listed before Magistrates’ Court for a first hearing 60% of all Magistrates’ Court breach cases to be concluded by day 25 Probation to apply for a Summons by day 10 in 90% of all breach cases at the Magistrates’ Court By day 25 (Indicator)* Please note: High risk cases are prioritised and enforcement proceedings may be initiated by an application for a warrant before a court following, in some circumstances, a single failure to comply. Offender’s 1st failure to comply with the Community Order or Suspended Sentence Order results in a Breach Notice letter (warning letter), as directed by legislation 1st failure to comply 41 Appendix 7: Offence types for which an AAMR was ordered, 31 July 2014 to 30 July 2015 Offence types Number of offences for which AAMR was ordered43 Driving Driving with excess alcohol 35 Failure to provide specimen for analysis 2 Being in charge of a motor vehicle while unfit through drink or drugs 1 Failing to stop after an accident 1 Violence Assault by beating or common assault 29 Assault on a Police Constable 11 Resisting/obstructing a Police Constable 1 Disorderly behaviour/ harassment Threatening words and behaviour 13 Causing a nuisance/disturbance without reasonable excuse on NHS premises 1 Drunk and disorderly conduct 3 Racially aggravated harassment or harassment 7 Damage/theft Criminal damage 9 Theft 4 Burglary 3 Aggravated taking of a vehicle 1 Possession – linked to the above offences, not standalone Possession of an offensive weapon (in combination with an offence listed above) 1 Possession of a bladed article (in combination with an offence listed above) 3 Possession of cannabis (in combination with an offence listed above) 1 Other Offensive/indecent/obscene/menacing message 1 Breach of an Anti-Social Behaviour Order (ASBO) 1 Total 128 43 Some cases involved multiple offences. 42 Appendix 8: Month by month breakdown of AAMRs issued by court Court Aug 1444 Sep 14 Oct 14 Nov 14 Dec 14 Jan 15 Feb 15 Mar 15 Apr 15 May 15 Jun 15 Jul 15 Total Croydon Magistrates’ Court 5 6 4 2 4 12 3 14 5 8 8 6 77 Croydon Crown Court N/A N/A N/A N/A N/A 1 1 0 1 0 1 1 5 Camberwell Green Magistrates’ Court 0 2 4 5 3 3 3 3 5 0 2 1 31 Total 5 8 8 7 7 16 7 17 11 8 11 8 113 44 August figures include one AAMR given on 31 July 2014. 43 Appendix 9: Headline PNC data on offenders sentenced to the AAMR, 31 July 2014 to 30 July 201545 Headline PNC data on offenders sentenced to the AAMR, 31 July 2014 – 30 July 201546 Arrests (The number of offence entries on PNC for an individual. This includes any outcome for the offence, guilty or not, and multiple arrests on the same occasion) Total number of arrests for AAMR cohort 1337 Average number of arrests 13 Range of number of arrests Upper 88 Lower 1 Average age of first arrest 24 Age range of first arrest Upper 54 Lower 11 Sanctions (The number of guilty entries on PNC. This includes court and non-court sanctions) Total number of sanctions for AAMR cohort 771 Average number of sanctions 8 Range of number of sanctions Upper 58 Lower 1 Average age of first sanction 25 Age range of first sanction Upper 54 Lower 11 Criminal versatility 21% have sanctions for 4 or more offence types Guilty sanction occasions at court (The number of court occasions where the outcome was a guilty verdict. Several offences could be tried at one court occasion) Total number of guilty sanction occasions at court for AAMR cohort 612 Average number of guilty sanction occasions at court 6 Range of number of guilty sanction occasions at court Upper 55 Lower 1 Average age of first guilty sanction occasion at court 26 Age range of first guilty sanction occasion at court Upper 54 Lower 12 45 Based on PNC records returned for 102 offenders 46 Based on PNC records returned for 102 offenders 44 Appendix 10: Total number of sanctions (including AAMR) of offenders on the AAMR 45 Appendix 11: Total number of previous guilty court occasions of offenders on the AAMR Number of previous guilty court occasions Count of AAMR cohort % 0 35 34 1 13 13 2 13 13 3 7 7 4 3 3 5 or more 31 30 Total 102 100 Appendix 12: OGRS3 scores for offenders on the AAMR OGRS3 score No. of AAMR offenders % Very low (0-24%) 40 39 Low (25-49%) 39 38 Medium (50-74%) 18 18 High (75-89%) 5 5 Total 102 100 46 Appendix 13: AAMR poster 47 Appendix 14: Offender views on effect of AAMR on their life at time of tag fitting and removal (tag fitting n = 44; tag removal n = 27) Better Worse No change N/A Tag fitting Tag removal Tag fitting Tag removal Tag fitting Tag removal Tag fitting Tag removal Family life in general 20 10 6 5 16 12 2 0 Relationship with partner 14 5 6 6 13 12 11 4 Relationship with children 11 6 4 2 12 7 17 12 Relationship with friends 10 6 7 7 26 11 1 3 Work 11 7 9 5 20 9 4 6 Education 4 5 3 3 16 10 21 9 Health 25 13 2 6 17 8 0 0 Attitudes to the police 11 6 3 3 27 15 3 3 Attitudes to probation/offender manager 11 9 3 2 25 14 5 2 Ability to comply with other orders/programmes N/A 7 N/A 0 N/A 10 N/A 10 Appendix 15: Hours required to implement a selection of requirements under a Community Based Order based on conversations with a small number of NPS/CRC staff at mid-point of pilot Assessment UPW = 5 mins Curfew = 5 mins Alcohol Treatment Requirement (ATR) = 30 mins AAMR = 5 – 10 mins Initial prep & review UPW Curfew ATR AAMR General CRC induction UPW, supervision, ATR etc. = 45 mins to 1.5 hours AAMR induction = 1 to 1.5 hours Day to day management UPW, supervision, curfew, ATR = 15 to 25 hours AAMR = 3 hours Enforcement UPW Supervision Curfew AAMR NPS Court Officer Responsible officers (1 x Probation Services Officer, 1 x Probation Officer) All 2 hours All 20 to 30 mins 48 References Andrews, M. (2010) Case for change – Commissioning Identification and Brief Advice to improve health and justice outcomes in offender populations London: Department of Health Billow, W. (2014) ‘Electronic Monitoring of Offenders: An Ethical Review’ Science and Engineering Ethics 20 (2): 505-518 Blakeborough, L. and Richardson, A. (2012) Summary of findings from two evaluations of Home Office Alcohol Arrest Referral pilot schemes London: Home Office Cattell, J., Mackie, A., Capes, T. and Lord, C. (2014a) Implementation of Community Orders: Results from the Offender Management Community Cohort Study London: Ministry of Justice Cattell, J., Kenny, T., Lord, C. and Wood, M. (2014b) Community Orders with Punitive Requirements; Results from the Offender Management Community Cohort Study London: Ministry of Justice Conservative Party (2015) The Conservative Party Manifesto 2015 London: The Conservative Party Dawson, P. (2005) Early Findings from the Prolific and Other Priority Offenders Evaluation Home Office Development and Practice Report 46 London: Home Office Dawson, P. and Cuppleditch, L. (2007) An Impact Assessment of the Prolific and other Priority Offender Programme Home Office Online Report 08/07 London: Home Office Dawson, P. and Stanko, E. (2013) ‘Implementation, Implementation, Implementation: Insights from Offender Management Evaluations’ Policing: A Journal of Policy and Practice 7 (3): 289-298 Dawson, P. and Williams, E. (2009) ‘Reflections from a Police Research Unit – an Inside Job’ Policing: A Journal of Policy and Practice 3 (4): 373-380 Department for Transport (2015a) Reported Road Casualties in Great Britain: Estimates for Accidents Involving Illegal Alcohol Levels: 2013 (final) and 2014 (provisional) London: Department for Transport Department for Transport (2015b) Self-reported drink and drug driving: Findings from the Crime Survey for England and Wales 2014/15 London: Department for Transport Dougherty, D., Charles, N., Acheson, A., John, S., Furr, R. and Hill-Kapturczak, N. (2012) ‘Comparing the Detection of Transdermal and Breath Alcohol Concentrations during 49 Periods of Alcohol Consumption Ranging from Moderate Drinking to Binge Drinking’ Experimental Clinical Psychopharmacology 20 (5): 373-381 Farrington, D. P. (2005) Integrated Developmental and Life-Course Theories of Offending New Brunswick, NJ: Transaction Gyateng, T., McSweeney, T. and Hough, M. (2010) Key Predictors of Compliance with Community Supervision in London London: London Criminal Justice Partnership Home Office (2013) Using Conditional Cautions with Sobriety Requirements London: Home Office Kilmer, B. and Humphreys, K. (2013) ‘Losing Your “License to Drink”: The Radical South Dakota Approach to Heavy Drinkers Who Threaten Public Safety’ The Brown Journal of World Affairs 20 (1): 267-277 Leffingwell, T., Cooney, N., Murphy, J., Luczak, S., Rosen, G., Dougherty, D. and Barnett, N. (2013) ‘Continuous Objective Monitoring of Alcohol Use: 21st Century Measurement Using Transdermal Sensors’ Alcohol Clinical Experimental Research 37 (1): 16-22 Loudenburg, R., Drube, G. and Leonardson, G. (2010) South Dakota 24/7 Sobriety Program Evaluation Findings Report Salem: Mountain Plains Evaluation Maguire, M. (2004) ‘The Crime Reduction Programme in England and Wales: reflections on the vision and the reality’ Criminal Justice 4: 213-237 McSweeney, T. (2015) ‘Calling Time on Alcohol-Related Crime? Examining the Impact of Court-Mandated Alcohol Treatment on Offending Using Propensity Score Matching’ Criminology and Criminal Justice 15 (4): 464-484 McSweeney, T., Gyateng, T., Moretti, A., Skrine, O., Turnbull, P. and Joliffe, D. (2013) An Evaluation of the Sentence Planning Approaches (SPA) Pilots London: ICPR and the University of Greenwich McSweeney, T., Webster, R., Turnbull, P. and Duffy, M. (2009) Evidence Based Practice? The National Probation Service’s Work with Alcohol Misusing Offenders London: Ministry of Justice Ministry of Justice (2015a) Prison and Probation Performance Statistics 2014 to 2015 London: Ministry of Justice Ministry of Justice (2015b) Criminal Justice System Statistics Quarterly: December 2014 London: Ministry of Justice 50 Ministry of Justice (2012) Proven Re-offending Statistics: Definitions and Measurement London: Ministry of Justice Ministry of Justice (July 2010) Conviction Histories of Offenders between the Ages of 10 and 52, England and Wales Ministry of Justice Statistics Bulletin London: Ministry of Justice Northamptonshire Police (8 December 2014) Northamptonshire Sobriety Bracelets Pilot Project: Update Northamptonshire: Northamptonshire Police Office for National Statistics (2015) Focus on Violent Crime and Sexual Offences 2013/14 Newport: ONS Public Health England (2014a) Alcohol Treatment in England 2013-14 London: Public Health England Public Health England (2014b) Local Alcohol Profiles for England Liverpool: Public Health England Raistrick, D., Heather, N. and Godfrey, C. (2006) Review of the effectiveness of treatment for alcohol problems London: National Treatment Agency for Substance Misuse 51 City Hall, More London, The Queen’s Walk, London SE1 2AA 020 7983 6532 enquiries@mopac.london.gov.uk www.london.gov.uk/MOPAC www.twitter.co.uk/MOPACLdn Alcohol Abstinence Monitoring Requirement Final Impact Evaluation December 2020 MOPAC Evidence and Insight Adele Harrison, Julia Yesberg, Michael Keenan Red Quadrant Tim McSweeney, Russell Webster Month and Year Authors MOPAC Evidence and Insight 1. Executive Summary In 2011, the Mayor’s Office for Policing And Crime (MOPAC) secured legislation to allow for the introduction of a new sentencing power – the Alcohol Abstinence Monitoring Requirement (AAMR) – to tackle the problem of alcohol related offending in London. The innovation has been constantly underpinned by evaluation, resulting in multiple publications encompassing performance, process and indicative impact of the original pilot scheme. The current report concludes the evaluation and presents final performance and impact findings. In terms of performance, over the entirety of the pan-London AAMR project (April 2016 to June 2018), a total of 1,173 AAMRs were issued by the London courts. Offenders were subject to the tag for an average of 64 days. Of the AAMR cases where breach information was available (n=1,014), there were only 60 cases where the offender had been returned to court and been found or pled guilty to the breach.1 This indicates an overall 94% compliance rate: a figure that has remained stable throughout the entirety of the project. In total, offenders were monitored for 71,584 days and were sober for 69,996 of these days. These figures indicate that in 98% of the days offenders were monitored, they did not consume alcohol. This is itself can be considered a considerable success of the innovation. To explore final impact upon offending, a statistical technique known as Propensity Score Matching was used to develop the counterfactual (i.e., what would have happened in the absence of the AAMR) which focussed upon examining reoffending within different time periods through the measures of both arrest and charges. As a result of this procedure, a total of 431 AAMR individuals were successfully matched to a control individual. There were no significant differences on any matching variable, confirming the success of the procedure. Overall, looking across both arrests and charge data, we see that the results mirror each other in that the level of reoffending for offenders on the AAMR is not significantly different to those on similar orders six months or 12 months post sentence commencement. However, they do commit more reoffences than the comparison group across both entire timeframes. One complication in this analysis is the markedly different lengths of the orders – AAMR offenders were subject to the tag for an average of 64 days (range 4 days - 121 days) whilst the control cohort spent on average 445 days on the order (range 180 days -1095 days). So essentially, the AAMR cohort had more time off the order to reoffend than the control group across both timeframes, and it is important that this is taken into account when interpreting these results. Looking closer at arrest, results indicate: • In terms of arrest data in the first six months of supervision, overall, there was no significant difference between exposure to the AAMR and the level of re-offending committed in the six months post-commencement of supervision (24.2% vs. 25.6%) χ²(1, N=718)=.19, p=.666; 1 For the purpose of measuring compliance, we have recorded an unsuccessful completion when alerts about violations on the tag led to enforcement action being taken by the Offender Manager that led to a breach conviction at Court. RR=1.057; CI=.82-1.36) or 12 months post commencement of supervision (35.5% vs. 40.1%; χ²(1, N=704) =1.55, p=.214; RR=1.12; CI=0.93-1.36). • During the six-month post commencement, the AAMR group had a significantly higher average number of reoffences compared to the comparison (a mean of 2.9 versus 1.45; p=.005). This was also observed during 12-month post commencement (a mean of 3.72 versus 1.83; p=.000). • The comparison group on average reoffended quicker than the AAMR group during the first six-months post commencement (73.5 days versus 85.7 days) although this was not a significant difference (p=.099). This pattern was reversed during 12-months post commencement, with the AAMR group reoffending sooner (140.88 days vs 145.9 days) but again the results were not significant (p=.691). • Overall, 35.5% of AAMR offenders offended (arrest) within 12 months since order commencement. However, only a minority committed their offence on the tag (4.8%, n=17). The clear majority of offenders committed their reoffences after the tag had been removed, but prior to sentence end (27%, n=95). There are statistically significant differences between the proportions of individuals being arrested in the following periods: on tag vs. remainder of sentence: (t=-8.61, p=.000, df=350, STD=.470), on tag vs. after sentence: (t=-7.08, p=.000, df=350, STD=.444). Therefore, we can say that offenders are significantly less likely to reoffend whilst on the tag and this is likely a deterrent effect for the period that the tag is worn. Key findings from the analysis of charge data indicate: • There was no significant difference between exposure to the AAMR and the proportion to re- offend (as measured by an MPS charge for a further offence) committed in the six-months post-commencement of supervision (13.9% vs. 11.1%; p=.217) or 12 months post commencement of supervision (21.6% vs. 18.9%; p=.364). • During the six-month post commencement, the AAMR group had a significantly higher average number of reoffences compared to the comparison (a mean of 2.7 versus 1.6; p=.005). This was also observed during 12-month post commencement (a mean of 3.1 versus 1.9; p=.001). • The comparison group on average reoffended quicker than the AAMR group during the first six-months post commencement (77.6 days versus 84.7 days). This was not a significant difference (p=.458). This pattern was reversed during 12-months post commencement, with the AAMR group reoffending sooner (139.5 days vs 152.5 days) but again the results were not significant (p=.417). • A total of 21.6% (N=81) of AAMR offenders offended (charge) in the 12 months since order commencement. However, within this, only a minority (4% of the total cohort, N=15) committed their offence during the AAMR, committing 27 offences between them. This means that 16.8% of the total cohort (N=63) offended after the tag had been removed and committed a total of 201 offences between them after this date. The average speed to first reoffence for these 63 offenders was 132.6 days, just over twice the time offenders were subject to the tag (64 days, range 4-121 days). • Analysis was conducted to attempt to predict the factors that were predicting overall offending. The factor with the largest effect on risk of reoffending was age: each year of increasing age exerted a protective effect, reducing the risk of reoffending by around three per cent. The other factor associated with risk of reoffending was order type; compared to those serving a Suspended Sentence Order, offenders serving a Community Order had a 32 per cent reduced risk of recidivism. Overall results indicate that there is little difference between the AAMR cohort and the comparison group. These echo similar findings to the earlier published pilot evaluation. Given that implementing innovation can bring risks, that the AAMR is broadly achieving the same results as similar orders over a shorter timeframe, should be considered in a positive light, especially when viewed with the previous process learning that staff and judiciary perceptions were, on the whole positive towards this order as it gives them different options and another tool in the toolbox to prevent offending. However, the cost benefit analysis does indicate that the costs associated with the AAMR do outweigh the costs of placing a similar cohort of offenders on a community sentence involving probation supervision by £257,314. In other words, the net cost to society of sentencing an individual offender on a AAMR is £2,573 more than sentencing the same individual to a community sentence involving probation supervision. These results should be treated as indicative as there were some limitations to the cost benefit analysis, which are discussed in detail within the report. Overall, the evaluation of AAMR has spanned 4 reports2 commencing with the process exploration of the initial pilot reporting in 2016, through annual reports, to the London wide rollout and examination of impact to this final impact study. This method of evidence-based working has enabled MOPAC and programme staff to base decisions over this time upon data and learning which has been a positive approach to implementing innovation and a manner of working that should be more frequent within the Criminal justice Service. 2 Pepper, M. & Dawson, P. (2016) The Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot, MOPAC. 2 Hobson, Z., Harrison, A. & Dawson, P. (2017) Alcohol Abstinence Monitoring Requirement – South London Pilot Indicative Impact Report,. MOPAC 2 Hobson Z,. Harrison A. & Duckworth, L. (2019) Alcohol Abstinence Monitoring Requirement – A review of Process and Performance from year 2. MOPAC 2. Introduction The Alcohol Abstinence Monitoring Requirement In 2011 the Mayor’s Office for Policing And Crime (MOPAC) secured legislation to allow for the introduction of a new sentencing power – the Alcohol Abstinence Monitoring Requirement (AAMR) – to tackle the significant problem of alcohol related offending in London. The AAMR gives the judiciary the statutory power to impose compulsory sobriety where an offence is alcohol related. The AAMR involves fitting a tag to the offender’s ankle and monitoring their alcohol consumption for up to 120 days. Following the positive learning from the initial AAMR proof of concept pilot3, a two-year pan London roll out of the AAMR was launched in April 2016, joint funded by MOPAC and the Ministry of Justice. The aim of this expansion was to test the impact of the AAMR, in line with the Conservative government’s 2015 manifesto commitment to make sobriety tags available across England and Wales. The pan London AAMR programme’s main delivery period ended at the end of March 2018, with a further three-month extension period until June 2018 to enable a managed closedown of the programme. The innovation has been routinely supported by evaluation exploring the performance, process and indicative impact of the tag (Hobson et al, 20174, Hobson et al, 20195) - refer to these reports for broader performance and process learning. Initial indicative analysis on the pilot scheme was published (Hobson, Harrison & Dawson, 20176) reporting no significant differences comparing the AAMR cohort and a matched comparison across reoffending. Following these results the technology has been rolled out nationally. current report provides a final performance update for the pan-London AAMR programme and presents findings assessing the impact of AAMR on offending. 3. Methodology To evaluate whether the AAMR has had an impact on reoffending, the AAMR cohort was matched using a statistical technique known as Propensity Score Matching to a similar group of offenders who had not received AAMR. This is well established and robust method of generating a counterfactual (Rosenbaum and Rubin, 1983). This section outlines the process behind this approach. AAMR Sample The overall AAMR cohort consisted of 1,165 individuals. However, to be included in the impact analysis, several exclusions had to be made. First, all cases sentenced after March 2018 were excluded (n=159) as no data was provided from probation for these cases during the closedown period (April- June 2018). Second, all NPS cases were excluded (n=150) because data for the control sample was only available for CRC cases (see next section). All cases with missing data (e.g., OGRS, age, ethnicity) were also excluded (n=126). Cases where a PNCID number was shared by more than one individual 3 Pepper, M. & Dawson, P. (2016) The Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot, MOPAC. 4 Hobson, Z., Harrison, A. & Dawson, P. (2017) Alcohol Abstinence Monitoring Requirement – South London Pilot Indicative Impact Report,. MOPAC 5 Hobson Z,. Harrison A. & Duckworth, L. (2019) Alcohol Abstinence Monitoring Requirement – A review of Process and Performance from year 2. MOPAC 6 Hobson, Z., Harrison, A. & Dawson, P. (2017) Alcohol Abstinence Monitoring Requirement – South London Pilot Indicative Impact Report,. MOPAC (n=39), or where a PNCID number was missing (n=7) were excluded. Finally, all cases where the AAMR tag was not successfully fitted were excluded (n=73), leading to an eligible AAMR sample of 611 tagged individuals. Control Sample To identify a control group, caseload data was obtained from the London CRC in March 2019. Offenders who were on a Community Order (CO) or Suspended Sentence Order (SSO) were initially selected (n=12,990). This sample was further reduced by selecting only those offenders with a moderate alcohol need, defined as: a score of ‘1’ on the overall OASys Alcohol misuse item (indicating alcohol is linked to offending) and a score of ‘1’ on the OASys Alcohol misuse sub-item 9.1 (indicating alcohol is a moderate problem). Based on these criteria, 1,149 offenders were identified. Cases were removed if the sentence had been imposed prior to April 2018 as MOPAC was not provided with any data for the AAMR cases during the closedown period (n=297), where relevant information to facilitate matching was missing (e.g., ethnicity, PNCID number; n=113) and for those who had received an AAMR (n=37). The final sample for matching consisted of 702 offenders managed by the CRC on a CO or SSO who had a moderate alcohol problem. Generating the counterfactual The first step of propensity score matching technique was to develop a regression model that is able to predict group membership (i.e., the propensity of receiving the AAMR treatment). Five theoretically relevant matching variables were selected and entered in a logistic regression model predicting group membership. The variables were: (1) OGRS score, (2) Age, (3) Gender, (4) Ethnicity, (5) Index Offence. The resulting model was significant, meaning that the procedure can continue. The approach then seeks to match pairs (i.e., control and AAMR) of individuals on their propensity score of receiving the treatment. To do so, an approach called nearest neighbour optimal matching method was used (Austin, 2011). A caliper was set for the propensity scores, meaning a control individual could only be matched to an AAMR individual if their propensity score fell within 0.2 of the standard deviation of the logit of the propensity score. The optimal matching method ensured the best-matched pairs were chosen. The matching procedure was conducted without replacement, meaning once a control individual was matched to an AAMR individual, they could not be matched again. As a result of this procedure, a total of 431 AAMR individuals were successfully matched to a control individual using CRIS charge data. A suitable match could not be found for 180 AAMR individuals, primarily due to an over-representation of driving-related index offences, a younger age profile, and lower OGRS scores in the AAMR cohort. Table 1 shows the match between the AAMR and control sample was strong. There were no significant differences on any of the matching variables, confirming the success of the PSM procedure. Table 1. AAMR and Control Cohorts Matching Criteria (N=862). Match based on the first approach taken using CRIS outcome data Matching Criteria AAMR N=431 Control N=431 Equivalence tests (t-tests or chi-square) OGRS 2-year score Average = 42.4 (SD = 23.8) Average = 41.1 (SD = 24.0) t(860) = 0.85, p = .396 Age (average) Average = 34.5 (SD = 10.2) Average = 35.0 (SD = 9.9) t(860) = -0.69, p = .488 Gender X2(1) = .039, p = .844 Male 86.3% (n=372) 85.8% (n=370) Ethnicity X2(4) = 3.34, p = .502 White 63.3% (n=273) 63.3% (n=273) Black 16.0% (n=69) 13.9% (n=60) Asian 11.8% (n=51) 13.0% (n=56) Mixed 4.4% (n=19) 6.5% (n=28) Other 4.4% (n=19) 3.2% (n=14) Index Offence Type X2(6) = .901, p = .989 Violence 53.4% (n=230) 53.4% (n=230) Driving 18.1% (n=78) 19.0% (n=82) Damage/Theft 13.9% (n=60) 13.7% (n=59) Harassment 9.0% (n=39) 9.5% (n=41) Public Order 3.5% (n=15) 3.0% (n=13) Breach 1.4% (n=6) 0.9% (n=4) Drugs 0.7% (n=3) 0.5% (n=2) Outcome measures and data matching The next stage was to collect the key outcomes measures (i.e., charge) from the MPS systems to compare the treatment and control. Charge data was adopted opposed to proven reoffending7 due to practicalities and timeliness – especially given the control sample were from a later period after the AAMR innovation ended, a necessary aspect given the AAMR was a pan-London initiative. Given this, the focus on the evaluation was charge data which covers a narrow timeframe, although research has discussed the broad equivalence between using charge data and proven reoffending (Ringland, 2013). To explore the impact of the AAMR, the evaluation sought to look at offending within a number of time frames post sentence commencement: 1) 6 months, 2) 12 months and 3) offending on and off the tag. However, upon collecting the police charge data difficulties were experienced in returning successful matches from CRIS. Successful matches were observed for CRIS data relating to 72% for the AAMR cohort and 67% for the comparison group. In order to improve the number of successful matches a variety of methods were employed – for example CRC quality assured all the PNC IDs, extraction dates for analysis were also widened in order to try and capture more trigger offences - the time period for data extraction from the CRIS system was widened to October 2015 to December 2019 for the AAMR cohort, and October 2017 to December 2019 for the control cohort. We then used fuzzy matching approaches to extract CRIS data, but this did not affect the number of successful matches returned from CRIS, the sample retrieved was still only 73% from the AAMR cohort and 69% from the Control cohort. We then explored the possibility of using arrest data from custody systems as there is no requirement to record drunk and disorderly information on CRIS and it was felt that this could affect the amount of successful matches being returned. MPS custody systems capture all offences including drunk and disorderly and approximately 17% (AAMR) and 19% (Control) of the offences within the arrest data returned fit this category. Fuzzy matching was also undertaken to improve the number of successful matches. Successful matches were observed for Custody data relating to 84% (n=402) of the AMMR cohort and 88% (n=422) for the control group. The percentage of successful matches has increased using arrest data - however, we can’t be sure that an outcome was generated for the arrest. Table 2 shows the breakdown of AMMR and Control cohort by outcome from the arrest data. Up to 29% generate no outcome information at all. This is higher for the AAMR group. 7 A proven re-offence is defined as any offence committed in a one year follow-up period and receiving a court conviction, caution, reprimand or warning in the one year follow up or a further six months waiting period. Table 2 – Outcomes from arrest – MPS custody data set AAMR Control No Further Action 17% 23% Charged and Bailed To Court 16% 17% Charged and Detained For Court 13% 17% Bailed To Police Station 8% 8% Detained For Court On Warrant 4% 4% Released Under Investigation 6% 3% Simple Caution 1% 2% To Court Breach of Bail 1% 2% Fixed Penalty Notice Issue 1% 2% Detained for Court - No Charge 0% 1% To Prison 0% 1% Processed on behalf of another force 0% 1% Cautioned 0% 1% To Court (Postal Requisitions) 1% 1% Bailed To Court On Warrant 0% 0% Other outcome 0% 1% Missing/Blank - Not Known 29% 16% The primary limitation on the comparison of the AAMR cohort with the control group is that the control group were formed at a later date than the AMMR group, which was by design to identify a control group given the AAMR was rolled out pan London. However, this has placed several constraints on the control group that should be acknowledged: 1. Sentence end date – a large portion of the control group have either not finished (4%, n=18) their sentence or had a more limited timeframe (less than 12 months) after their sentence date (15%, n=62), reducing the number of cases that could be included in the analysis. 2. COVID-19 lockdown period – this period of time saw a very unnatural crime landscape where the volume of offences reduced drastically. Several of the control group sentences and/or post sentence periods have run into the lockdown period (taken as 23/03/2020 onwards). 181 cases were removed from the analysis due to this reason. In order to conduct a comparison between the AAMR cohort and control group both individuals in a matched pair needed to a) both have data for the particular period being looked at; b) the period must be complete (i.e., for a time period this must not end after the data download date and for a sanction period this must be complete); and c) the period must not run into the COVID-19 lockdown period. This has produced a base size for the arrest analysis which is 359. Arrest extraction dates for analysis were widened in order to try and capture more trigger offences - the time period for data extraction from custody system was October 2015 to September 2020 for the AAMR cohort and October 2017 to September 2020 for the control cohort. Any matched pair who both returned data for the period, and who both held at least 6 months data post sentence that was unaffected by the COVID-19 lockdown were included in the analysis. This equated to 359 matched pairs. In terms of the CRIS charge data, all charges that occurred within 12-months of an offender’s sentence commencement date was recorded as a re-offence. Thirteen percent of the assembled matched comparison group (n=56) had not been supervised for a 12-month period at the point CRIS data were extracted in December 2019. These cases and their corresponding AAMR matches were therefore excluded from our analyses of offending outcomes at 12 months (N=375 matched pairs). Given the issues we have encountered with data quality and MPS recording practices, plus impact of Covid-19 on the control cohort data extraction period, combining the above approaches provides the best methodology in understanding the impact of AAMR upon reoffending. The report therefore focusses on findings from both approaches. Table 1a. AAMR and Control Cohorts Matching Criteria (N=718) based on the second approach using Arrest data (18 months prior thru to 6 months post sentence) Matching Criteria AAMR Control Equivalence tests (t- tests or chi-square) N=359 N=359 OGRS 2-year score Average = 45.14 Average = 44.50 t(716) = -0.368, p = .375 (SD = 23.03) (SD = 23.57) Age Average = 34.36 Average = 34.66 t(716) = 0.406, p = .299 (SD = 10.27) (SD = 9.75) Gender X2(1) = .050, p = .823 Male 86.9% (n=312) 87.5% (n=314) Ethnicity X2(5) = 4.010, p = .548 White British 42.1% (n=151) 46.2% (n=166) White other 19.5% (n=70) 16.2% (n=58) Black 17.5% (n=63) 15.3% (n=55) Asian 12.3% (n=44) 12.0% (n=43) Mixed 5.0% (n=18) 7.2% (n=26) Other 3.6% (n=13) 3.1% (n=11) Index Offence Type X2(6) = 1.575, p = .954 Violence 50.4% (n=181) 52.1% (n=187) Driving 17.8% (n=64) 18.1% (n=65) Damage/Theft 15.6% (n=56) 15.3% (n=55) Harassment 10% (n=36) 10.3% (n=37) Public Order 3.3% (n=12) 2.5% (n=9) Breach 1.9% (n=7) 1.1% (n=4) Drugs 0.8% (n=3) 0.6% (n=2) 4. Performance Results (April 2016-June 2018) Insights on the AAMR Tag and performance Over the entirety of the pan-London AAMR project (April 2016 to June 2018), a total of 1,173 AAMRs were issued by the London courts (91% by Magistrates’ Courts, n=1,070). Overall, Croydon Magistrates’ Court was the most active in imposing the AAMR (15%, n=178), followed by Highbury Corner (12%, n=141), Camberwell Green (9%, n=111) and Uxbridge (9%, n=106) Magistrates’ Courts. The majority of AAMRs were imposed for Violence (45%, n=456) or Driving Offences (29%, n=291), followed by Damage/Theft (10%, n=106) and Harassment (9%, n=88).8 AAMRs were usually imposed as a requirement of a Community Order (73%, n=851) rather than a Suspended Sentence Order (27%, n=322). Most orders received multiple requirements (69%, n=702), most commonly a Rehabilitation Activity Requirement (n=350), Unpaid Work (n=-340), Prohibited Activity (n=197), Accredited Programme (n=89), and Curfew (n=54).9 In 86% of cases, EMS received the notification from the Courts on the same day as sentencing (n=834), or the following day (n=172). However, only 43% (n=436/1006) of offenders (where EMS was notified on the same or following day) were tagged on the same day or within one day of notification. On average, offenders were tagged within four days of EMS receiving the notification. In 10% of cases the tag was not fitted (n=118). This could be due to a variety of reasons for example they had lied about their drinking and were actually not suitable, they absconded or refused to be tagged or they immediately changed address, became homeless or moved out of London and hence the sentence became unworkable. For those that were fitted, fitting was done on either the first (69%, n=724) or second attempt (20%, n=210). Offenders were subject to the tag for an average of 64 days (range 4 days - 121 days). The control cohort spent on average 445 days on the order (range 180 days -1095 days). Of the AAMR cases where breach information was available (n=1,014), there were only 60 cases where the offender had been returned to court and been found or pled guilty to the breach.10 This indicates an overall 94% compliance rate with the AAMR: a figure that has remained stable throughout the entirety of the project. In total, offenders were monitored for 71,584 days and were sober for 69,996 of these days. These figures indicate that in 98% of the days offenders were monitored, they did not consume alcohol. This is itself can be considered a considerable success for the innovation. Tag Wearer Demographics The majority of offenders who were sentenced to an AAMR were male (87%, n=1,018), with an average age of 33 years old (ranging from 18–73 years). Three-fifths of offenders (60%, n=698) were aged between 18 and 34 years at the time of sentencing.11 Over half of offenders were white (63%, 8 Offence type was missing for 159 cases. 9 Additional requirements were missing for 159 cases. 10 For the purpose of measuring compliance we have recorded an unsuccessful completion when alerts about violations on the tag led to enforcement action being taken by the Offender Manager that led to a breach conviction at Court. 11 Age was missing for 4 cases. n=555), 17% were black (n=149), 12% were Asian (n=105) and the remainder were Mixed or Other ethnic group (n=66).12 The NPS and CRC consider the risk an offender poses in two ways – risk of serious harm and risk of reoffending. Nearly two thirds of offenders were considered to be a medium risk of serious harm (63%, n=610), with just 1% (n=14) being considered a high risk of serious harm.13 The average OGRS 2-year14 score was 36% (ranging from 0 to 97%), indicating that just over one third of offenders would be predicted to reoffend within two years.15 This figure shows that those receiving the AAMR are broadly identified as having a low risk of reoffending. This has remained a constant throughout the duration of the AAMR evaluation. 5. Impact Analysis Results To explore impact of the AAMR, the evaluation focussed upon offending behaviours across a variety of timeframes (6 months post commencement of supervision, 12 months post commencement of supervision, offending on and off the tag, predicting the drivers of reoffending and an economic analysis). This section works through these measures to seek to understand the impact, if any of the AAMR. Offending in the first six months post commencement of supervision The first timeframe explores offending (as measured by both arrest and charge data) within the first six months post commencement of the order. In terms of arrest data in the first six months of supervision, overall, there was no significant difference between exposure to the AAMR and the level of re-offending committed in the six months post- commencement of supervision (24.2% vs. 25.6%) χ²(1, N=718)=.19, p=.666; RR=1.057; CI=.82-1.36). See table 2a. There had been a total of 388 reoffences leading to an MPS arrest in the six months post- commencement of supervision, with an average (mean) of 2.2 offences (Mdn=1.5, SD=1.9, R=1-14, n= 179). There is a significant difference in arrests between the AAMR group and the control group six months post sentence commencement, with the AMMR group being subject to a higher number of arrests (253 vs 135). The number of reoffences was significantly higher for the AAMR group (M=2.9, Mdn=2.0, SD=2.42, R=1-14, n=87) relative to the comparison group (M=1.45, Mdn=1.0, SD=0.75, R=1- 5, n=92) (t=-3.30, p=.005). 12 Ethnicity was missing for 298 cases. 13 Risk of harm was missing for 200 cases. 14 OGRS is a predictor of re-offending based only on static risks – age, gender and criminal history. It allows probation, prison and youth justice staff to produce predictions for individual offenders even when the use of dynamic risk assessment tools (e.g. The Offender Assessment System (OASys) or Asset) is not possible. 15 OGRS Year 2 score was missing for 209 cases. The overall average (mean) time to first re-offence was 79.53 days (Mdn=75.5, SD=48.98, R=1-183, n=92), with the comparison group reoffending sooner (M=73.5, Mdn=66.0, SD=47.91, n=135) than the AAMR group (M=85.7, Mdn=88.0, SD=49.57, n=87), but not significantly so (t=-1.66, p=.099). Table 2a: Offending in the six months post-commencement of supervision (arrest) AAMR Cohort (N=359) Control Cohort (N=359) Arrests Arrests % (number) of offenders who reoffended 24.2% (87) 25.6% (92) Total number of reoffences 253* 135 Minimum number of reoffences 1 1 Maximum number of reoffences 14 5 Average reoffences 2.90* 1.45 Average speed to first reoffence (days) 85.7 73.5 Range (days) 1-178 6-183 Turning to charge data for an offence in the first six months subsequent to the AAMR. Overall, there was no significant difference between exposure to the AAMR and the level of re-offending committed in the six months post-commencement of supervision (13.9% vs. 11.1%; χ²(1, N=862)=0.15, p=.217; RR=1.29; CI=0.86-1.94). See table 2b. There was a total of 238 reoffences leading to an MPS charge in the six months post-commencement of supervision, with an average (mean) of 2.2 offences (Mdn=1.0, SD=2.0, R=1–14, n= 108). The number of reoffences was significantly higher for the AAMR group (M=2.7, Mdn=2, SD=2.4, R=1-14, n=60) relative to the comparison group (M=1.6, Mdn=1.0, SD=1.2, R=1-8, n=48) (t=-2.89, p=.00). The average (mean) time to first re-offence was 81.6 days (Mdn=80.0, SD=49.0, R=0-180, n=108), with the comparison group reoffending sooner (M=77.6, Mdn=71.0, SD=49.8, n=48) than the AAMR group (M=84.7, Mdn=89.5, SD=48.6, n=60), but not significantly so (t(106)=-.745, p=.458). Table 2b: Offending in the six months post-commencement of supervision (charge) AAMR Cohort (N=431) Control Cohort (N=431) Charges Charges % (number) of offenders who reoffended 13.9% (60) 11.1% (48) Total number of reoffences 160* 78 Minimum number of reoffences 1 1 Maximum number of reoffences 14 8 Average reoffences per offender 2.7* 1.6 Average speed to first reoffence (days) 84.7 77.6 Range (days) 0-180 5-175 There were no significant difference between the AAMR (84.7 days) and the comparison group (77.6 days) in the average time to first reoffence within six months. Offending in the 12 months post-commencement The next timeframe explores a longer follow on period, that of offending (again as measured by both arrest and charge data) within the first 12 months post commencement of the order. There was no association between exposure to the AAMR and the level of re-offending (as measured by arrest for a further offence) committed in the 12 months post-commencement of supervision (35.5% vs. 40.1%; χ²(1, N=704) =1.55, p=.214; RR=1.12; CI=0.93-1.36). There had been a total of 723 reoffences leading to an MPS arrest in the 12 months post- commencement of supervision, with an average (mean) of 2.72 offences (Mdn=2.00, SD=2.65, R=1- 22, n= 266). There is a significant difference in arrests between the AAMR group and the control group 12 months post sentence commencement, with the AMMR group being subject to a higher number of arrests (465 vs 258). The number of reoffences was significantly higher for the AAMR group (M=3.72, Mdn=3.00, SD=3.33, R=1-22, n=125) relative to the comparison group (M=1.83, Mdn=1.00, SD=1.33 R=1-7, n=141) (t=-6.025, p=.000). The average (mean) time to first re-offence was 143.53 days (Mdn=121.00, SD=102.37, R=1-364, n=266), with the AAMR group reoffending sooner (M=140.88, Mdn=120.00, SD=97.70, n=125) than the comparison group (M=145.92, Mdn=123.00, SD=106.73, n=141), but not significantly so (t=.397, p=.691). Table 3a: Offending in the 12 months post-commencement of supervision (arrest) AAMR Cohort (N=352) Control Cohort (N=352) Arrests Arrests % (number) of offenders who reoffended 35.5% (125) 40.1% (141) Total number of reoffences 465* 258 Minimum number of reoffences 1 0 Maximum number of reoffences 22 7 Average reoffences per offender 3.72* 1.83 Average speed to first reoffence (days) 140.88 145.92 Range (days) 1-355 6-364 In terms of charge data over the longer 12 month period, again, there were no significant difference between exposure to the AAMR and the level of re-offending committed in the 12 months post- commencement of supervision (21.6% vs. 18.9%; χ²(1, N=750)16=0.82, p=.364; RR=1.18; CI=0.82-1.68). There was a total of 391 reoffences leading to an MPS charge in the 12 months post-commencement of supervision, with an average (mean) of 2.6 offences (Mdn=2.0, SD=2.4, R=1–16, n= 152). The number of reoffences was significantly higher for the AAMR group (M=3.1, Mdn=2, SD=2.6, R=1-14, n=81) relative to the comparison group (M=1.9, Mdn=1.0, SD=2.0, R=1-16, n=71) (t(147)=3.2, p=.001). The average (mean) time to first re-offence was 145.6 days (Mdn=122.5, SD=98.1, R=0-361, n=152), with the AAMR group reoffending sooner (M=139.5, Mdn=116.0, SD=98.1, n=81) than the comparison group (M=152.5, Mdn=145.0, SD=98.3, n=71), but not significantly so (t(150)= -.813, p=.417). Table 3b: Offending in the 12 months post-commencement of supervision (charge) AAMR Cohort (N=375) Control Cohort (N=375) Charges Charges % (number) of offenders who reoffended 21.6% (81) 18.9% (71) Total number of reoffences 255 136 Minimum number of reoffences 1 1 Maximum number of reoffences 14 16 Average reoffences per offender 3.1 * 1.9 Average speed to first reoffence (days) 139.5 152.5 Range (days) 0-361 5-346 16 One AAMR case which lacked details on supervision start date was excluded from all subsequent outcome analyses. Which factors were predictive of reoffending in the 12 months post-commencement of supervision? The evaluation considered nine key variables in our exploratory analyses of factors associated with the likelihood of overall reoffending (across both treatment and control) at 12 months: age; gender; ethnicity, main offence; whether exposed to an AAMR; number of different requirements imposed; and type of order17. These factors were chosen based on previous research by the Ministry of Justice, which identified that there are key static factors which can influence reoffending1819. On the basis of the results from the analysis, the following variables were excluded from multivariate modelling, as they were not associated with offending: • gender (reference was ‘male’) (χ²= .530, p=.467), • whether exposed to an AAMR (reference was ‘yes’) (χ²= .922, p=.337), • ethnicity (reference was ‘White’) (χ²= 1.034, p=.309); and • main offence (reference was ‘violence’) (χ²= 1.256, p=.262). Only the three factors associated with a reoffence leading to a charge within 12 months of commencing supervision were included in the final multivariate model. These were (in order of entry based on p-values and Wald statistic): age (χ²= 9.8, p=.002); type of order (reference was ‘Suspended Sentence Order’) (χ²= 6.215, p=.013); and number of requirements (χ²= 3.96, p=.047). The final model identified two of the three factors as being significantly predictive of a reoffence. In this model, adjusting for covariates, the factor with the largest effect on risk of reoffending was age. Each year of increasing age exerted a protective effect, reducing the risk of reoffending by around three per cent. Compared to those serving a Suspended Sentence Order, offenders serving a Community Order had a 32 per cent reduced risk of recidivism. The results of the model are set out in Table 4. 17 For categorical variables the model value served as the reference category. We used p=.25 as a cut-off for inclusion in multivariate analysis. 18 Transforming rehabilitation: A summary of evidence on Reducing Reoffending, MoJ Analytical Series, 2013. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/243718/ evidence-reduce-reoffending.pdf 19 The factors associated with proven reoffending following release from prison, MoJ Analytical Series 2013. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/491119/ re-offending-release-waves-1-3-spcr-findings.pdf Table 4: Results of a Cox proportional hazards model of reoffending in the 12 months post- commencement of supervision (N=750) Covariate Hazard ratio (HR) (95% CI) Adjusted HR (95% CI) χ² p Age .973 (.956 - .990) .974 (.958 - .991) 8.60 .003 Gender [reference was 'male'] .836 (.517 – 1.353) - Ethnicity [reference was White] 1.072 (.937 – 1.226) - Main offence [reference was ‘violence’] .957 (.887 – 1.033) - Whether exposed to an AAMR [reference was 'no'] 1.169 (.850 – 1.608) - Number of requirements imposed 1.055 (1.001 – 1.111) 1.041 (.987 – 1.098) 2.16 .141 Type of order [reference was 'Suspended Sentence Order'] .656 (.472 - .914) .680 (.488 - .949) 5.15 .023 Offending on and off the AAMR tag So far, the analysis has explored offending within set timeframes (be it 6 or 12 months) reporting that there was no overall difference in proportion to offend across these periods, but higher levels of offences conducted by the AAMR group. One complication in this analysis is the markedly different lengths of the orders – offenders were subject to the tag for an average of 64 days (range 4 days - 121 days) whilst the control cohort spent on average 445 days on the order (range 180 days -1095 days). So essentially, in any period under study - the AAMR will be more likely to have completed compared to the control. With this in mind, the analysis now turns to exploring whether the AAMR offending occurred during or after the tag was removed. For the AAMR group, the time spent on the tag was usually a relatively short period at the start of their sentence (between 2 weeks to 2 months). Overall, 35.5% of AAMR offenders offended (arrest) within 12 months since order commencement, however, only a minority committed their offence on the tag itself (4.8%, n=17). The clear majority of offenders committed their reoffences after the tag had been removed, but prior to sentence end (27%, n=95). There are statistically significant differences between the proportions of individuals being arrested in the following periods: On tag vs. Remainder of sentence: (t=-8.61, p=.000, df=350, STD=.470), On tag vs. After sentence: (t=-7.08, p=.000, df=350, STD=.444). There is no statistically significant difference between the remainder of sentence period and post sentence period. Those with a low OGRS score (30 or under) are significantly more likely to have 0 arrests in the remainder of their sentence (86%, n=91) compared to medium (score 31-60, 71%, n=107) or high (score 61 or over, 62%, n=59) OGRS scoring individuals (X2(2) = 14.879, p = .001). Similarly, those whose index offence was driving are significantly more likely to have 0 arrests for the remainder of their sentence (91%, n=58) compared to all other offences (69%, n=199) (X2(6) = 12.853, p = .045). Table 5: AAMR cohort arrests whilst on tag and rest of sentence Period on tag (N=35120) Period off tag on sentence (N=351) Period 6 months after sentence end (N=351) Arrests Arrests Arrests % (number) of offenders who reoffended 4.8% (17) 27.0% (95) 21.9% (77) Total number of reoffences 52* 324 141 Minimum number of reoffences 1 1 1 Maximum number of reoffences 13 14 12 Average reoffences 3.05 9.26 1.83 20 N is reduced to 351 (instead of 359) due to missing tag dates (either fitted or removed). Returning to the charge data a similar pattern can be seen. Overall, we know that 21.6 % (N=81) of offenders on the AAMR offended (charged) in 12 months since order commencement out of a total of 375. Within this, only a minority (4% of the total cohort, N=15) committed their offence during the AAMR, committing 27 offences between them. This means that 16.8% of the total cohort (N=63) offended after the tag had been removed and committed a total of 201 offences between them (3 offenders had no tag removal date and therefore were excluded from the analysis). The average speed to first reoffence for these 63 offenders was 132.6 days (range = 3-309). This is just over double the average time the offenders were subject to the tag (64 days, range = 4-121 days). Unfortunately, given the longer timeframes of the order received by the control group, it is not possible to explore an equivalent offending ‘off’ the order for this group. The nearest comparison to draw upon would be MoJ reoffending statistics which show that 28.5% of adult offenders reoffended during April 2017 to March 2018, which is significantly higher than our AAMR cohort measured via arrest or charge. However, we have been able to further explore the proportion of offenders arrested on a monthly basis for both the AAMR and Control. Figure 1 below compares the monthly average arrests across the AAMR cohort and the Control cohort, clearly showing a dip in arrests during the tagging period for the AAMR cohort compared to the ‘on sentence’ period for the control cohort.21 Figure 1: proportion of sample with arrests pre, during and post sentence date 21 AAMR criteria for inclusion - at least 1 month on AAMR tag and at least 1 month sentence post tag. This aided calculating monthly offending rates for comparison with other monthly periods. Due to limitations on the Control group (sentence end date not reached or 4 month post period going into COVID-19 lockdown), only those with shorter sentences can be included. Those included have an average sentence length of 11.65 months vs. 17.09 months for those who can't be included. This is likely to be related to general seriousness / volume of offending prior to the index sentence. the two groups are therefore not matched based on the PSM pairings. Pre AAMR Tagging Post AAMR Tag removal Economic Break-Even Analysis The final analysis within the evaluation sought to explore the economic break-even point of the AAMR. That is, what level of reduced reoffending would need to be observed in order for the AAMR to ‘break even’ on the costs. MOPAC commissioned Red Quadrant to conduct this analysis. Calculating costs for delivery of the AAMR included both tagging and supervision22. The total cost of the AAMR group is calculated by adding the direct costs of delivering the AAMR (100 x £1409.76 = £140,976) to the economic costs of predicted reoffending (£346,91923) making a total of £487,895. The equivalent costs for a matched group of offenders on a community sentence involving just probation supervision24 would be the direct costs of delivering that supervision (100 x £599.56 = £59,956) added to the economic costs of predicted reoffending £170,62525) making a total of £230,581. Therefore, currently the total costs of placing 100 offenders on a AAMR outweigh the costs of placing a similar cohort of offenders on a community sentence involving probation supervision by £257,314. In other words, the net cost to society of sentencing an individual offender on a AAMR is £2,573 more than sentencing the same individual to a community sentence involving probation supervision. If we are to base our breakeven analysis on the findings of this impact study which finds that the economic cost (comprising the cost of delivering the sentence with the economic consequences of the 22 Given the uncertainty of future costings and the fact that any future rollout of AAMRs in London will not require dedicated resources from MOPAC, we have decided to exclude this additional cost and have based this CBEA on a unit cost per AAMR of £1391.25 (Tagging, which includes renting of monitoring equipment, monitoring costs and delivery costs) + £18.51 (Supervision) = £1409.76. 23 Since this information relates to charges rather than convictions we have assumed that all charges are prosecuted but reduced the total cost in line with the latest national conviction rate of 87% to leave a revised total economic cost of £1,300,946. Finally, for the purpose of making cost comparisons easier to follow, we have produced a figure for the economic cost of reoffending for a cohort of 100 individuals sentenced to a AAMR. Assuming 21.6% of this cohort offends, based on results from our impact analysis involving a matched AAMR and comparison group, the total economic cost of the reoffending of a cohort of 100 offenders placed on a AAMR is £1,300,946 ÷ 81 x 21.6) £346,919. 24 To calculate the level of supervision Red Quadrant have assumed that the supervising officer spends an average of 30 minutes per week supervising an individual on a community sentence and requires that individual to attend every week. The assumption of 30 minutes is based on the findings of Tidmarsh (2019) who observed 47 supervision interviews undertaken by offender managers working in a CRC as part of an ethnographic study and recorded the average length of those interviews as 17 minutes. We have allowed an additional 13 minutes for recording and other administrative duties. Using this methodology, the unit cost for the supervision element of a one year community sentence would be 26 hours of the offender manager’s time at an hourly rate of £23.06, a total of £599.56. 25 Since this information relates to charges rather than convictions we have assumed that all charges are prosecuted reduced the total cost in line with the latest national conviction rate of 87% to leave a revised total economic cost of £640,972. Finally, for the purpose of making cost comparisons easier to follow, we have produced a figure for the economic cost of reoffending for a comparison cohort of 100 individuals. Assuming 18.9% of this cohort offends, based on the findings from our impact analysis, the total economic cost of the reoffending of a cohort of 100 offenders placed on a community sentence involving supervision is £640,972÷ 71 x 18.9) £170,625. higher reoffending rate) of sentencing 100 offenders in London to a AAMR is an additional £257,314 compared to sentencing those same offenders to a community sentence involving supervision only, the breakeven point would be achieved by a cohort of 100 AAMR cases committing, for example, around 24 fewer offences of violence against the person. We have chosen to illustrate the breakeven point by this offence group because it is the most commonly committed offence by the AAMR cohort and has a high economic cost to society as calculated by the Ministry of Justice of £10,667 per offence. However, the results should only be viewed as indicative as successful matches were not returned for all of the AAMR and control cohort offenders from CRIS. It is also worth noting that the control group offenders may have also received support such as a community payback scheme or treatment programmes as part of a condition associated with their community sentence, however, these costs were not accounted for in the CBA, costs were based on a community sentence requiring supervision only. 6. Conclusion The AAMR was introduced as a pilot concept in 2014 to address the significant problem of alcohol related offending in London. This sentencing power enabled courts to impose, as part of a Community Order or Suspended Sentence Order, a requirement that compelled an offender to abstain from alcohol for a fixed time period and be regularly tested via electronic monitoring. Previous reports present the findings from a series of performance and process evaluations, highlighting the implementation success, positive views of both staff and individuals wearing the tag, and learning generated from proof of concept to pan-London roll-out26. Similarly, early results exploring impact were published in 2017 focussing upon the pilot cohort, reporting no significant differences comparing the pilot AAMR cohort and a matched comparison across reoffending. This report concludes the AAMR evaluation in taking a final look at the impact of the AAMR on reoffending behaviour over the course of the pan-London roll-out (April 2016-June 2018). Following these results the technology has been rolled out nationally. Prior to discussing the results, it is worth highlighting that there were some limitations with the data - successful matches were only observed for police data relating to 72% for the AAMR cohort and 67% for the comparison group. Possible reasons for a lack of match for both cohorts include the inaccuracy of PNC IDs at point of recording, missing PNC IDs, wider data error or MPS crime reporting processes. When comparing the AAMR matched sample with those that did not retrieve a successful match from CRIS we saw no significant differences between the samples across the demographics and offence type matching criteria, however, there was a small difference between OGRS 2 scores and as a result a variety of means were attempted to increase the match; such as quality assuring all PNC IDs, fuzzy matching using PNC ID, name and date of birth, and exploring wider data sources to capture offences which are not recorded on CRIS systems, all of which increased the successful hits to 84% (AAMR) and 88% (control). Even with this challenge, the current analysis of AAMR remains one of the largest explorations to date of the technology. In terms of the results, this analysis has demonstrated that the AAMR is broadly comparable to other Community Orders/Suspended Sentence Orders in terms of reoffending behaviour as measured by both arrests or charges and echoes previous findings from our early impact analysis27. To illustrate, across both 6 and 12 months subsequent to the order, the proportions of AAMR and control that offend (be it arrest or charge) are not significantly different. The difference that is observed is that the AAMR cohort have a higher average number of offences (arrest or charge) than the control cohort, across both 6 and 12 months - indicating that when they do reoffend, they commit more (at an average of 1 offence more in both 6 – 12 months). This 1 offence more was a statistically significant difference, but overall questions could be posed as to whether this difference has a practical significance as well. Where offending did occur for the AAMR group, it was far more likely to occur once the tag was removed. To further illustrate this, the average speed to first reoffence for the AAMR offenders was 132.6 days (range = 3-309), which is just over double the time the individuals were subject to the tag (64 days, range = 4 -121 days). Whilst both the AAMR and the control cohort do reoffend to some 26 Hobson Z,. Harrison A. & Duckworth, L. (2019) Alcohol Abstinence Monitoring Requirement – A review of Process and Performance from year 2. MOPAC 27 Hobson, Z., Harrison, A. & Dawson, P. (2017) Alcohol Abstinence Monitoring Requirement – South London Pilot Indicative Impact Report,. MOPAC degree, it is also worth reflecting that offending rates are lower for the AAMR cohort and control than the wider reoffending rates throughout England and Wales (28.7%)28, confirming that both cohorts are at low risk of reoffending in general. Overall, reflecting upon the offending results - given the far shorter length of the AAMR order, which brings with it a far longer ‘at risk’ period (i.e., no AAMR order) relative to the control group, such comparisons over 6 and 12 months should be viewed as an overly rigorous test of success for the AAMR. However, these results do demonstrate the need for continued support for this group subsequent to tag removal to address potential offending. Analysis also sought to identify factors that had the largest effect on risk of reoffending; these were age (older less likely to offend) and type of order the offender was sentenced to (community orders had a reduced risk of recidivism), which is something that has been echoed by other studies29 30. Whilst we are seeing similar findings to other Community Orders/Suspended Sentence Orders in terms of reoffending behaviour, the financial analysis does indicate that the AAMR is more costly than these orders. If we are to base our breakeven analysis on the findings of this impact study which finds that the economic cost (comprising the cost of delivering the sentence with the economic consequences of the higher reoffending rate) of sentencing 100 offenders in London to a AAMR is an additional £257,314 compared to sentencing those same offenders to a community sentence involving supervision only, the breakeven point would be achieved by a cohort of 100 AAMR cases committing, for example, around 24 fewer offences of violence against the person. This report concludes the evaluation of the AAMR. The evaluation has been active since 2014, monitoring the AAMR from a small-scale proof of concept to the pan-London implementation. The evaluation has covered performance, process, impact and economic aspects with findings being routinely fed back into practice. Overall, findings appear positive in terms of the technology itself, the implementation, and the perceptions of both staff and individuals. This evaluation has also highlighted the challenges of reducing reoffending, particularly for low risk offending groups and the importance of providing adequate support to offenders at the end of their orders. Given the inherent risks of delivering innovation, to be broadly achieving the same results as longer term, albeit less costly orders, the AAMR innovation should be seen as a firm foundation to build upon. 28 Proven Reoffending Statistics Quarterly Bulletin, England and Wales, January 2018- March 2018 https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/861994/pr oven_reoffending_bulletin_January_to_March_18.pdf 29Howard, P., Francis, B., Soothill, K. and Humphreys, L. (2009) OGRS 3: The revised Offender Group Reconviction Scale. Ministry of Justice Research Summary 7/09. Available at http://www.justice.gov.uk/publications/docs/oasys-research-summary-07-09-ii.pdf 30 A, Mews et al (2015) The risk of short custodial sentences, community sentences and suspended sentence orders on reoffending. Ministry of Justice Analytical Series. https://pdfs.semanticscholar.org/6051/a8528d7a7f8c11b299575c0913497edf740b.pdf REFERENCES Austin. P (2011). An Introduction to propensity score methods for reducing the effects for confounding in observational studies. Multivariate Behavioural Research, 46, 399-424.Toylor and Frances Group, LLC. A, Mews et al (2015) The risk of short custodial sentences, community sentences and suspended sentence orders on reoffending. Ministry of Justice Analytical Series. https://pdfs.semanticscholar.org/6051/a8528d7a7f8c11b299575c0913497edf740b.pdf Evaluation of the AAMR tagging pilot: 2 year process evaluation finding, NatCen, 2019. http://www.hlnycrc.co.uk/wp-content/uploads/sites/4/2020/01/Process-Evaluation-Year-2-Interim- Report-Final.pdf Hobson, Z., Harrison, A. & Dawson, P. (2017) Alcohol Abstinence Monitoring Requirement – South London Pilot Indicative Impact Report,. MOPAC Hobson Z,. Harrison A. & Duckworth, L. (2019) Alcohol Abstinence Monitoring Requirement – A review of Process and Performance from year 2. MOPAC Howard, P., Francis, B., Soothill, K. and Humphreys, L. (2009) OGRS 3: The revised Offender Group Reconviction Scale. Ministry of Justice Research Summary 7/09. http://www.justice.gov.uk/publications/docs/oasys-research-summary-07-09-ii.pdf Proven Reoffending Statistics Quarterly Bulletin, England and Wales, January 2018- March 2018 https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file /861994/proven_reoffending_bulletin_January_to_March_18.pdf Ringland, C. (2013). Measuring recidivism: Police versus court data. BOCSAR NSW Crime and Justice Bulletins, 12. Transforming rehabilitation: A summary of evidence on Reducing Reoffending, MoJ Analytical Series, 2013.https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_da ta/file/243718/evidence-reduce-reoffending.pdf The factors associated with proven reoffending following release from prison, MoJ Analytical Series 2013.https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_da ta/file/491119/re-offending-release-waves-1-3-spcr-findings.pdf Pepper, M. & Dawson, P. (2016) The Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot, MOPAC. Appendix A – Propensity Score Matching Procedure The first step of propensity score matching technique was to develop a regression model that is able to predict group membership (i.e., the propensity of receiving the AAMR treatment). Five theoretically relevant matching variables were selected and entered in a logistic regression model predicting group membership. The variables were: (1) OGRS score, (2) Age, (3) Gender, (4) Ethnicity, (5) Index Offence. The resulting model was significant, meaning that the procedure of using this approach can continue as the variables are significant in predicting AAMR group membership. The classification table confirms the model is correctly predicting group membership for 64.5% of the cohort. 472 cases were correctly predicted to be within the control and 407 were correctly predicted to be within the AAMR cohort. The variables in the equation table shows that age, Index offence and OGRS2 were significant in predicting AAMR group membership. The approach then seeks to match pairs (i.e., control and AAMR) of individuals on their propensity score of receiving the treatment. To do so, an approach called nearest neighbour optimal matching method was used (Austin, 2011). A caliper was set for the propensity scores, meaning a control individual could only be matched to an AAMR individual if their propensity score fell within 0.2 of the standard deviation of the logit of the propensity score. The optimal matching method ensured the best-matched pairs were chosen. The matching procedure was conducted without replacement, meaning once a control individual was matched to an AAMR individual, they could not be matched again. The above histogram shows a normal distribution for both AAMR and control propensity scores included in the match, with a symmetrical distribution and near superimposition of both groups, proving that the match is robust and there is a low probability of confounding. 0 50 100 150 200 250 300 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Propensity Scores for AAMR and Control Matches AAMR Matches Control Matches Total Matches 1 Alcohol Abstinence Monitoring Requirement A review of process and performance from Year 2 July 2018 Dr. Z. Hobson, A. Harrison & L. Duckworth MOPAC Evidence and Insight July 2018 MOPAC Evidence and Insight 2 Contents Executive Summary ............................................................................................................. 3 1. Introduction .................................................................................................................... 5 Sobriety Programmes and Interventions............................................................................... 5 Alcohol Abstinence Monitoring Requirement ....................................................................... 6 The 2014/2015 Pilot .......................................................................................................... 7 Pan London Roll Out .......................................................................................................... 8 2. Methodology ................................................................................................................. 10 3. Results ........................................................................................................................... 12 Using the AAMR: Performance Learning ............................................................................ 12 Using the AAMR: Process Learning from Stakeholders ....................................................... 15 Using the AAMR: Process Learning from Offenders ........................................................... 20 4. Discussion ...................................................................................................................... 24 References ......................................................................................................................... 26 Appendices ........................................................................................................................ 29 3 Executive Summary In 2011 the Mayor’s Office for Policing And Crime (MOPAC) secured legislation to allow for the introduction of a new sentencing power - the Alcohol Abstinence Monitoring Requirement (AAMR) - to tackle the significant problem of alcohol related offending in London. The AAMR gives the Judiciary the statutory power to impose compulsory sobriety, or to stop an offender drinking alcohol, where their offence is alcohol related. The AAMR involves fitting a tag to the offender’s ankle and monitoring their alcohol consumption for up to 120 days. When this is not complied with, the offender will be breached and punished further. Following the positive learning from the initial AAMR proof of concept pilot (Pepper & Dawson, 2016), a two year pan London roll out of the AAMR was launched in April 2016, joint funded by MOPAC and the Ministry of Justice. The aim of this expansion was to test the impact of the AAMR, in line with the Conservative government’s 2015 manifesto commitment to make sobriety tags available across England and Wales. However, sobriety tags were not included in the 2017 Conservative Manifesto and there are no immediate plans from central government to fund a national rollout of the AAMR. As a result, the pan London AAMR programme’s main delivery period ended at the end of March 2018, with a further three month extension period until June 2018 to enable a managed closedown of the programme. This report focuses only on those cases imposed during the main two year delivery period, until 31 March 2018. Key Findings: Performance learning • A total of 1,014 AAMRs have been imposed between April 2016 and March 2018, with an average length of 61 days tagged. • The vast majority of offenders who completed the AAMR did so successfully, indicating a 94% (n=880) compliance rate with this requirement1. This has remained consistent since the introduction of the AAMR in London. • In total, the 1,014 offenders were monitored for 71,584 days in the two year period and were sober for 69,996 of those days. These figures indicate that in 98% of the days offenders were monitored they did not consume alcohol. • AAMRs were usually imposed as part of a Community Order (73%, n=740), and standalone AAMRs accounted for 29% (n=297) of all Orders. Multiple requirement Orders accounted for 69% (n=702), usually consisting of AAMR and Unpaid Work (19%, n=193). • AAMRs were given for a variety of crime types, most commonly in relation to violence (45%, n=456) or drink driving offences (29%, n=291), which is higher than reported in the interim report (31%, n=115 and 22%, n=82 respectively). • AAMRs have been imposed across London, most frequently from Magistrates Courts (91%, n=922). Magistrates Courts in the South London LJA (Croydon and Camberwell Green Magistrates) were responsible for a quarter of all AAMRs in both the first (25%, n=87) and second year (26%, n=150/587). This is unsurprising given that they had already gained momentum and understanding of the requirements from being part of the pilot initiative in 2014/15. However, some courts that came on later also imposed a significant proportion of AAMRs (e.g. Highbury Corner imposed 13% of all AAMRs despite coming on in October 2016). 1 For the purpose of measuring compliance we have recorded an unsuccessful completion when alerts about violations on the tag led to enforcement action being taken by the Offender Manager that led to a breach conviction at Court. 4 • Findings in Year 2 are generally comparable with MOPAC’s previous learning around the AAMR in terms of usage and offences and it is encouraging to see that the compliance rate has remained consistently high since the AAMR pilot. Process learning • To gather learning about the AAMR programme, stakeholders, including Magistrates, National Probation Service (NPS) and London Community Rehabilitation Company (CRC) staff across London were surveyed (N=44) and interviewed (N=24). Additionally, offenders who received an AAMR were asked to complete a survey when the tag was fitted (N=412) and when it was removed (N=407). • Overall the AAMR has been welcomed across London and supported by stakeholders across London, as a tool tailored to specifically addressing alcohol related offending behaviour. • Stakeholders who received specific AAMR training had generally positive feedback on the training. However, some reported not receiving any relevant training, despite it being available pan-London. This may be due to the high turnover of staff and frequent rotation of magistrates. Providing on-going /refresher training could address this in future. • When AAMRs have been used, stakeholders have a good understanding of the eligibility criteria. However, many sentencers noted they often had cases who resided outside of London and would have liked the AAMR to be available for these cases as well. • Delays in tagging the offender have continued in Year 2 of the programme, with only 42% (n=382) tagged within 24 hours of receiving their Court Orders. This is a significant reduction from the pilot (82%) and Year 1 (50%), and illustrates the continued scale up challenges of covering the whole of London. Tagging at Source was supposed to mitigate some of these issues, but due to staffing and infrastructure issues very few offenders (n=8) were actually tagged at Court. • Positively, the AAMR has had little impact on the workload of stakeholders, who spoke generally positively about the requirement, noting that it was a “useful tool” to their role (82%, n=36). • It was felt that the AAMR, whilst a punitive measure, should also be viewed as a rehabilitative requirement, as it provides the opportunity to reflect on one’s behaviour. Stakeholders indicated that the AAMR has the potential to have a positive impact on the lives of the offenders, particularly around reducing their alcohol consumption (57%, n=25) and reoffending (86%, n=16). • Offenders were generally optimistic about the requirement, and felt that the AAMR had a positive impact on their lives, particularly around their health, wellbeing and offending behaviour. However, practical concerns were raised around the size and design of the electronic tag itself and the stigmatisation that wearing it may cause. • The majority of stakeholders welcomed a national roll out of the AAMR (82%, n=36) to allow for more consistency in sentencing across the UK. • Overall, performance results and learning from stakeholders and offenders reflect the positive findings reported in the interim report. This report sits as part of a wider, holistic evaluation to test the impact of the pan-London AAMR programme. As we come to the end of the programme, further research in the form of proven reoffending analysis to explore the impact of AAMR on offending behavior and a full cost benefit analysis will continue. These final elements will be reported on in the final evaluation report in Spring 2019. 5 1. Introduction It is widely understood that alcohol use contributes to criminal behaviour, particularly around violent crime and public disorder in the UK. It was last estimated that the total social and economic cost of alcohol related harm was £21.5 billion (Public Health England, 2018). Alcohol is recognised as a major cause of attendance at Accident and Emergency departments and accounts for over 1 million hospital admissions each year (Public Health England, 2016) and more than 23,000 related deaths in the year 2015/16 (Public Health England, 2017). Historically, as measured by the Crime Survey for England and Wales (CSEW), approximately half of violent incidents are related to the influence of alcohol. Across England and Wales in the year ending March 2017, victims recognised the offender as having consumed alcohol within 40% of violent crimes (ONS, 2018). Whilst there has been a slight decline in violent crime related to alcohol (from 53% in 2013/14 to 40% in 2015/16) and alcohol-related road traffic accidents (Public Health England, 2016), it remains clear that there is a longstanding resistant association between alcohol and violence. Recent trends suggest that alcohol is increasingly present among violent incidents that are likely to occur in a pub or club (93%), at the weekend (64%), and during the evening/night time (65%) in comparison to the previous year (ONS, 2018). Victims of alcohol related violent incidents are also more likely to receive greater injuries (ONS, 2015). Wider data also contributes to the picture - around two-fifths (40%) of those who are worried about ASB in their local area feel that people being drunk or rowdy in public places is a problem (MOPAC Public Attitude Survey (PAS), 12 months to Quarter 1 2018-19)2. Similar views are held by London businesses who perceive people being drunk / rowdy in their local area to be a problem (26%) (MOPAC Business Attitude Survey, 2014 – 2016). Criminal behaviour and alcohol are intrinsically linked, with large costs to both the public purse and public health and wellbeing. Many attempts have been made to address alcohol use over the years from both a criminal and health related perspective. This report will review relevant interventions and further discuss the introduction of a compulsory sobriety programme introduced across London to address alcohol related offending. Sobriety Programmes and Interventions Following a number of popular drug and alcohol sobriety programmes (e.g. 12-step Programme, Alcoholics Anonymous), the focus within sobriety interventions has shifted from addressing addiction, towards specific behavioural implications such as offending. The primary approach has been the monitoring of alcohol-use through various methods; from random sobriety check-points and ignition interlocks in addressing driving under the influence (DUI) (Bergen, Pitan, Shults & Sleet, 2012; Blais & Dupont, 2005; Kerns, 2017; Roth, Marques & Voas, 2009; Vanlaar et al., 2017) to continuous monitoring via transdermal tags, urine and blood testing to reduce alcohol- fuelled offending (Dougherty et al., 2012). The focus of many of these programmes has been around assessing the efficacy of the equipment and compliance with the programmes’ ethos. 2 The PAS explores the views of the residents across London around crime, ASB and policing issues via a face to face interview with over 12,800 respondents per year. 6 While few studies have explored the impact of such interventions, those that have, have published promising results. Most notably, the South Dakota 24/7 Sobriety Programme (Kilmer, Nicosia, Heaton & Midgette, 2013) sought to reduce DUI offences using transdermal tags and/or twice- daily breathalyser tests to encourage complete abstinence from alcohol. Primarily targeting repeat offenders, the programme combines constant alcohol monitoring with ‘swift’ and ‘modest’ sanctions - those who breach are immediately taken into custody or court (Kilmer et al., 2013). Compared to counties where a 24/7 sobriety programme was not implemented, results suggested a 12% reduction in DUI repeat-arrests and a 9% reduction in arrests related to domestic abuse across the 5 years following the intervention. Recent research in North Dakota further identified reduced DUI offence rates following completion of the 24/7 Sobriety program. Interestingly, the longer period an individual was enrolled in the program, the less likely they were to re-commit DUI (Kubas and Vachal, 2017). A similar project assessing the use of Secure Continuous Remote Alcohol Monitor (SCRAM) tags (Flango & Cheesman, 2009) had mixed results but provided vital learning. Within this study a small difference, albeit not significant (+2.8%), in recidivism across two years was seen between those who wore a SCRAM tag and matched controls, increasing by over 10% when restricted to prolific offenders with at least two prior convictions (+12.9%). While unable to infer impact from these results, the data provided learning in terms of the timeliness of the order. Those who wore the tag reoffended more quickly than controls and often to a greater extent following tag- removal. Investigating this trend further, Flango and Cheesman (2009) found this effect was mediated by the length of the order; those who wore the tag for at least 90 days reoffended at around half the rate of those who did not wear the tag (10% vs. 21% respectively) whereas recidivism for those who wore the tag for less than 90 days was almost equal to controls. However, whilst the implementation of transdermal alcohol monitoring devices have been noted to reduce alcohol consumption, the physical appearance and discomfort of the device has been perceived by users to be problematic (Alessi et al., 2017; Averill et al., 2018) and in turn could impact upon co-operation with the intervention. It is unclear at this time what the longer term behaviour effects are from participating in a programme of enforced sobriety and Axdahl (2013) suggested that behavioural effects beyond tag removal may be short-lived. Comprehensive evidence around the efficacy of enforced sobriety is both lacking and mixed. Despite this, an intervention of this kind has been well-implemented across a number of US counties and states (Kilmer et al, 2013) and the Alcohol Abstinence Monitoring Requirement (AAMR), inspired by the approach in South Dakota, provides an opportunity to address alcohol related offending under UK criminal justice legislation. Alcohol Abstinence Monitoring Requirement In 2012, a new sentencing power was introduced as part of the then Mayor’s manifesto pledge to address the significant problem of alcohol related offending in London. Under the Legal Aid, Sentencing and Punishment of Offenders (LASPO) Act 2012, the Judiciary are allowed to impose the punitive Alcohol Abstinence Monitoring Requirement (AAMR); a requirement that necessitates offenders abstain from alcohol for a fixed time period of up to 120 days. To be eligible to receive an AAMR, offenders must meet the following criteria: 7 • Consumption of alcohol must be an element of the offence or an associated offence, or the court must be satisfied that consumption of alcohol was a factor that contributed to the offender committing the offence or an associated offence; • The offender must not be dependent on alcohol3; • The court must not include an alcohol treatment requirement (ATR) in the order (ATRs are for dependent drinkers only); • The offender must live in London; and • The offence must not have involved domestic abuse (at the present time). Compulsory sobriety is measured via regular testing via a transdermal alcohol monitoring devise (a tag around the offender’s ankle) as part of a Community or Suspended Sentence Order4, and when this is not complied with, the offender will be prosecuted under breach proceedings and punished further. The 2014/2015 Pilot An initial 12 month pilot of the AAMR was commissioned in South London by the Mayor’s Office for Policing And Crime (MOPAC) commencing in July 2014. This pilot was subject to a process and performance evaluation5. Over the course of the pilot, 113 AAMR Orders were imposed by the Courts. The AAMR pilot had a final compliance rate6 of 92%7 which compares favourably with other orders - analysis by the NPS in 2014 estimated a compliance rate of 61% for other community based Orders it managed with the Community Rehabilitation Company (CRC)8. Variation was also seen with different requirements - 82% of offenders completed Unpaid Work Orders successfully9 in London compared to Alcohol Treatment requirements (80%) and Drug Rehabilitation requirements (67%) (Ministry of Justice, 2015). The research indicated the AAMR was received well, particularly by the judiciary and stakeholders, who recognised the AAMR as an important ‘tool in the box’. This can in part be attributed to the 3 The Probation Court Team officer will assess “suitability” with alcohol AUDIT Tool. This assessment is used to determine the offender’s alcohol dependency levels. 4 Providing the offender is deemed both suitable and eligible. 5 https://www.london.gov.uk/sites/default/files/aamr_final.pdf 6 Caution needs to be applied when interpreting the completion and compliance rate of AAMR – this was a pilot study with a small sample size enabling the project manager to provide some assurances that the large majority of the AAMRs were enforced when failures to comply arose. This may not necessarily be the case with other Orders and requirements. 7 The AAMR pilot had a compliance rate of 92%, based on the number of cases (n=9) who were returned to court and convicted on breaching their AAMR as a proportion of all cases imposed. Of these nine, five had their AAMR revoked and failed to complete, and the remaining 4 completed their AAMR following their return to court. This gives a final completion/compliance rate of 95% (Pepper, & Dawson, 2016). 8 However there are caveats to be considered such as a direct ‘like for like’ comparison is not possible due to different offence types, offender characteristics, breach processes and the length of the orders themselves. 9 These figures should be caveated however as the AAMR project manager recently reviewed other requirements on Orders (i.e. UPW), finding there were numerous occasions when breaches were not enforced and cases were simply closed, which may distort the actual compliance rate. 8 strength of the design and implementation of the programme. There were clear toolkits and training provided, effective partnership working and a project management team in place with relevant experience in this area. The effectiveness and certainty provided by the technology, as well as a strong understanding of the aims of the pilot and how the AAMR works in practice amongst both offenders receiving the order and stakeholders involved in its delivery also helped. In addition, there were a number of associated positive consequences of the pilot, including, but not limited to: the period of abstinence gave offenders a ‘pause’ in their drinking; it also provided time for reflection of their alcohol consumption and the impact it has on offending behaviour, work and relationships; and an opportunity was provided for offenders to break their cycle of routine drinking. The AAMR was also used as a ‘teachable moment’ in some instances, with products such as tailored advice and relevant literature supplied by the service providers to support offenders further. The report also stressed the potential challenge in sustainability and the scale up challenge in moving from a small scale pilot to a pan-London approach. This initial pilot research was not able to explore robust impact on reoffending due to sample size and follow- up time limitations. Pan-London Roll Out The pilot was considered to be a success and, following the Conservative government’s 2015 manifesto commitment to make sobriety tags available across England and Wales, MOPAC and the Ministry of Justice agreed to joint fund the roll out of the AAMR pan London from the 1 April 2016 until the 31 March 2018. The pan-London roll out provided the opportunity to test the effectiveness of the South London pilot on a much larger, more complex scale and to measure the impact of the AAMR in line with central government’s commitment to roll this out nationally.10 To achieve this the core elements of the South London proof of concept pilot were maintained, including the use of transdermal tags which enabled a full and comparative evaluation to be completed. The London Sobriety Project aimed to test: - Learning from the original pilot; - Take up of the AAMR requirement by the Judiciary; - Compliance with the AAMR; - Completion rates of the AAMR; and - Re-offending rates. Building upon the findings from the MOPAC pilot study, this report outlines findings from the two years of the pan-London roll out of the AAMR. It details the implementation process of the AAMR through the views and experiences of stakeholders involved in delivery and offenders 10 However, sobriety tags were not included in the 2017 Conservative Manifesto and there are no immediate plans from central government to fund a national roll out. MOPAC believes that it is the responsibility of central government to fund the AAMR, just like any other sentence requirement. 9 sentenced to wear the alcohol tag and analysis of performance metrics. These findings sit as part of a wider holistic evaluation around the use of the AAMR across London, which also includes proven reoffending analysis to explore the impact of AAMR on offending behaviour and a full cost benefit analysis. These elements will be reported on in the final report in spring 2019. 10 2. Methodology Both quantitative and qualitative research methods were employed to triangulate learning and gain an understanding of the AAMR working processes, how the AAMR is performing and experiences of offenders and stakeholders. A variety of methodologies were used to collect data11, including: • Stakeholder surveys: An online survey was completed by 44 stakeholders to explore their understanding and experiences of AAMR. The survey was distributed electronically to all Local Justice Areas and Probation Trusts across London, with a follow up email to encourage responses. Respondents were largely Magistrates (55%, n=24), or working for the National Probation Service (NPS) or Community Rehabilitation Company (CRC) (25%, n=11) (Appendix A). • Stakeholder interviews: To gain a more in-depth understanding of stakeholders’ views telephone interviews were conducted. A total of 24 semi-structured interviews were conducted with a variety of stakeholders at practitioner and strategic levels across London12 (Appendix B). Topics included: understanding the rationale, partnership working, implementation, usage, decision making, suitability, perceived impacts, good practice, lessons learned, challenges, and broader attitudes to the equipment. • Offender surveys: Offenders who received an AAMR were asked to complete a survey at the time the electronic tag was fitted, and when it was subsequently removed. The two surveys sought to understand their first impressions of the tag, perceptions of what life may be like whilst wearing the tag and once it has been removed. Surveys were given to the offender by the Electronic Monitoring Services (EMS) tag fitter13 or the Probation Office if tag was fitted at source (e.g. the Court house – Westminster Magistrates or Bromley Magistrates only). Completing the survey was not compulsory, and some offenders chose not to participate. In total, 412 (out of a possible 915, 45%) completed the survey at the time when the tag was initially fitted, and 407 (out of a possible 83714, 49%) completed it during tag removal15. 11 Given the size of the research cohort (e.g. the number of respondents to the stakeholder survey/offender surveys), caution should be used when considering the results. Response base size is provided, however this varies as not all respondents answered every question. 12 Potential interviewees were identified with the AAMR project manager and contacted via email by the researchers. There was no obligation to participate, therefore participants were self-selecting. Detailed notes were taken in all interviews, and analysed to draw out themes. 13 Whilst this method of distribution has its limitations, this was the most practical approach available for obtaining insightful data on offender perceptions and experiences. 14 This accounts for offenders who have completed their AAMR requirement and are no longer an ‘active’ case. 15 Due to the way the data was anonymously collected, it is not possible to link survey responses to know whether offenders who completed the initial survey also completed the removal survey. 11 • Performance monitoring data: A range of performance data was gathered from both the NPS/CRC and EMS - the company which conducted the field delivery and assisted in data collection of performance metrics about the tag. Performance metrics included: numbers of AAMRs given, types of offences, court details, demographics on who received the tag, number of breaches, days of sobriety and compliance with tag. 12 3. Results Using the AAMR: Performance Learning Imposing the Requirement Throughout the two years of the pan-London AAMR project (April 2016 – March 2018), 1,014 AAMRs were issued by the London courts, with Magistrates’ Courts imposing the requirement most frequently (91%, n=922). There were an additional 133 cases imposed during the closedown period between April and June 2018, but these cases will not be touched on in this report. Following an initial pilot period in the South London Local Justice Area, the AAMR programme was rolled out in a phased approach between April 2016 and January 2017. As Figure 1 demonstrates, there was a steady increase in the use of the AAMR across London over the following year where all Local Justice Areas had the option to impose the requirement. Figure 1. Number of AAMRs imposed across the two year period. In the second year of the project, with all areas ‘live’, nearly double the number of AAMRs were imposed compared to the first year (647 AAMRs vs 367 AAMRs). Overall, Croydon Magistrates’ Court was the most active in imposing the AAMR (14%, n=144). The Magistrates’ Courts in the South London LJA (Croydon and Camberwell Green Magistrates) were responsible for a quarter of all AAMRs in the first (25%, n=87/342) and second year (26%, n=150/587). This is unsurprising given they had already gained momentum and understanding of the requirements from being part of the pilot initiative. Their enthusiasm for the requirement appears to have continued throughout the project. Magistrates’ Courts in the North West London LJA also frequently imposed AAMRs, totalling 18% (n=106/587) of all AAMRs imposed in Magistrates’ Court. This is likely to be a reflection of the North West London LJA having five Magistrates’ Courts, whereas most other Local Justice Areas have approximately two Magistrates’ Courts. Therefore, it may be expected the throughput of AAMRs to be greater in this area (see Appendix C for a full breakdown of Courts). 0 10 20 30 40 50 60 70 80Apr-16May-16Jun-16Jul-16Aug-16Sep-16Oct-16Nov-16Dec-16Jan-17Feb-17Mar-17Apr-17May-17Jun-17Jul-17Aug-17Sep-17Oct-17Nov-17Dec-17Jan-18Feb-18Mar-18 Roll-out of AAMR complete across London 13 To be considered for an AAMR, the offence must be alcohol related; however, this leaves a broad scope of the type of offences committed. In total, 1,014 AAMRs were issued for 22 different offence classifications (see breakdown of offences in Appendix D). The majority of AAMRs were imposed for violence (45%, n=456) or driving offences (29%, n=291), which is unsurprising in light of the previous literature (e.g. McSweeney, 2015). Violence can cover a range of offences, however in terms of those offences where an AAMR was received, they tended to be low level common assault (57%, n=261) or resisting/assaulting a police officer in the execution of their duty (17%, n=76). Only 2% (n=10) of offences were classed as wounding or grievous bodily harm (GBH). When compulsory sobriety was imposed, this was usually as a requirement of a Community Order (73%, n=740) rather than a Suspended Sentence Order (27%, n=274). The majority of Court Orders received multiple requirements (69%, n=702) with 29% (n=297) of offenders receiving a standalone AAMR, reflecting the Year 1 findings. Again, this may indicate that the Court and Probation (who recommend sentencing options) may feel that the AAMR requires the support of additional requirements (such as the Rehabilitation Activity Requirement), or that the AAMR only addresses a specific element of their offending (alcohol use), so AAMR is paired with additional requirements to address other issues. When the AAMR was used in conjunction with other requirements, there were large variations in the types of additional requirements that it was paired with. Table 1 details the most frequently combined requirements with an AAMR. Table 1. Additional requirements paired with AAMR. Requirement Number of Orders % Standalone AAMR 297 29% AAMR + Unpaid Work (UPW) 193 19% AAMR + Rehabilitation Activity Requirement (RAR) 136 13% AAMR + RAR + Accredited Programme 55 5% AAMR + UPW + Prohibited Activity 53 5% AAMR + RAR + UPW 52 5% AAMR + Prohibited Activity 44 4% Other combination of requirements16 184 18% Total 1,014 100% The AAMR Tag The requirements of the AAMR, to remain abstinent from alcohol, start immediately once the sentence has been imposed, despite the actual monitoring equipment (the tag) being fitted subsequently. Where possible, EMS (the tagging company) should receive notification of the sentence, and hence the request to tag an individual, within 24 hours of the offender receiving the Order. In 91% of cases, EMS received the notification from the Courts on the same day as sentence (n=759) or the following day (n=159). Although this indicates a clear and effective 16 This includes cases where there the requirements were not recorded by CRC/NPS (n=15). Other requirement combinations included Attendance Centre, Specified Activity, Exclusion and Curfew. 14 communication between the Courts and EMS, only 42% (n=382/918) of offenders (where EMS were notified on the same or following day) were tagged on the same day or within one day of notification17. On average, offenders were tagged within four days of EMS receiving the notification. This is a significant reduction (p<.05) from the findings of the pilot (82%) and interim report (50%) which may reflect further resourcing issues deriving from the scaling up of the project, having to fit almost double the number of tags in the second year compared to the first and the EMS tag fitters ability to effectively cover a much larger geographical area. Despite these issues, tags were usually fitted (82%, n=831) on either the first (n=646) or second attempt (n=185). To address this scale up issue, the operational model was changed in the second year of the programme. This introduced ‘tagging at source’ where offenders would be tagged immediately after sentence either at the Court house or nearby Probation Office instead of within their own homes. The tagging was conducted by Probation Officers instead of EMS colleagues. As this was a new element, it was tested within two Magistrates’ Courts, Westminster and Bromley. Unfortunately, due to staff absences and the difficulties with the infrastructure required to tag at source (e.g. access to the court Wi-Fi), there were delays in getting the pilot fully up and running, resulting in tagging at source being rarely used, with only eight offenders receiving their tag at the Court house by Probation. Therefore this will have had no impact on EMS’ ability to effectively cover the large geographical area required to tag offenders and provides an explanation as to why it is taking EMS longer than anticipated to tag offenders. Stakeholders who answered the survey indicated mixed views on Tagging at Source. Whilst some suggested that it may be more efficient as it provides the tag more immediately and more cost effectively, others felt that tagging at the offender’s home would be more reliable, particularly in terms of ensuring that the equipment was correctly installed, and raised concerns around facilities at the two courts not being readily available. Such views may provide an insight in to the reasons why Tagging at Source was rarely used. In total, 1,014 AAMRs were imposed over the two year period, and at the end of the time period, 7% (n=74) of the AAMR requirements remained active. Of the remaining 940 AAMR cases where the Order was complete18, there have only been 60 breaches where the offender has been returned to Court and found/pled guilty to the breach19. This indicates an overall 94% compliance rate with the AAMR - a figure that has remained stable throughout the entirety of the pilot and project. In total, the 1,014 offenders were monitored for 71,584 days in the two year period and were sober for 69,996 of those days. These figures indicate that in 98% of the days offenders were monitored they did not consume alcohol. Of the completed Orders where the offender was tagged (n=837), offenders were subject to the tag for an average of 61 days (range 3 days - 144 days). This reflects a 6 day reduction in the average length of requirements since the interim report (average 67 days), and a significant 17 EMS must attend on the day for all Orders issued before 3pm, or within 24 hours for Orders received after 3pm. 18 Order complete means that the tag has been removed from the offender, or the tag was never fitted/offender never inducted. 19 For the purpose of measuring compliance we have recorded an unsuccessful completion when alerts about violations on the tag led to enforcement action being taken by the Offender Manager that led to a breach conviction at Court. 15 reduction since the pilot of AAMR (average 75 days). It is unclear at this time why such a reduction in tag duration has occurred. Offender Demographics The majority of offenders who were sentenced to an AAMR were male (86%, n=876), and white (55%, n=555) with an average age of 33 years old (ranging from 18 – 73 years). Reflecting the findings in the interim report (male: 86%, n=314; white: 47%, n=171; average age: 33 years), nearly two thirds of offenders (61%, n=617) were aged between 18 and 34 years (see Appendix E for full breakdown of demographics). The NPS and CRC consider the risk an offender poses in two ways – risk of serious harm and risk of reoffending. Risk of serious harm assessments were conducted on offenders who received either a Community or Suspended Sentence Order. Nearly two thirds of offenders were considered to be a medium risk of serious harm (60%, n=610), with just 1% (n=14) being considered a high risk of serious harm. As AAMR is designed to be used with low to medium risk of serous harm offenders, this finding would suggest that the AAMR is being used correctly. The Offender Group Reconviction Scale version 3 (OGRS3) scores for the AAMR pan-London cohort were calculated at the point of receiving a Court Order by the NPS or CRC. OGRS uses static factors such as age at sentence, gender, offence committed and criminal history to predict the likelihood of proven reoffending within either one or two years after starting their Court Order. Offenders with a high OGRS score are at greater risk of reoffending. As a group (n=964)20, the average OGRS 2 year score was 36%, (ranging from 0 to 97%) – indicating that just over one third of offenders would be predicted to reoffend within two years (see Appendix F). This is comparable to the OGRS 2 year score of the pilot AAMR cohort (35%) and when it was calculated for the interim report (covering requirements issued in April 2016 – March 2017) (37%). This shows that those receiving the AAMR are broadly identified as having a low risk of reoffending, and align to the general offending population of the UK, particularly those who receive community sentences (Farrington, 2005; Ministry of Justice, 2015b). Using the AAMR: Process Learning from Stakeholders As the AAMR pan-London roll-out reaches the two year stage, stakeholders were asked to participate in interviews and complete a survey to gain an understanding of the wider London perspective around the requirement and the impact that it may have. This section details the views and experiences of stakeholders (including staff at the NPS, CRC, Magistrates’ Courts) around training, using the AAMR, the effects of the AAMR on their workload and their views on the impact the AAMR has had on offenders. All respondents and interviewees reported that they had an overall good understanding of AAMR (86%, n=38) and its aim and objectives from the interviews. 20 OGRS Year 2 was not provided for 50 offenders and therefore for OGRS calculation these have been removed from the analysis, reducing sample size from N=1,014 to n=964). 16 Training As each Local Justice Area rolled out the AAMR, MOPAC and AMS delivered training to approximately 1,500 individuals, including a wide variety of professionals engaged with the Court and Probation Service. Overall, attendees found it a positive and useful experience (97%, n=238/245), particularly around understanding how they could apply the AAMR in their job role (93%, n=509/545)21 (Hobson, Dangerfield & Harrison, 2017). At the time of the Year 1 report (July 2017), training sessions were still relatively fresh in people’s minds, with some individuals only just having completed the training. However, there were mixed responses at the end of Year 2. The majority of survey respondents reflected the Year 1 findings and reported being satisfied with the training received (66%, n=29) and feeling that they attained enough training and support information from MOPAC (77%, n=34). Conversely, many interviewees thought that the training was “lacking” or they had not received any training at all. It was suggested that because of the high turnover of staff within organisations there needs to be some form of continuous training and refresher training to enable new staff to have a full understanding of the AAMR rather than relying on colleagues to pass on information. One interviewee commented that the AAMR would be used more if people were more knowledgeable about it, through regular refresher training. In addition to training, a range of supplementary material has been produced to inform stakeholders about the AAMR, such as toolkits, leaflets and posters. However, there remains concern that these documents have not been as widely publicised as perhaps necessary. Participants gave a mixed response to whether they had seen the material (Toolkit = 39%, n=17; Poster = 43%, n=19; Leaflet = 59%, n=26; Website = 14%, n=6); however, those who had sight of the leaflet considered it to be a “concise document”. Using the AAMR As the performance data indicates, the AAMR has been used widely across London. All interviewees (with relevant job roles) reported having experience of recommending the AAMR, and two thirds of survey respondents indicated having used the AAMR in their work (68%, n=30/44), although this was relatively infrequently, with respondents suggesting on average they had been involved in four AAMRs. There appears to be a good understanding of the eligibility criteria and in fact, of the survey respondents who had not recommended an AAMR, the main reason was that they had not come across any eligible cases (n=11/14). Interviewees also noted their frustration that due to some court locations numerous cases were heard where the offender would benefit from receiving an AAMR, however, because they resided just outside of a London borough, they were no longer eligible (e.g., Romford court that serve a London borough and Essex). Additionally, interviewees thought that the requirement would have been more impactful 21 Not all attendees at the training completed every question on the training feedback; therefore base sizes vary for each question and are reported. “We’re asked about the AAMR by the defendants or solicitors, we can’t expand on anything too much because what we know is based on quite limited information that MOPAC has given us” Interviewee 17 if the alcohol audit score was lower so that more offenders would be able to take part in the scheme and that it targeted a “very niche” cohort of individuals. Further challenges were mentioned by stakeholders working specifically within the central London courts. Here it was anticipated that they would have a relatively high number of AAMR cases given the locality and amenities, but several factors have meant this was not the case. These have included the large homeless population in the area with no fixed abode so they would not be able to participate in AAMR, those with mental health difficulties and offenders entrenched in drug and alcohol misuse. Additionally, due to the high transient population of central London (i.e. people descend on central London for employment/entertainment but not residential), many of the violent offences that occur in the area are committed by people residing outside of London and therefore not eligible to participate in AAMR. Across London, concern was also felt about the judicial process. In order to assess whether an offender is suitable for an AAMR, probation pre-sentence reports must be provided prior to the Magistrates or Judges deciding on a sentence. Often, especially within a Magistrates court, this means that reports are requested and the sentence is dealt with on another day in front of a different bench, who may be thinking along different lines for the sentence and the Probation recommendation is not always followed and an alternative sentence imposed. This may reflect a lack of understanding from judges/magistrates that could be addressed with further training or ensuring that benches are regularly reminded of all the sentencing options available to them. The AAMR is designed to act as a punitive measure, which was, in general, understood by interviewees and survey respondents. Therefore, it is unsurprising that stakeholders indicated that they would most likely recommend or impose one of the other punitive sentencing options as an alternative. Survey respondents reported they would suggest either a curfew (n=7/23) or Unpaid Work (n=6/23). However, interviewees felt that these alternative options did not directly deal with alcohol problems in the same way that AAMR has the potential to. It was also noted that the AAMR should be viewed as a rehabilitative requirement rather than just a punitive measure (77%, n=34). In fact, many felt that the AAMR could be more successful if it was delivered in combination with other requirements (70%, n=31) and some interviewees have used the AAMR in conjunction with other measures such as a Rehabilitation Activity Requirement (RAR), allowing the offender to address in more depth their alcohol use and the impact on their lives. One area where it was felt that the AAMR would benefit from being delivered in combination with another requirement is when it was imposed for a domestic abuse offence. Overall, “The vast majority of alcohol-relating offending coming before our courts is committed by alcohol-dependent people. Therefore the AAMR is not very useful to us.” Survey Respondent “I have sat on cases where AAMR would be ideal but once put off for reports it’s a different sentencing bench. I think benches should be reminded more often about its use.” Survey Respondent 18 stakeholders were of the opinion that the AAMR should be used in domestic abuse cases (70%, n=31), especially where there was a clear correlation between the drinking of alcohol and the domestic violence. However, it was strongly noted that in such instances, the AAMR would need to be alongside a specialised programme, such as ‘Building Better Relationships’, to specifically address the thinking and behavioural causes of domestic abuse offending and the ongoing risk of further abuse and harm. Within the pan-London roll out, a feasibility study was conducted by Standing Together Against Domestic Violence, exploring the use of AAMR with domestic abuse perpetrators. The study provided learning about how to implement this work, particularly around addressing the management of such cases to ensure the safety of victims. However, uptake of AAMR with domestic abuse offenders was incredibly low (N=4) despite the fact that alcohol was a factor in 33% of the domestic abuse cases appearing at the pilot Courts. It is unclear as to why uptake of the AAMR was low with domestic abuse perpetrators, although anecdotally it is likely to be because of PSR writers not being able to contact the survivor (part of the protocol) within the five day time scale and a lack of understanding about the use of AAMR for domestic abuse perpetrators. Further testing would be needed before any meaningful conclusions can be reached about the impact of AAMR on domestic abuse cases. Potential Effects on Stakeholders The AAMR has the potential to impact many people, and it is important to also consider the effects that the AAMR may have on stakeholders’ workloads. Overall interviewees and survey respondents indicated that AAMR had not made a difference to workloads22 despite comments that it was labour intensive in terms of paperwork. However, many participants did suggest that the AAMR was a “useful addition” to their role (82%, n=36), reflecting previous comments noted in the pilot and interim report. Indeed, other stakeholders spoke positively about the AAMR, noting that this requirement “deals with the root cause of offending”, helping people to make the link between alcohol use and alcohol related offending, as well as sending a clear message to their peers about the potential consequences of their actions. It was raised by some of the Probation Officers interviewed that they would have been interested in receiving updates about the AAMR throughout the project; for example, so they could understand the usage across London, how many offenders were completing the AAMR, breaches and good news stories from people who have undertaken the order. It was also thought that this would have enabled them to provide the Courts with additional information and a pan-London view that would inform their sentencing. It was also mentioned that once the offender has finished with the stand-alone AAMR Orders, then the Probation Officer does not have any further contact with the offender, so it is hard to comment specifically on any impact that the AAMR may have had. In order for the AAMR to work effectively, it is essential that there are good working partnerships between agencies such as CRC, NPS, the judiciary and EMS. However, throughout the pan- 22 Responses to the questions: “The AAMR has had no impact on my work “– Agree = 45%, n=20; Disagree = 45%, n=20, “The AAMR has increased my workload” – Agree = 9%, n=4; Disagree = 64%, n=28 and “The AAMR has decreased my workload” – Agree = 9%, n=4; Disagree = 48%, n=21. 19 London programme, few stakeholders reported that the AAMR had enabled them to develop relationships with new partners (20%, n=9) or improve relationships with existing partners (20%, n=9). This is a significant reduction (p<.05) from the interim report findings (49% and 55% respectively), which may reflect the different respondents in this current survey (mostly Magistrates) compared to previously when there was a much wider spectrum of participants. Despite new and existing partnerships not being developed drastically over the two years, there was no indication that relationships had deteriorated or had a negative effect on the implementation and delivery of the AAMR. Potential Effects on Offenders In addition to recognising the effects of the AAMR on their own workloads, stakeholders were also able to identify the potential impact that the AAMR may have on the offender. The AAMR requires offenders to be abstinent from alcohol for a set duration of time and it was generally perceived amongst interviewees and survey respondents that compulsory sobriety would have a positive impact on offenders. In particular it was believed that the AAMR would reduce alcohol consumption (57%, n=25) as well as reoffending (36%, n=16) as AAMR is usually imposed for low level offences; however, given these are key aims of the AAMR, it is unclear why more stakeholders do not feel the AAMR would reduce alcohol use or reoffending. The benefits of the AAMR were highlighted, including the fact that it allows offenders to continue with their lives, whereas the alternative of Unpaid Work can sometimes be disruptive. Indeed, interviewees felt that offenders were often pleased to be offered an alternative that was different from either an electronic Curfew or hours of Unpaid Work. However, there was some concern with a few interviewees stating their scepticism of the requirement as they felt that the type of people receiving AAMRs were unlikely to reoffend anyway, with one describing the AAMR as a “middle class order” and that they do not think there is the stigma attached to the AAMR that may already exist with Curfew or Unpaid Work requirements. Whilst the AAMR has been used as a punitive measure imposed by the Courts, interviewees and survey respondents felt that there were potential rehabilitative benefits that may derive from compulsory sobriety for a period of time. The majority of survey respondents indicated the AAMR would help in others area of the offenders lives (64%, n=28) such as work, family, short term health benefits and highlighting to the offender that alcohol use can become an issue. Indeed, it was felt being on the AAMR would help people play a more positive role in society (52%, n=23). There was an overall feeling that the AAMR would be most beneficial to younger people who regularly consume alcohol on a social basis with very few previous convictions, as the AAMR is an opportunity to provide offenders with information on alcohol education and signposting to alcohol interventions and support groups so they know where and how they can access support if desired. Additionally, stakeholders’ thoughts turned to the wider landscape noting that the potential rehabilitative elements may extend to substance users, and a “The general attitude is that they will do whatever it takes and they do see the positives of having the tag when you outline it to them.” Interviewee “it’s a really good additional option and I particularly like that it is both punitive (i.e. alcohol abstinence) and rehabilitative (there is counselling offered).” Interviewee 20 device similar to the AAMR specifically to monitor compulsory drug abstinence could be designed. Offenders who have been subjected to the tag and AAMR have been given the opportunity to express their views about the requirement – their views are heard next. Using the AAMR: Process Learning from Offenders To gain a better understanding of the impact of the AAMR tag on offenders, at the time of fitting and removing the tag, offenders were asked to complete a short survey. In total, over the two years of the project, 412 offenders completed the entry survey when the tag was fitted, and 407 offenders also completed the exit survey when the tag was removed (see Appendix G for a full breakdown of responses). Although it cannot be guaranteed that those offenders who responded to the entry survey also completed the exit survey, this large sample still provides a unique voice to the discussion. Entry Survey – Expectations and Concerns Overall, the majority of tags were fitted for the first time within the offender’s home (99%, n=408), and only on three occasions was the offender initially tagged within a court building – a pilot occurring in Westminster Magistrates Court or Bromley Magistrates Court. One respondent was having the tag refitted following breach proceedings. At the start of their Orders, the majority of those surveyed reported having a good23 relationship with their close family (93%, n=383) and friends (94%, n=386), to be in good physical health (88%, n=363), have suitable accommodation (86%, n=356), and have a relatively good sense of well-being (75%, n=31124). Whilst 62% (n=257) stated that they have a job they enjoy, a third of offenders agreed that their current financial situation is difficult (36%, n=150), possibly reflecting the experiences of many in the current economic climate. Despite receiving an AAMR as part of their court order, less than half of offenders agreed that drinking alcohol has a negative effect on their life in general (43%, n=178) and the majority of offenders did not feel that socialising with their friends caused them problems (62%25, n=254). There were mixed views around whether offending behaviour caused problems, with just over a third stating that it did not (39%, n=161), whilst a further third agreed that it did (39%, n=160). However, a strong majority understood why they had received the AAMR (92%, n=379), how to comply (96%, n=396), and felt confident they would successfully complete the requirement (97%26, n=398) – this is reflective of the high compliance rate seen with this requirement (94%). 23 Based on collated ‘Agree’ and ‘Strongly Agree’ responses. 24 Measured by those who ‘Agreed’ or Strongly Agreed’ to the statement ‘I am happy most of the time’. 25 Based on collated ‘Strongly Disagree’ and ‘Disagree’ responses. 26 Based on collated ‘Fairly Confident’ and ‘Very Confident’ responses. 21 Although not the original intention of the AAMR, wearing the tag does provide some offenders with the chance to reflect on their lives and make changes, as highlighted by the stakeholders. When first receiving the tag, offenders were broadly optimistic that the AAMR could improve aspects of their lives (‘life in general’: 56%, n=230), including specific facets such as relationship with family (46%, n=188), friends (38%, n=156), current financial situation (48%, n=196), physical health (57%, n=234), mental well-being, (50%, n=206) and their offending behaviour (53%, n=219). This may be indicative of the recognition of AAMR’s ability to infiltrate other aspects of people’s lives, beyond their offending behaviour. Whilst potential benefits from compulsory sobriety were recognised, offenders also felt being tagged may impact their lives negatively in some aspects. Their main concerns were around what their friends and family would think of the alcohol tag (39%, n= 161 were worried what others would think) and offenders thought it would make their ability to socialise worse (21%, n=87). Concerns were raised about the stigmatisation offenders may feel from having to wear the tag, particularly extending to employment or education, with 10% (n=43) suggesting that wearing the tag would make their employment situation worse. Furthermore, there were practical concerns raised about the actual tag itself, with offenders frequently commenting on the size and weight of the tag and the disruption it would cause to their everyday lives, particularly around health and well-being, travel, participating in exercise and having to ensure they return home at set times to register on the data box. Exiting the AAMR – were concerns realised? Once an offender had completed the AAMR, the tag was removed and the offenders were asked to participate in another short survey. Of those who agreed to complete the survey, nearly all tags were removed within the offender’s home (n=405), with the remaining two tags removed at Westminster Magistrates Court (n=2). At the end of their AAMRs, offenders reported having a good sense of well-being (70%, n=286), reflecting how they felt at the start of the requirement. Overall, they were positive about their lives, although this was to a significantly lesser extent than when they initially received the tag (p<.05). Offenders still reported having a good relationship with family (76%, n=311*27) and friends (78%, n=318*), suitable accommodation (75%, n=304*) and good physical health (75%, n=307*). There was no significant difference between offenders indicating they were happy with 27 *=Result significant at 95% level of confidence. “I think this should help me to improve my overall life as a whole.” Offender “…I am a student and it is so big on my leg and I am worrying what my fellow students and Lecturers would say? This is another stress of my life.” Offender “The device is too large and may cause me difficulties.” Offender 22 their job (59%, n=242), and offenders felt that their current financial situation was manageable (36%, n=14728). Other changes also occurred around offenders’ behaviour, with significantly fewer offenders (p<.05) disagreeing that ‘going out socialising with my friends causes me problems’ after being subject to the AAMR (51%, n=208*) suggesting that they have had more positive experiences socialising whilst abstaining from alcohol and reflected by significantly fewer offenders indicating that drinking alcohol has a negative effect on their life (33%, n=133*). Furthermore, just over a quarter now felt that their offending behaviour caused them problems (28%, n=112), significantly fewer than at the start of the requirement, which may indicate they have moved away from criminal activity and sought a more prosocial journey. When asked whether they felt the AAMR had made things better, worse or had no impact, the trend in the data suggested that the AAMR had no negative effects (i.e. had not made anything worse). At the start, offenders noted that the AAMR may have an impact on their relationships with their families and their financial situation, when in fact it appears to have made no impact. However, some expectations were met, with offenders feeling their life in general had got better following the AAMR, as well as their physical and mental well-being and offending behaviour (Table 2). Indeed, some offenders spoke positively about the AAMR, highlighting how it had made an impact to them particularly around their alcohol use, indicating that the AAMR can achieve one of its predominant aims. 28 Measured by those who ‘Disagree’ or ‘Strongly Disagreed’ to the statement ‘My current financial situation is difficult’. “Having the tag on my leg was the best thing for me and now I can finally say I’m totally free from alcohol.” Offender “It’s good because it made me look at alcohol in another light.” Offender 23 Table 2. Entry and Exit Survey results There continued to be ongoing concern around some of the practicalities of the tag itself. In particular, three-quarters of offenders reported that the tag was uncomfortable to wear for the duration of the requirement (74%, n=303), with many stating that it was too large and the design was flawed. Additionally, offenders commented that it had impacted on other practicalities of their daily lives, such as being unable to go swimming, bathe or use certain products such as perfumes/aftershave. There also continued to be concern around the stigmatisation of wearing a tag that were initially raised, with 44% (n=181) reporting being concerned about what their friends and family thought of the tag. Indeed, offenders spoke about the difficulties they had in trying to hide the tag from friends and colleagues through purchasing new clothes and the expense they incurred from having to do so. Overall, the offenders perceived the AAMR positively with around half reporting that it has had a positive influence in their health and offending behaviour. However, practical concerns were raised about the imposition of having to wear the tag itself, with many stating that the tag was too large and uncomfortable, and concerns that this may lead to stigmatisation. I think that being on the AAMR will make my: Worse No Impact Better Life in general 10% 24% 56% Relationship with family 5% 40% 46% Relationship with friends 8% 45% 38% Money/current financial situation 6% 36% 48% Offending behaviour 1% 27% 53% Physical health 4% 31% 57% Mental Wellbeing 8% 33% 50% Housing situation 2% 57% 29% Employment situation 10% 48% 26% Ability to 'go out'/socialising 21% 44% 24% Educational situation 1% 49% 18% I think that being on the AAMR has made my: Worse No Impact Better Life in general 10% 32% 44% Relationship with family 6% 43% 36% Relationship with friends 8% 46% 32% Money/current financial situation 8% 42% 36% Offending behaviour 2% 35% 44% Physical health 5% 32% 50% Mental Wellbeing 8% 35% 43% Housing situation 3% 57% 26% Employment situation 11% 52% 23% Ability to 'go out'/socialising 19% 41% 26% Educational situation 3% 49% 21% Exit Survey Entry Survey 24 4. Discussion In 2014, the AAMR was introduced as a pilot concept to address alcohol related offending, by imposing compulsory abstinence from alcohol for a set period of time. Following the success of the pilot, a two year programme was funded and from April 2016, in a phased approach, the AAMR was rolled out across all of London, enabling every court to have this sentencing power. At the end of the two years, there is now the opportunity to reflect on the learning gained from this programme and the effectiveness of the AAMR on a larger, more complex scale. This report brings together the evidence around compulsory sobriety electronic monitoring, particularly around how it has performed and details a process evaluation generating learning through the views and experiences of stakeholders involved in delivery and offenders who were sentenced to the AAMR. Positively the AAMR has been used widely throughout London, gaining support from both the judiciary and Probation staff, who have welcomed a tool tailored to specifically addressing alcohol related offending behaviour. There appeared to be a good understanding of the technology and requirement itself, with little self-reported impact on professionals’ workload. The Year 1 interim report noted that AAMR training was well attended and satisfactory, and whilst some stakeholders still held this view, conversely others reported that being new to their role, they had not received any training, instead learning about the AAMR from colleagues. It was clear that such individuals would have welcomed on-going or refresher training throughout the second year of the programme. In the second year of service, the AAMR has settled well into Probation report writers’ “toolbox” and has been, on the whole, recommended appropriately. However the requirement is not without its frustrations and stakeholders frequently commented on the limitations of the eligibility criteria, particularly that offenders must reside within a London borough and score appropriately on the alcohol audit score, excluding many offenders from this sentencing option. Overall this has meant that the AAMR has not been recommended as often as some stakeholders would have liked. One key aim of the AAMR was to prevent people from committing further offences and whilst it was recognised that the AAMR is designed as a punitive measure, stakeholders strongly felt that the requirement should also be viewed as a rehabilitative requirement, as it provides the opportunity to reflect on one’s behaviour particularly their alcohol consumption and offending behaviour. It is unclear at this time what actual effects the AAMR will have on offenders’ reoffending behaviour, and research would suggest that the behavioural effects may be short- lived beyond the removal of the tag (Axdahl, 2013); however, there is still a general perception that the AAMR offers more than a punishment to the offender. This view is reflected by the offenders themselves, who were often optimistic about the requirement and felt it would have a positive impact on their health, wellbeing and offending behaviour. Much of the learning encountered over the two years of the AAMR being available pan-London reflects the initial pilot research. Challenges were seen at the end of Year 1 with scaling up the programme from a small pilot (four boroughs) to a pan-London approach – challenges such as delays in getting offenders tagged. These have clearly not been resolved, and it remains unclear 25 why such delays occur. Although a trial of tagging the offender at court was started, due to mitigating circumstances around staffing and infrastructure issues, very few offenders were tagged in this way, which meant this had little impact on addressing the delays. Other challenges previously noted were around partnership working. Whilst stakeholders here did not recognise improved relationships with partners or opportunities to build new partnerships, this also does not appear to have caused issues in the implementation and delivery of the AAMR and may in fact reflect the current landscape of offender management in London. Currently the AAMR has only been available within London-based courts for offenders who reside within London boroughs. However, there was strong support for the AAMR to be rolled out nationally (82%, n=36). A variety of reasons for a UK wide compulsory abstinence requirement were emphasized, including that it would make sentencing more consistent, and allow all Courts to impose the requirement regardless of the resident location of the offender – an issue that is perceived as causing difficulties in some London courts currently. Another suggestion to improve the service going forward would be to enhance the technology associated with the AAMR. That is, improve the electronic monitoring tags to ensure they are smaller and more comfortable to wear as well as make the tags themselves fully waterproof. Many offenders complained not only of the size and shape of the unit, but also that they could not take a bath or swim, which impacted on their daily lives. Stakeholders thoughts have also turned to the wider landscape noting that there needs to be a better way to address drug related offending and it was proposed that a similar device to the AAMR is required – a tag that can detect low level drug use to enable a requirement dictating compulsory drug abstinence to address low level drug use and related offending behaviour. The pan-London rollout of the AAMR was funded specifically for two years to test the wider implementation and impact of the AAMR. However, due to the change in central government policy regarding a national rollout, there are currently no plans to extend the project further. Therefore, as of June 2018, the AAMR will cease to be an imposable requirement by the Courts. All offenders will therefore have completed the AAMR by October 2018 at the latest. This report sits as part of a wider, holistic evaluation to test the impact of the pan-London AAMR programme. Further research to explore proven reoffending and cost benefit of the AAMR will be included in the final evaluation report in Spring 2019. 26 References Alessi, S.M., Barnett, N.P. & Petry, N.M. (2017). Experiences with SCRAMx alcohol monitoring technology in 100 alcohol treatment outpatients. Drug & Alcohol Dependence, 178, 417-424. Averill, F., Brown, T.G., Robertson, R.D., Tchomgang, A., Berbiche, D., Nadeau, L. & Ouimet, M.C. (2018). Transdermal alcohol monitoring combined with contingency management for driving while impaired offenders: a pilot randomised controlled study. Traffic Injury Prevention. Advance Online Publication. Axdahl, L. (2013). Analysis of 24/7 Sobriety Program SCRAM Participant DUI Offense Recidivism. Mountain Plains Evaluation, LLC. USA: Salem. Bergen, G., Pitan, A., Shults, R, Qu S. & Sleet, D. (2012). Current evidence on publicised sobriety checkpoint programmes: are they still effective? Injury Prevention, 18, A45-A46. Blais, E. & Dupont, B. (2005). Assessing the capability of intensive police programmes to prevent severe road accidents. British Journal of Criminology, 45(6), 914-937. Cattell, J., Kenny, T., Lord, C. & Wood, M. (2014b). Community Orders with Punitive Requirements: Results from the Offender Management Community Cohort Study. London: Ministry of Justice. Department for Transport (2016). Reported road casualties in Great Britain: Estimates for accidents involving illegal alcohol levels: 2014 (final) and 2015 (provisional). London: Department for Transport. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/543627/rrc gb-drink-drive-final.pdf Dougherty, D., Charles, N., Acheson, A., John, S., Furr, R. & Hill-Kapturczak, N. (2012). Comparing the Detection of Transdermal and Breath Alcohol Concentrations during Periods of Alcohol Consumption Ranging from Moderate Drinking to Binge Drinking. Experimental and Clinical Psychopharmacology, 20(5), 373 – 381. Farrington, D.P. (2005). Integrated Developmental and Life-Course Theories of Offending. New Brunswick, NJ: Transaction. Fell, J.C., Tanenbaum, E. & Chelluri, D. (2018). Evaluation of a combination of community initiatives to reduce driving while intoxicated and other alcohol-related harms, Traffic Injury Prevention, 19, 176-179. Flango, G. & Cheesman, F. (2009). The effectiveness of the SCRAM alcohol monitoring device. Drug Court Review, 6, 2, 109-134. Hobson, Z., Dangerfield, B. & Harrison, A. (2017). Alcohol Abstinence Monitoring Requirement. The Pan London roll out: A Review of Process and Performance from Year 1. London: MOPAC Evidence & Insight. 27 Kerns, T. (2017). Effectiveness of an ignition interlock device in reducing alcohol-impaired driving recidivism and alcohol-impaired motor vehicle crashes in Maryland. Baltimore: University of Maryland https://archive.hshsl.umaryland.edu/handle/10713/6751. Kilmer, B., Nicosia, N., Heaton, P. & Midgette, G. (2013). Efficacy of frequent monitoring with swift, certain, and modest sanctions for violations: Insights from South Dakota’s 24/7 Sobriety Project. American Journal of Public Health, 103(1), e37–e43. Kubas, A. & Vachal, K. (2017). Does the 24/7 Sobriety Program positively influence driver behaviours in North Dakota? North Dakota: Upper Great Plains Transportation Institute. McSweeney, T. (2015). Calling time on alcohol-related crime? Examining the impact of court- mandated alcohol treatment on offending using propensity score matching. Criminology and Criminal Justice, 15(4), 464-483. Ministry of Justice (2015). Prison and Probation Performance Statistics 2014 to 2015. London: Ministry of Justice. Ministry of Justice (2015b). Criminal Justice System Statistics Quarterly: December 2014. London: Ministry of Justice. Ministry of Justice (MOJ) & Public Health England (2017). The effect of community-based drug and alcohol treatment on reoffending. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_d ata/file/674858/PHE-MoJ-experimental-MoJ-publication-version.pdf Ministry of Justice (MOJ) (2017) Criminal Justice Statistics quarterly, England and Wales, 2017. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_d ata/file/707935/criminal-justice-statistics-quarterly-december-2017.doc.pdf Nicosia, N., Kilmer, B. & Heaton, P. (2016). Can a criminal justice alcohol abstinence programme with swift, certain and modest sanctions (24/7 Sobriety) reduce population mortality? A retrospective observational study. The Lancet Psychiatry, 3(3), 226-232. Office for National Statistics (ONS) (February 2016). Overview of violent crime and sexual offences, Nature of Crime Table 3.2. Office for National Statistics (ONS) (2018). The nature of violent crime in England and Wales: year ending March 2017. https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/thenatureo fviolentcrimeinenglandandwales/yearendingmarch2017#what-is-happening-to-trends-for- different-types-of-violent-crime. Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. London: MOPAC Evidence and Insight. Public Health England (2016). The Public Health Burden of Alcohol and the Effectiveness and Cost-Effectiveness of Alcohol Control Policies, An evidence review. London: Public Health England 28 Public Health England (2017). Adult substance misuse statistics from the National Drug Treatment Monitoring System (NDTMS). London: Public Health England. Public Health England (2018). Alcohol and drug prevention, treatment and recovery: why invest? https://www.gov.uk/government/publications/alcohol-and-drug-prevention- treatment-and-recovery-why-invest/alcohol-and-drug-prevention-treatment-and-recovery- why-invest. Roth, R., Marques, P. & Voas, R. (2009). A note on the effectiveness of the house-arrest alternative for motivating DWI offenders to install ignition interlocks. Journal of Safety Research, 40(6), 437–441. Vanlaar, W.G., Hing, M.M. & Robertson, R.D. (2017). An evaluation of Nova Scotia’s alcohol ignition interlock program. Accident Analysis & Prevention, 100, 44-52. 29 Appendices Appendix A: Stakeholder Survey Respondents Full breakdown of respondents who completed the Stakeholder Survey (N=44). Job Role No. of respondents % Magistrate 24 55% District Judge 3 7% Probation Officer 11 25% Other 6 14% Total 44 100% Organisation No. of respondents % Magistrates / Crown Court 14 32% National Probation Service 1 2% CRC 11 25% Her Majesty's Courts and Tribunal Service 12 27% Ministry of Justice 4 9% Not recorded 2 5% Total 44 100% Local Justice Area Represented No. of respondents % North London LJA 4 9% North West LJA 14 32% East London LJA 2 5% Central LJA 2 5% South East LJA 6 14% South London LJA 4 9% South West LJA 4 9% West London LJA 4 9% Not Recorded 2 5% Other 2 5% Total 44 100% 30 Appendix B: Interviewee Respondents Full breakdown of those who agreed to be interviewed (N=24). Organisation Job Role No. of Interviewee's NPS Probation Officer / Senior Probation Officer / Court Officer 13 Probation Prosecutor 1 CRC Probation Officer / Senior Probation Officer 4 Courts Deputy Justice Clerk 2 Legal Team Manager 2 EMS / AMS / SCRAM Systems Manager 2 Total 24 31 Appendix C: Breakdown of Courts that have imposed AAMRs. Magistrates Court No. of AAMRs imposed % Croydon 144 14% Highbury Corner 129 13% Uxbridge 99 10% Camberwell Green 93 9% Barkingside 80 8% Westminster 75 7% Thames 58 6% Bromley 50 5% Hendon 47 5% Wimbledon 44 4% Bexley 28 3% Hammersmith 23 2% Willesden 21 2% Ealing 20 2% Stratford 5 0% City of London 3 0% Feltham 3 0% Total 922 91% Crown Court No. of AAMRs imposed % Isleworth CC 18 2% Croydon CC 14 1% Central Criminal Court 9 1% Wood Green CC 8 1% Blackfriars CC 7 1% Harrow CC 7 1% Romford MC 7 1% Southwark CC 7 1% Woolwich CC 6 1% Inner London CC 4 0% Snaresbrook CC 4 0% Kingston CC 1 0% Total 92 9% 32 Appendix D: Breakdown of Offences This table details the offence type for which an AAMR was imposed between April 2016 and March 2018. Offence Type No. of offences % Damage/Theft Burglary 11 1% Criminal Damage 69 7% Theft 17 2% Unauthorised taking of a motor vehicle 9 1% Driving Driving or attempting to drive whilst unfit through drink/drugs 240 24% Failing to provide specimen 27 3% Other driving 24 2% Drugs Failure to cooperate (drugs) 3 0% Supply/Possession of drugs 7 1% Harassment Harassment 18 2% Racially aggravated harassment 34 3% Threatening words or behaviours 36 4% Public Order Drunk and disorderly conduct 6 1% Offences against Public Order 7 1% Other offences 27 3% Sexual Exposure 6 1% Sexual Assault 17 2% Violence Assault (beating, common assault, ABH, GBH) 338 33% Assault Police Officer 72 7% Possession of a weapon 36 4% Resisting/Obstructing a Police Officer 4 0% Other Violence 6 1% Total 1,014 100% 33 Appendix E: Offender Demographics Offender ethnicity Offender Ethnicity No. of AAMR Offenders % Asian or Asian British (including Bangladeshi, Chinese, Indian, Pakistani, Other) 105 10% Black or Black British (including African , Caribbean, Other) 149 15% Mixed (including White & Asian, White & Black African, White & Black Caribbean, Other) 38 4% White (including British, Welsh, Scottish, Northern Irish, Irish, Gypsy or Irish Traveller, Other) 555 55% Other Ethnic Group 28 3% Refusal 13 1% Not recorded 126 12% Total 1,014 100% Offender Age 0 10 20 30 40 50 60 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 63 66 68 No. of AAMR offenders Offender Age 34 Appendix F: Offender Group Reconviction Scale (OGRS) 2 Year scores: OGRS Year 2 No. of AAMR offenders % Very low (0 - 24%) 360 37% Low (25% - 29%) 305 32% Medium (50% - 74%) 209 22% High (75% - 89%) 80 8% Very high (90% +) 10 1% Total 964 100% 35 Appendix G: Offender Entry and Exit Survey Results Entry Survey Results (N=412) Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Don't Know I have a good relationship with my close family 5 1% 11 3% 12 3% 109 26% 274 67% 1 0% I have a good relationship with my friends 2 0% 10 2% 9 2% 142 34% 244 59% 5 1% I have good physical health 7 2% 14 3% 26 6% 167 41% 196 48% 2 0% I have a nice place to live 5 1% 15 4% 29 7% 138 33% 218 53% 7 2% I have a job which I enjoy 11 3% 30 7% 42 10% 107 26% 150 36% 72 17% Going out socialising with my friends causes me problems 107 26% 147 36% 77 19% 42 10% 24 6% 15 4% My current financial situation is difficult 45 11% 93 23% 102 25% 99 24% 51 12% 22 5% My offending behaviour cases me problems 75 18% 86 21% 60 15% 116 28% 44 11% 31 8% Drinking alcohol has a negative effect on my life in general 60 15% 76 18% 76 18% 114 28% 64 16% 22 5% I am happy most of the time 10 2% 21 5% 58 14% 159 39% 152 37% 12 3% Agree Disagree Blanks I understand why I received the AAMR 379 92% 20 5% 13 3% I understand what I must do to comply with the AAMR 396 96% 2 0% 14 3% I am worried about what my friends and family will think of the AAMR tag 161 39% 209 51% 42 10% Not at all confident Not very confident Fairly Confident Very Confident Blanks How confident are you that you will successfully complete the AAMR 2 0% 5 1% 46 11% 352 85% 7 2% 36 I think being on the AAMR will make my: Worse No Impact Better N/A Life in general 42 10% 99 24% 230 56% 41 10% Relationship with family 20 5% 165 40% 188 46% 39 9% Relationship with friends 31 8% 184 45% 156 38% 41 10% Money/current financial situation 25 6% 150 36% 196 48% 41 10% Offending behaviour 5 1% 110 27% 219 53% 78 19% Physical health 18 4% 127 31% 234 57% 33 8% Mental Wellbeing 31 8% 136 33% 206 50% 39 9% Housing situation 10 2% 235 57% 118 29% 49 12% Employment situation 43 10% 198 48% 108 26% 63 15% Ability to 'go out'/socialising 87 21% 181 44% 97 24% 47 11% Educational situation 3 1% 202 49% 76 18% 131 32% Exit Survey Results (N=407) Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree Don't Know I have a good relationship with my close family 5 1% 14 3% 39 10% 110 27% 201 49% 38 9% I have a good relationship with my friends 3 1% 9 2% 41 10% 124 30% 194 48% 36 9% I have good physical health 7 2% 15 4% 44 11% 132 32% 175 43% 34 8% I have a nice place to live 7 2% 10 2% 49 12% 124 30% 180 44% 37 9% I have a job which I enjoy 15 4% 20 5% 77 19% 103 25% 139 34% 53 13% Going out socialising with my friends causes me problems 76 19% 132 32% 101 25% 42 10% 18 4% 38 9% My current financial situation is difficult 52 13% 95 23% 122 30% 80 20% 24 6% 34 8% My offending behaviour cases me problems 61 15% 100 25% 96 24% 78 19% 34 8% 38 9% Drinking alcohol has a negative effect on my life in general 45 11% 94 23% 101 25% 83 20% 50 12% 34 8% I am happy most of the time 7 2% 21 5% 62 15% 149 37% 137 34% 31 8% 37 Agree Disagree I was worried what my friends and family thought of the alcohol tag 181 44% 226 56% The alcohol tag felt comfortable to wear 104 26% 303 74% The AAMR guidance document I received was useful 285 70% 122 30% I think that being on the AAMR has made my: Worse No Impact Better N/A Life in general 39 10% 130 32% 179 44% 59 14% Relationship with family 26 6% 175 43% 148 36% 58 14% Relationship with friends 32 8% 187 46% 132 32% 56 14% Money/current financial situation 31 8% 172 42% 147 36% 57 14% Offending behaviour 8 2% 144 35% 181 44% 74 18% Physical health 21 5% 132 32% 203 50% 51 13% Mental Wellbeing 33 8% 141 35% 177 43% 56 14% Housing situation 12 3% 230 57% 105 26% 60 15% Employment situation 43 11% 210 52% 95 23% 59 14% Ability to 'go out'/socialising 78 19% 165 41% 105 26% 59 14% Educational situation 13 3% 199 49% 84 21% 111 27% 1 Alcohol Abstinence Monitoring Requirement South London Pilot Indicative Impact Report April 2017 Zoё Hobson, Adele Harrison & Paul Dawson MOPAC Evidence and Insight 2 Contents 1. Background........................................................................................................................... 3 2. Key Findings ......................................................................................................................... 4 3. Methodology ........................................................................................................................ 5 3.1 Creating a matched control .............................................................................................. 5 4. Results .................................................................................................................................. 8 4.1 Offending on the Order ..................................................................................................... 8 4.2 Offending in the first 12 months since Order imposed ...................................................... 8 4.3 Offending after completion of the Order ......................................................................... 11 5. Conclusion .......................................................................................................................... 13 Appendices ............................................................................................................................. 15 3 Alcohol Abstinence Monitoring Requirement South London Pilot Indicative Impact Report 1. Background In 2011 the Mayor’s office secured legislation to allow for the introduction of a new sentencing power, the Alcohol Abstinence Monitoring Requirement (AAMR) to tackle the significant problem of alcohol related offending in London. The AAMR gives the Judiciary the statutory power to stop an offender drinking alcohol (Compulsory Sobriety), where their offence is alcohol related. The AAMR involves fitting a tag to the offender’s ankle and monitoring their alcohol consumption for up to 120 days. When this is not complied with, the offender will be breached and punished further. MOPAC commissioned a compulsory alcohol sobriety trial in South London which aimed to test the monitoring equipment, understand the take up, and test compliance and enforceability of the requirement. A 12 month pilot period commenced 31 July 2014, over which time 113 AAMR Orders were imposed by the Courts. The AAMR pilot1 had a final completion compliance rate of 95%2 which compares favourably with other similar orders - analysis by the NPS in 2014 estimated a compliance rate of 61% for other community based Orders it managed with the Community Rehabilitation Company (CRC)3. Variation was also seen with different requirements - 82% of offenders completed their Unpaid Work Orders successfully4 in London compared to Alcohol Treatment requirements (80%) and Drug Rehabilitation requirements / Drug Testing and Treatment Orders (67%)5. The AAMR was received well, particularly by the judiciary and professionals, who recognised the AAMR as an important ‘tool in their box’. This can in part be attributed to the strength of the design and implementation of the programme. There were clear toolkits and training provided, effective partnership working and a project management team in place with relevant experience in this area. The effectiveness and certainty provided by the technology, as well as a strong understanding of the aims of the pilot and how the AAMR works in practice amongst both offenders receiving the order and stakeholders involved in its delivery also helped. In addition, there were a number of associated positive consequences of the pilot, including but not limited to; the period of abstinence gave offenders a ‘pause’ in their drinking; it also provided time for reflection of their alcohol consumption and the impact it has on offending behaviour, work and relationships; and an opportunity was provided for offenders to break their cycle of routine drinking. The AAMR was also used as a ‘teachable moment’ in some instances with products such as literature and advice tailored by the service providers in order to help offenders further. The technology working as intended and the strong implementation resulted in a pan London extension of AAMR, launched on 1st April 2016. The initial MOPAC research report6 explored process and performance, but it was not possible to assess any impact of AAMR on offending due to a limited follow up time and relatively small numbers. Given there is now a longer follow up period, this report focuses on an indicative analysis of offending behaviours of the AAMR pilot cohort compared to a comparison group. These findings sit as part of a wider, holistic evaluation around the roll out of the AAMR across London. Additional research is currently being conducted around a process evaluation, cost benefit analysis and performance metrics which will be reported on in summer 2017. Following this, in 2018, once enough time has elapsed; proven reoffending will be explored to demonstrate the potential impact of AAMR. 1 Caution needs to be applied when interpreting the completion and compliance rate of AAMR – this was a pilot study with a small sample size enabling the project manager to provide some assurances that the large majority of the AAMRs were enforced when failures to comply arose. This may not necessarily be the case with other Orders and requirements. 2 The AAMR pilot had a compliance rate of 92%, based on the number of cases (n=9) who were returned to court and convicted on breaching their AAMR as a proportion of all cases imposed. Of these 9, 5 had their AAMR revoked and failed to complete, and the remaining 4 completed their AAMR following their return to court. This gives a final completion/compliance rate of 95%. Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC 3 However there are caveats to be considered such as a direct ‘like for like’ comparison is not possible due to different offence types, offender characteristics, breach processes and the length of the orders themselves. 4 These figures should be caveated however as the AAMR project manager recently reviewed other requirements on Orders (i.e. UPW), finding there were numerous occasions when breaches were not enforced and cases were simply closed, which may distort the actual compliance rate. 5 Ministry of Justice (2015). Prison and Probation Performance Statistics 2014 to 2015. London: Ministry of Justice. 6 Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC 4 2. Key Findings  Indicative analysis was conducted on the pilot AAMR cohort (n = 109) and a well matched comparison group (Control cohort; n = 250) to explore offending behaviours over three timeframes.  The Control cohort consisted of offenders who received Community or Suspended Sentence Orders over the same time frame as the AAMR cohort, were from matched boroughs, with similar OGR Scores and of comparable age at the time Order was imposed.  Examining offending behaviours on the order, in the first 12 months from order start date and 9 months since order completion, there does not appear to be any significant differences between the AAMR and comparison group. - Whilst subject to an Order, a slightly higher proportion of the control group were either arrested (40%, n = 100 vs. AAMR: 31%, n=34) or convicted (31%, n = 77 vs. AAMR: 27%, n = 29): this was not a significant difference (and likely explained by the longer length of the comparison Order). - There was no significant difference in the proportion of individuals who went on to reoffend in the first year subsequent to their Order (calculated as offence within 12 months, with 3 months lag for convictions: AAMR: 36%, n = 40 vs. Control: 34%, n = 84). These are broadly comparable to proven reoffending rates throughout England and Wales for Community Orders and Suspended Sentence Orders (32%)7. - There was no significant difference comparing the proportions of offenders arrested in the subsequent nine months for further offences after the end of their Orders (AAMR: 30%, n = 24 vs. Control: 26%, n = 42).  The majority of reoffences for both cohorts have been for ‘Offences Against Police, Courts and Prisons’. There have been very few guilty convictions for the offence of ‘Driving whilst under the influence of alcohol’ (an alcohol related offence) since the start of the Order for either cohort.  It is acknowledged that the analysis has limitations and should be viewed as indicative given the relatively small numbers and timeframes under analysis - but the weight of evidence so far indicates that the impact upon offending behaviours of the AAMR is comparable to pre-existing requirements (e.g. Unpaid Work). This should be seen as a positive given the inherent risks of delivering innovation. The AAMR is also shorter than comparable orders raising a wider narrative around achieving the same outcomes in a shorter timeframe in terms of efficient justice and potential cost savings. This will be explored in future reports.  The next phase of AAMR research will focus upon the pan London roll out, providing a holistic appraisal of the project including performance metrics, and a process, impact and economic evaluation. The pilot cohort will continue to be tracked and new analysis completed when proven reoffending is able to be calculated. 7 Ministry of Justice (2017). Proven Reoffending Statistics: Quarterly Bulletin April 2014 – March 2015. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/585908/proven-reoffending-quarterly-bulletin.pdf 5 3. Methodology The initial MOPAC report8 explored process and performance, but it was not possible to assess impact of AAMR on offending due to a limited follow up time and relatively small numbers. Whist the numbers remain small, there is now a longer follow up period. In such – this report focuses on an indicative analysis of offending behaviours of the AAMR pilot cohort9 compared to a comparison group of 250 offenders. 3.1 Creating a matched cohort Initially, to identify a suitable matched control cohort, 1,000 offenders were identified from the National Probation Service / CRC based on the following criteria:  Sentencing took place between July 2014 and July 2015;  Offence was committed in either Ealing, Redbridge, Islington, Haringey, Tower Hamlets, Hackney, Bexley, or Kingston10;  Offender received a Community Order or Suspended Sentence Order; and  OGRS3 Score: between 6-8211. From this sample, Police National Computer (PNC) data12 was used to identify a cohort of offenders who matched most closely to the AAMR cohort based on a number of measures including:  OGRS3 Year 2 score13;  Number of previous convictions in 5 years prior to order;  Age at first conviction;  Age at order start date; and  Seriousness of offending14. For each of the above criteria, the average of the AAMR cohort was compared to the sample; those individuals who fell within the top 50% of similarity15 were selected as the Control cohort. The match between AAMR and control was strong. See the below tables that illustrate the similarities around the key matching criteria, demographics, court Orders and associated requirements16. 8 Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC 9 AAMR cohort in this report = 109 offenders. It was not possible to obtain PNC data for all 113 AAMR cases reported in the initial pilot. From this cohort, 6 Orders were returned to court in the interest of justice. However, if they received the AAMR tag and completed days wearing the tag, they have been included in the 109 cohort reported on here. 10 Boroughs were selected based on nearest neighbour calculator (containing a host of factors) to select top ten matched boroughs to each of the pilot boroughs. These were then filtered down for number of TNO’s 2014/15 and 2015/16, and average no, of re-offences per re-offender for all offences July 2013 – June 2014. 11 Taken from the range of OGRS3 scores for the AAMR pilot cohort detailed in the AAMR Pilot Evaluation Report. 12 Downloaded on 02/12/2016. 13 Offender Group Reconviction Scale version 3 (OGRS3 2 year) score uses static factors, such as age at sentence, gender, offence committed and criminal history to predict the likelihood of proven reoffending within a given time (e.g., either one or two years after starting their Community Order. This research reports the two year score). Offenders with a higher OGRS score are at greater risk of reoffending. Risk of reoffending scores can be categorised as: 0% - 24% = Very Low, 25% - 49% = Low, 50% - 74% = Medium, 75% - 89% = High. 14 Measured using a scale of offence seriousness identified by the Ministry of Justice. Mason, T., de Silva, N., Sharma, N., Brown, D. & Harper, G. (2007). Local Variation in Sentencing in England and Wales. Ministry of Justice. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/217971/local-variation-sentencing-1207.pdf 15 This meant that each individual scored 12/25 or higher. Whilst this is a higher percentage match than would necessarily be ideal, 50% was chosen to provide a robust sample size of 250 offenders. (If this score had been reduced, to create a closer matched score (e.g.: 15/25), this would have provided a sample of 92, which would not have been a robust sample size). 16 Matching criteria did not include whether the original offence was alcohol related for this indicative piece of work due to limitations with the NPS database filtering system. Retrospective analysis has since indicated that 25% of the control sample convictions were alcohol related. In future waves we shall again seek to include it if possible. None-the-less, the overall match is satisfactory. 6 Table 1: AAMR and Control cohort results for Matching Criteria (calculated at start of Order) Matching Criteria AAMR Control OGRS3 Year 2 Score (before Order started) 39% 44% Average number of convictions in 5yrs before Order started 8 10 Average age of 1st conviction 25 24 Average age Order started 33 33 Average offence seriousness (before Order started) 7 7 Table 2: Demographics of the AAMR and Control cohorts Demographics AAMR Cohort (N = 109) Control Cohort (N = 250) Gender M = 88% (n = 98) M = 100% (n = 250) Ethnicity White = 56% (n = 61) White = 50% (n = 124) BME = 44% (n = 48) BME = 50% (n = 125) Table 3: Court Order requirements AAMR Cohort Control Cohort Multiple Requirement 61% (66) 58% (146) Standalone AAMR 39% (43) N/A Standalone RAR N/A 3% (8) Standalone Supervision N/A 7% (18) Standalone UPW N/A 31% (78) Table 4: Number of requirements per offender No. of Requirements AAMR cohort Control cohort Standalone 1 requirement 39% (43) 42% (104) Multiple Requirements 2 requirements 51% (56) 44% (111) 3 requirements 9% (10) 12% (30) 4 requirements 0 2% (4) 5 requirements 0 0.4% (1) 7 Of the AAMR cohort who received a multiple requirement (n = 66), as well as the AAMR, additional requirements included17:  UPW requirement (64%, n = 42);  Supervision requirement (23%, n = 15);  Electronic Monitoring/Curfew (6%, n = 4);  RAR (12%, n = 8); and  Accredited Programme (6%, n = 4). Of the 146 offenders in the Control cohort who received multiple requirements, a combination of requirements imposed included:  Supervision requirement (77%, n = 113);  Accredited Programme (38%, n = 55);  Specified Activity (21%, n = 31); and  RAR (18%, n = 27). Three timeframes are explored in this report to explore offending behaviours during and subsequent the Order: 1) Offending on the Order = an offender is arrested (and convicted) of an offence whilst subject to a Court Order. 2) Offence in the first year since order start = an offender is deemed to have reoffended if PNC data shows that they have committed an offence within the first 12 months since receiving a Court Order, which resulted in a conviction at court within 15 months (12 + 3 months). This method differs from that used by the Ministry of Justice to calculate official reoffending rates, where an 18 month (12 + 6 months) conviction follow up period is permitted. However, in this instance, given the time frame for this analysis, reducing the follow up period to 15 months allows for a larger and all offenders from both cohorts to be included in the analysis and previous research demonstrates that a 12 + 3 months method closely tracks that of a 12 + 6 month analysis18. 3) Offence after order end = measured via arrests in the nine months subsequent to the order concluding. This timeframe was selected as it was the optimal length to track the majority of the AAMR timeframe for the longest period. 17 For some offenders, who received more than two requirements, they may have received a combination of these requirements, and therefore the proportion (%) per requirement do not equal 100%. 18 Dawson, P., Stanko, E., Higgins, A. & Rehman, U. (2011). An evaluation of the Diamond Initiative: year two findings. London. Metropolitan Police Service; London Criminal Justice Partnership. https://www.london.gov.uk/sites/default/files/slp_reducing_reoffending_board_-_may_2011_-_info_item_- _diamond_year2_final_120411.pdf 8 4. Results 4.1 Offending on the Order The first analysis was to explore offending whilst on the order. This should be caveated given that the Control cohorts Orders were on average far longer than the AAMR order (Control = 315 days, AAMR = 194 days)19. In terms of results, slightly more of the Control cohort were either arrested (AAMR = 31%, n=34; Control 40%, n=100) or convicted (AAMR 27%, n = 29; Control 31%, n = 77) whilst subject to their court Order. These differences were not statistically significant. Again, in terms of the amount of offending, there was no significant difference in the number of arrests and convictions during the Order (Control cohort 3.1 arrests/person & 2.4 convictions/person, compared to the AAMR cohort 2.8 arrests/person & 1.9 convictions/person). See table 5. Table 5: Arrests and Convictions whilst on the Order AAMR Cohort (N = 109) Control Cohort (N = 250) Arrests Convictions Arrests Convictions % (number) of offenders who reoffended 31% (n = 34) 27% (n = 29) 40% (n = 100) 31% (n = 77) Total number of reoffences 97 57 318 190 Minimum number of reoffences 1 1 1 1 Maximum number of reoffences 12 5 15 8 Average reoffences per offender 2.8 1.9 3.1 2.4 Average speed to first reoffence (days) 124 130 132 136 Range (days) 9 - 327 9 - 327 1 - 451 1 - 675 Analysis was also conducted for a variety of splits (i.e., requirement or age of order) within the data to understand any offending differences. No significant differences were found. However, this should be caveated heavily given the low numbers (see Appendix A and B). 4.2 Offending in the first 12 months since Order imposed (plus 3 months for convictions) Indicative analysis indicates that there were no differences in reoffending between AAMR and the Control cohorts in the first 12 months subsequent to the Order being imposed (AAMR = 36%, n=40; Control = 34%, n=84). Indeed, these findings indicate that both cohorts appear to be offending above their predicted one year OGRS score (Average OGRS Y1: AAMR = 26%, Control = 29%). See Table 6. There were also no significant differences comparing the proportion of individuals arrested within the first 12 months (AAMR 45%, n=49: Control 40%, n=99) or the average number of arrests (AAMR 3.6 vs. control 3.3). Figure 1. also illustrates the relatively low level of offending before and after the order. 19 The average Order length has not been calculated for the whole AAMR cohort (N = 109) because for 10 of the AAMR cohort, their Order is still ongoing, or they are subject to a warrant. When calculating the average Order length, these have been excluded as their Order data is not verified at this time. 9 Table 6: Reoffending during the 12 months (+ 3 months) since the Order was imposed AAMR reoffences Control reoffences % (number) of offenders to reoffend 36% (n = 40) 34% (n = 84) Total number of reoffences 113 236 Minimum number of reoffences 1 1 Maximum number of reoffences 17 11 Average reoffences per offender 2.9 2.8 Average speed to first reoffence 166 days 149 days Range 14 - 415 days 1 – 447 days Figure 1: Average number of reoffences pre and post start of Order To seek further understanding of the reoffending in the first 12 months since the Order started, below details a number of splits in the data analysis, including requirements of the Order, age and offence type. The small sample sizes makes robust conclusions difficult. 4.2.1 Requirement There was little difference in reoffending rates in the 12 (+3) months since the Order started despite the combination of requirements were imposed on the Order. Approximately one third of the cohorts who received either a standalone AAMR Order or standalone UPW Order reoffended (AAMR = 40%, n = 17 vs. Control = 32%, n = 25), however the standalone AAMR cohort appear to be reoffending slightly more frequently, with 4.1 convictions per offender compared to 2.9 convictions per offender for the standalone UPW cohort (see Appendix C). When a multiple requirement Order was imposed, there was no difference in number of offenders who reoffended. These were not significant differences. 10 4.2.2 Age at start of Order Younger offenders were slightly more likely to reoffend in the 12 (+3) months following the start of their Order (AAMR = 44%, n = 27 vs. Control = 34%, n = 52) but this was not a significant difference. Table 7. Reoffending in the 12 (+3) months since start of Order split by age at Order start AAMR Cohort (mean age = 33) Control Cohort (mean age = 33) Younger Age Group Older Age Group Younger Age Group Older Age Group 18 - 32 years (57%, n = 62) 33 - 63 years (43%, n = 47) 18 – 33 years (60%, n = 151) 34 – 62 years (40%, n = 99) % (no.) of offenders 44% (27) 28% (13) 34% (52) 34% (34) Total number of reoffences 73 42 151 98 Average reoffences per offender 2.7 3.2 2.9 2.9 4.2.3 Offence Type All reoffending in the 12 (+3) months since the Order started could be attributed to a variety of different offence types. For both cohorts the main reoffences were committed towards Police, Courts and Prisons (AAMR: 37%, n = 42; Control: 39%, n = 91). Other prevalent reoffence types included Theft (AAMR: 15%, n = 17; Control: 14%, n = 33) and a category known as ‘Other’ (AAMR: 22%, n = 25; Control: 18%, n = 42) which includes a variety of driving offences (e.g. Drinking Under the Influence of Alcohol, Driving with No Insurance) and Racially/Religiously Aggravated Harassment/Alarm/Distress by Words/Writing. See Appendix D for a full breakdown of offence type. When offenders reoffend with same offence type as their ‘trigger’ offence20, there were no discernible differences between the AAMR and Control cohorts, with the exception of those who previously committed offences towards Police, Courts and Prisons. Here, a larger proportion of the AAMR cohort (39%) reoffended within this offence type (Control cohort = 26%), although this was not a significant difference (see Appendix E). Specifically focusing on offences known to be related to alcohol, since the start of the Order, overall there have been very few guilty convictions for the offence of ‘Driving whilst under the influence of alcohol’21. From both the AAMR and Control cohort, only 5 offenders (in each cohort) have committed a drink related driving offence since the start of their Order. For both the AAMR and Control cohort, if their ‘trigger’ offence was for ‘Driving whilst under the influence of alcohol’ (AAMR – 38 offenders, 40 offences of drink drive; Control – 22 offenders, 23 offences of drink drive), analysis indicates that they have not reoffended since the start of their Order. 20 The ‘trigger’ offence refers to the offence that occasions the Community Order/Suspended Sentence Order. 21 Only driving whilst under the influence of alcohol was used as a measure of alcohol related offending as it is not possible to confidently identify whether alcohol was a related factor of other offences. This information is poorly recorded on MPS CRIS data. Using the trigger offence as an example (where we know the offence was alcohol related), only 25% of all Offences Against the Person were flagged as being related to alcohol. 11 4.3 Offending after completion of the Order There is a limited timeframe to track offending subsequent to Order. To develop insights, indicative analysis has been conducted tracking the number of arrests in the nine months subsequent to the order coming to conclusion. This timeframe was selected as the optimal selection as 81% (n = 80/109) of the AAMR cohort had at least nine months follow up between completion of their Order and data collection (02/12/2016). In comparison, 64% (n = 159/250) of the Control cohort also had 9 months from the completion of their Order when they could potentially reoffended (see Appendix F). Overall, there was no significant difference between the proportions of offenders arrested after they had completed their Order for the AAMR (30%, n=24) or Control (26%, n=42). However, the AAMR cohort appear to present almost double the number arrests per individual and on average were arrested over 6 weeks earlier than the Control cohort. See Table 8 and Appendix G. Table 8: Arrests in 9 months after Order was completed AAMR Cohort (N = 80) Control Cohort (N = 159) % (number) of offenders 30% (n = 24) 26% (n = 42) Total number of arrests 95 89 Minimum number of arrests 1 1 Maximum number of arrests 15 13 Average arrests per offender 3.9 2.1 Average speed to first arrest 84 days 129 days Range 2 - 217 days 10 - 262 days Analysis explored a variety of splits – including type of requirement and age at start of order. Overall – there were no significant differences, this is likely hampered by low base sizes. Table 9 presents the base sizes if attempting to look at requirement and age. Table 9: Arrests in the 9 months since Order completed split by age and requirement for AAMR cohort Standalone AAMR (mean age = 32, n = 43) AAMR Multiple Req. (mean age = 33, n = 66) Younger Age Group Older Age Group Younger Age Group Older Age Group 0 - 32 years (60%, n = 26) 33 - 63 years (40%, n = 17) 0 - 33 years (55%, n = 36) 34 – 56 years (45%, n = 30) % (no.) of offenders arrested 31% (8) 24% (4) 25% (9) 23% (7) Total number of convictions 37 29 22 12 Minimum number of convictions 2 1 1 1 Maximum number of convictions 14 15 6 4 Average convictions per offender 4.6 7.3 2.4 1.7 12 4.3.1 Offence Type After offenders had completed their Order, very few from either cohort went on to commit further offences that are directly attributable to alcohol (e.g.: Drink driving, Drunk and Disorderly), although at this level the sample size is too low to confidently draw conclusions. Whilst violence against the person accounted for nearly half of reconvictions of those who reoffended from the AAMR cohort (46%, n = 6), this reflects a very small sample and it is not possible to confidently know if these offences are related to alcohol22. Only a quarter of those who reoffended from the Control cohort (24%, n = 8) were convicted of violent crimes against the person. 22 Where the trigger offence for the AAMR cohort was violence against the person (and hence 100% of these would have been related to alcohol by the fact they received a AAMR Order for this offence), CRIS was used to identify how many of these had alcohol flagged on the crime record. Results indicated that the police had only flagged 25% of these offences as being alcohol related and as such, it is not possible to infer how many of this crime type may be attributable to alcohol. 13 5. Conclusion The AAMR was introduced as a pilot concept in 2014 to address the significant problem of alcohol related offending in London. This new sentencing power enabled courts to impose, as part of a Community Order or Suspended Sentence Order, a requirement that compelled an offender to abstain from alcohol for a fixed time period and be regularly tested via electronic monitoring. A previous report23 conveys the findings from the initial performance and process evaluation, highlighting the implementation success and initial learning generated from this proof of concept. The findings presented here build on this and take an indicative look at the impact of the AAMR on reoffending behaviour so far. The analytics need to be caveated by the relatively short timeframes and the small sample sizes with the AAMR cohort overall and especially within the different analytic splits. It is also worth highlighting that impacting any offending or behaviour change is difficult and can take a long time. None-the-less, such analysis can be useful in gaining insights and helping to plot the future course of the AAMR. In terms of main results, using a robustly matched Control cohort, this analysis has demonstrated that the AAMR is broadly comparable to other Community Orders/Suspended Sentence Orders in terms of reoffending behaviour (i.e., proportion of offenders, frequency of reoffending) both during and subsequent the imposition of the requirement and Order. Furthermore, the findings here are equivalent to wider reoffending rates throughout England and Wales for both Community Orders and Suspended Sentence Orders (proportion of offenders = 32%)24. Overall, such results should be interpreted as a positive – there are genuine risks in establishing innovation. For example, the Diamond IOM 25 initiative, was hampered in the early stages by implementation issues, which matured in terms of multi-agency delivery but never fully reconciled ‘working tensions’ between the police and other partners. For AAMR to be achieving broadly the same results as longstanding (and longer) orders is a firm foundation to build upon. The findings reiterate that it can always prove challenging to reduce reoffending, especially for low risk offender groups26. Indeed, previous research27 has suggested that interventions with low risk offenders can in fact increase the probability of recidivism. The issue that these comparable orders to AAMR are far longer is also worth highlighting – especially within a narrative of efficient justice and achieving the same results within a shorter timeframe. To illustrate, the average length of a AAMR requirement is considerably shorter in duration, and potentially requires fewer resources than other similar requirements making it a more efficient form of justice with the potential for cost savings28. The AAMR tag is an additional punitive requirement available to the judiciary and provides, where there is currently no provision, a means of specifically addressing alcohol related offending. Additionally, this short term requirement provides an opportunity to reflect on their offending behaviour, with minimal disruption to the offender’s everyday lives, for example in terms of family life or employment. The AAMR can provide the offenders with a ‘pause’ in their drinking behaviour, and a teachable moment to address alcohol use29, as well as providing reassurance to professionals that the technology underpinning the AAMR is working as intended. Therefore, this requirement potentially has the same impact on offenders offending behaviour, but with efficiency and behavioural benefits. 23 Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC. 24 Ministry of Justice (2017). Proven Reoffending Statistics: Quarterly Bulletin April 2014 – March 2015. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/585908/proven-reoffending-quarterly-bulletin.pdf 25 Dawson, Stanko, Higgins and Rehman (2011). An evaluation of the Diamond Initiative: year two findings. London. Metropolitan Police Service; London Criminal Justice Partnership. 26 Minisrty of Justice (2013). Transforming Rehabilitation: a summary of evidence on reducing reoffending. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/243718/evidence-reduce-reoffending.pdf. 27 Bonta, J. (2009). What to do with low risk offenders? Public Safety Canada, 14, 4. 28 A cost analysis will be conducted as part of the pan London roll out evaluation. 29 Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC. 14 This report is just one element of a larger evaluation taking place around the roll out of AAMR across London. Whilst this original AAMR cohort can still be tracked, emphasis now moves to the pan London rollout. Within this are additional elements such as ‘Tagging at Source’ which aims to administer the electronic monitoring tag to the offender at the Court or probation office rather than their home address, and using the AAMR requirement with offenders of domestic violence, both of which will be evaluated by MOPAC Evidence & Insight. The pan London research includes process aspects (focusing on large scale implementation, stakeholder views, offender experience), cost benefit analysis and performance metrics to gain a greater understanding of the use of AAMR. A report will be produced in summer 2017 providing a comprehensive 12 month rollout report to share the learning and findings from the evaluation focusing on performance and process. In spring 2018 an update report will be produced focusing upon performance and process, and early impact on the pan London cohort (e.g. 6 + 3 months reoffending). A final impact report, focusing on proven reoffending and cost analysis will be produced in winter 2018. 15 Appendix A: Reoffending (arrests and convictions) whilst on Order per requirement Arrests During Order per Requirements AAMR Cohort Control Cohort Multiple Requirement Stand Alone AAMR Multiple Requirement Stand Alone RAR Stand Alone Supervision Stand Alone UPW Total % (number) of cohort who received this Order 61% (66) 39% (43) 58% (146) 3% (8) 7% (18) 31% (78) % (number) of offenders arrested whilst on Order 24% (26) 7% (8) 25% (63) 2% (4) 2% (6) 11% (27) Total number of arrests 72 25 177 22 28 91 Minimum number of arrests 1 1 1 1 1 1 Maximum number of arrests 11 12 12 14 10 15 Average arrests per offender 2.7 3.1 2.8 5.5 4.6 3.3 Average speed to first arrest 140 days 74 days 147 days 50 days 101 days 115 days Range 9 - 327 days 14 - 260 days 1 - 451 days 8 - 113 days 3 - 214 days 8 - 433 days AAMR Cohort Control Cohort Guilty Convictions During Order per Requirement Multiple Requirement Stand Alone AAMR Multiple Requirement Stand Alone RAR Stand Alone Supervision Stand Alone UPW Total % (number) of cohort who received this Order 61% (66) 39% (43) 58% (146) 3% (8) 7% (18) 31% (78) % (number) of offenders convicted whilst on Order 18% (20) 8% (9) 18% (46) 2% (4) 2% (4) 9% (23) Total number of convictions 40 17 118 9 6 57 Minimum number of convictions 1 1 1 1 1 1 Maximum number of convictions 5 5 8 3 2 6 Average convictions per offender 2 1.9 2.5 2.2 1.5 2.4 Average speed to first conviction 166 days 93 days 146 days 94 days 84 days 114 days Range 27 - 342 days 14 - 351 days 1 - 436 days 22 - 239 days 41 - 177 days 18 - 473 days 16 Appendix B: Reconvictions during Order split by age and requirement Standalone AAMR (mean age = 32, n = 43) AAMR Multiple Req. (mean age = 33, n = 66) Control Standalone UPW (mean age = 33, n = 78) Control Multiple Req. (mean age = 33, n = 146) Younger Age Group Older Age Group Younger Age Group Older Age Group Younger Age Group Older Age Group Younger Age Group Older Age Group 0 - 32 years (60%, n = 26) 33 - 63 years (40%, n = 17) 0 – 33 years (55%, n = 36) 34 – 56 years (45%, n = 30) 0 - 33 years (62%, n = 48) 34 - 53 years (38%, n = 30) 0 - 33 years (60%, n = 87) 34 - 63 years (40%, n = 58) % (no.) of offenders 19% (5) 24% (4) 42% (15) 17% (5) 27% (13) 33% (10) 30% (26) 33% (19) Total number of convictions 7 10 30 10 32 25 67 46 Minimum number of convictions 1 1 1 1 1 1 1 1 Maximum number of convictions 2 5 5 4 6 6 5 8 Average convictions per offender 1.4 2.5 2 2 2.5 2.5 2.6 2.4 The other standalone requirements for the Control cohort (RAR and Supervision) had very low sample size, making then unsuitable to be split further by age. 17 Appendix C: Reoffending (arrests and convictions) in the 12 (+3) months from start of Order ARRESTS +12m from start of Order AAMR Cohort Control Cohort Multiple Requirement Stand Alone AAMR Multiple Requirement Stand Alone RAR Stand Alone Supervision Stand Alone UPW Total % (number) of cohort who received this Order 61% (66) 39% (43) 58% (146) 3% (8) 7% (18) 31% (78) % (number) of offenders arrested 45% (30) 44% (19) 40% (58) 63% (5) 39% (7) 37% (29) Total number of arrests 87 87 181 28 31 90 Minimum number of arrests 1 1 1 1 1 1 Maximum number of arrests 11 18 19 14 12 15 Average arrests per offender 2.9 4.6 3.1 5.6 4.4 3.1 Average speed to first arrest 146 days 141 days 122 days 93 days 125 days 115 days Range 9 - 327 days 40 - 319 days 1 - 351 days 8 - 268 days 3 - 265 days 8 - 314 days CONVICTIONS 12 (+3)m from start of Order AAMR Cohort Control Cohort Multiple Requirement Stand Alone AAMR Multiple Requirement Stand Alone RAR Stand Alone Supervision Stand Alone UPW Total % (number) of cohort who received this Order 61% (66) 39% (43) 58% (146) 3% (8) 7% (18) 31% (78) % (number) of offenders convicted whilst on Order 33% (22) 40% (17) 34% (50) 50% (4) 28% (5) 32% (25) Total number of convictions 43 70 135 18 11 72 Minimum number of convictions 1 1 1 2 1 1 Maximum number of convictions 5 17 10 7 4 6 Average convictions per offender 2 4.1 2.7 4.5 2.2 2.9 Average speed to first conviction 167 days 164 days 163 days 94 days 148 days 129 days Range 27 - 342 days 14 - 415 days 1 - 447 days 22 - 239 days 41 - 407 days 18 - 436 days 18 Appendix D: Offense type for reoffences in the 12 (+3) months since start of Order. Offence Type AAMR Cohort Control Cohort Total number of reoffences 113 236 Drugs 6% (7) 11% (27) Offences Against the Person 4% (5) 8% (19) Offences Against the Property 5% (6) 4% (9) Offences Against Police, Courts and Prisons 37% (42) 39% (91) Other 22% (25) 18% (42) Public Disorder 4% (5) 1% (3) Sexual Offences 4% (4) 1% (2) Theft 15% (17) 14% (33) Weapons 2% (2) 1% (2) Fraud 0 3% (6) 19 Appendix E: Reoffences by specific crime type in 12 (+3) months after start of Order Drugs AAMR Control Offences Against the Person AAMR Control Offences Against Property AAMR Control Offences Against Police, Courts & Prisons AAMR Control % (no) of offenders with reoffence for specific crime type prior to start of Order (Sub-Cohort) 29% (32) 34% (85) 43% (47) 57% (143) 37% (40) 31% (78) 38% (41) 49% (122) % (no) of Sub Cohort offenders with reoffence of any type post start of Order 53% (17) 42% (36) 43% (20) 32% (46) 55% (22) 41% (32) 51% (21) 42% (51) Number of reoffences 69 112 70 127 77 85 79 148 % (no) of these reoffences for specific crime type: 7% (5) 24% (27) 4% (3) 13% (17) 8% (6) 9% (8) 39% (31) 38% (56) % (no.) of Sub-Cohort with reoffence for specific crime type: 9% (3/32) 13% (11/85) 6% (3/47) 6% (9/143) 13% (5/40) 10% (8/78) 39% (16) 26% (32) Average no. of reoffences each (for sub-cohort) 4.1 3.1 3.5 2.8 2.4 2.7 3.8 2.9 Other Offences AAMR Control Public Order Offences AAMR Control Theft Offences AAMR Control % (no) of offenders with reoffence for specific crime type prior to start of Order (Sub-Cohort) 66% (72) 48% (119) 37% (40) 25% (62) 37% (40) 43% (108) % (no) of Sub Cohort offenders with reoffence of any type post start of Order 35% (25) 39% (47) 40% (16) 42% (26) 55% (22) 43% (46) Number of reoffences 83 154 32 69 80 145 % (no) of these reoffences for specific crime type: 24% (20) 19% (30) 3% (1) 3% (2) 23% (18) 23% (34) % (no.) of Sub-Cohort with reoffence for specific crime type: 13% (9) 12% (14) 3% (1) 3% (2) 15% (6) 18% (19) Average no of reoffences each (for sub-cohort) 3.3 3.3 2 2.7 3.6 3.2 *Highlighted row in table denotes reoffences for specific crime type where there was also a pre-conviction for this same crime type. 20 Appendix F: Percentage of available cohorts to measure reoffending post completion of Order 21 Appendix G: Arrests in the 9 months after completion of the Order AAMR Cohort Control Cohort Multiple Requirement Stand Alone AAMR Multiple Requirement Stand Alone RAR Stand Alone Supervision Stand Alone UPW % (number) of offenders arrested in 9m since Order complete 15% (n = 12) 15% (n = 12) 16% (n = 25) 1% (n = 2) 4% (n = 6) 6% (n = 9) Total number of arrests 29 66 54 6 11 18 Minimum number of arrests 1 1 1 2 1 1 Maximum number of arrests 6 15 13 4 3 4 Average arrests per offender 2.4 5.5 2.2 3 1.8 2 Average speed to first arrest 85 days 83 days 137 days 116 days 137 days 102 days Range 8 - 151 days 3 - 217 days 21 - 262 days 85 - 146 days 15 - 256 days 10 - 224 days London Gang Exit Evaluation Progress Report October 2016 Emily Southall & Lynne Grossmith MOPAC Evidence and Insight 2 Contents Executive Summary ................................................................................................................ 1 1. Introduction ........................................................................................................................ 2 Background ............................................................................................................................ 2 London Gang Exit and its aims ............................................................................................... 3 Evaluation overview ............................................................................................................... 4 2. Performance monitoring ................................................................................................... 5 Process model ........................................................................................................................ 5 Current cohort demographics ................................................................................................. 5 Offending data ....................................................................................................................... 5 Referrals & Interventions ........................................................................................................ 6 Rejections & Referrals ............................................................................................................ 7 The need for wider support .................................................................................................... 7 3. Findings to date ................................................................................................................. 8 Understanding of LGE programme ......................................................................................... 8 Initial implementation: Training and Resourcing .................................................................... 8 Separation of the referral centre (CRC) .................................................................................. 9 Cross agency referrals ............................................................................................................ 9 Managing expectations ........................................................................................................ 10 Poor data quality and missing referral information ............................................................... 10 Specific intervention strands ................................................................................................ 11 The Cost of LGE ................................................................................................................... 11 4. Conclusion and next steps .............................................................................................. 12 Moving forward: changes to the LGE referral process .......................................................... 12 Evaluation: next steps .......................................................................................................... 13 Appendices ............................................................................................................................ 14 References ............................................................................................................................. 17 1 Executive Summary Commencing in February 2016 and running until October 2017, London Gang Exit (LGE) is a £1.5M multi-agency intervention, jointly commissioned by MOPAC and the Community Rehabilitation Company (CRC), which aims to reduce the harm caused by gangs; reduce involvement in gang offending; and to assist exit from a gang related lifestyle. LGE aims to provide a service tailored to identify needs of individuals via delivery partners. This interim LGE report provides a brief overview of the evaluation methodology, highlighting ongoing risks and presenting key performance data to date. Learning around implementation via feedback from the first practitioners’ survey, and qualitative semi-structured interviews carried out with relevant LGE staff are included. Any measurable impact of key outcomes will be detailed in future reports (as per the evaluation timeline provided). A final report, including potential analysis of offending proxies and cost will be available in May 2018. Findings to date are summarised below: What is working well? • Training has been well received, particularly its thoroughness. Up-skilling practitioners offers a sustainable approach to tackling gang culture and desistance from it. • Referral numbers are increasing, indicating there is a good relationship with the boroughs. • Although the majority of referrals are from the 16-24 age cohort, there have been some made and accepted for those younger than 16, indicating a need for support from a wider cohort. • There is a shared understanding between the referral centre (CRC) and LGE of the barriers between referral and intervention; several strategies with the aim of making the process more streamlined have been actioned because of this disjoin. Challenges so far • The name of the programme – ‘London Gang Exit’ is considered misleading, potentially secluding certain groups, .e.g. young females, from being referred. • Further engagement is needed with ‘harder to reach’ groups, for example those at risk from CSE (Child Sexual Exploitation), and the VCS (Voluntary Community Sector) to increase referrals. • There are barriers associated with the separateness of the referral centre and LGE, affecting the ability to retain the integrity of the programme. Greater sharing of knowledge and experience is recommended to ensure a more streamlined feedback loop. • A shortage of LGE staff caused initial problems, with specialist caseworkers taking on clients in the absence of support workers. In some cases, only one specialist worker has been assigned to a specific intervention. • Missing data on referral forms and inconsistencies between the forms in circulation have led to delays in the client moving from the referral team to LGE, delaying client access to interventions. • The quality of data has been inconsistent, particularly around offending data where it remains unknown if many of the current cohort is on the MPS Gangs Matrix. 2 1. Introduction Background Youth violence and gang related crime in London continues to be a significant concern. Although only an estimate of a complex picture, Metropolitan Police Service (MPS) data suggests there are currently around 182 gangs, made up of approximately 3,500 individuals identified across the capital. These individuals are estimated to be responsible for a disproportionate amount of London’s crime; 9% of all personal robbery, 16% of serious youth violence, and 40% of shootingsi. Criminal career analysis, conducted by the MOPAC Evidence & Insight team indicated a typical gang member had on average nine proven offences in their history, and were on average 15 years old when they committed their first offenceii. After initial reductions in 2012, gang flagged crime has been increasing since (1,579 gang flagged offences in 2013 (CY), compared to 2,094 in 2015 (CY))iii. In response to recent rises across the board in proxies of ‘gang related crime’ (such as knife and gun crime offences) there is recognition more needs to be done to target not only those involved in the gang-lifestyle, but youth at risk of joining. The latest Mayoral manifesto acknowledges the daily reality of gang related threats of violence, and knife crime in particular, that many communities face. It states the intention to develop anti-gang strategies with local authorities, schools and youth services, whilst working with community organisations to further youth engagement. It is recognised individuals involved with gang violence are also likely to have complex needs – experiencing higher levels of victimisation and a higher incidence of mental health difficulties, above both the general population and other entrants to the criminal justice systemiv. As gang members are disproportionately involved in violence, as both victims and offendersv it important for any programme promoting desistance from gang culture to be inclusive of both, and recognises the support needed for young people living with the threat of violence. There is limited research around ‘what works’ in terms of leaving gangs. Studies have highlighted the average age of a gang member is around 20vi and the upper age limit is around 25vii, suggesting there may be a maturational ‘turning point’ which exists in a criminal career and requires an active cognitive element of wanting to get outviii. It is recognised that exiting a gang will most likely be a difficult processix, as just as gang entry is multi-dimensional, so is gang-exit, therefore any intervention must share this naturex. As leaving often involves a great number of people and requires facing an ever-greater challenge in terms of dealing with the negative consequences of exclusion and labelling, the process may become involuntaryxi. A successful intervention programme will be reactive to these changes. Other factors thought to promote the likelihood of leaving include pivotal life events such as parenthood and new opportunities e.g. a good jobxii, with a central need for an individual to have a sense that things can be different in the futurexiii. Therefore, gang exit is considered a process, which occurs over time and involves increasing ties to ‘conventional activities and institutions’xiv. In contrast, there is little evidence to support the criminal justice system as a reason linked to exitxv; the emphasis instead lies on gang experiences (such as violence), social processes and fundamentally a level of motivation from the individual to feel ‘ready’ in order for it to be successful. Despite a wealth of academic researchxvixvii and promising localised work across the third sector and community interventions (e.g. The Hackney Gangs Interventions Project, Southwark’s SERVE 3 programme, Waltham Forest’s Gang Prevention Programme, and Lewisham’s Trilogy+1), London services are still provided at a borough level. National and regional initiatives such as the Tackling Gangs Action Programme (2007), and Ending gang and youth violence (2015), have led to policy changes, such as mandatory sentences for knife and gun crimesxviii. However, there is still a lack of consistency around the opportunities on offer across London’s boroughs. Some commission no gang specific services at all and many outer London areas do not have the resources or experience to deal with the complex issue of gangs, particularly given the reported increased fluidity in territorial and criminal activitiesxix. The requirement of a ‘collaborative pan-London gang exit service’2 to replicate need and be free from the constraints often associated with borough specific services has become increasingly obvious. Without a pan-London response, the continual reliance on varied local service provision will not guarantee young people receiving a consistent approach or an equal opportunity to access services. In 2016, the government refreshed its national approach to gangs; the Home Office are reforming their ending gang and youth violence programme into a new approach to end gang violence and exploitationxx. The Pan London Gangs Exit (LGE) service remains linked at a strategic level with these developments. London Gangs Exit and its aims Commencing in February 2016, the LGE is a £1.5M multi-agency intervention, jointly commissioned by MOPAC and the Community Rehabilitation Company (CRC). Safer London (SL) is leading a consortium of organisations to deliver the LGE with key delivery partners Redthread and Only Connect, working with the CRC’s new referral centre. The Pan London service was designed to complement and enhance existing local services, filling gaps in provision. Specific aims are to: • Reduce offending • Increase stable and secure housing • Increase education, training and/or employment • Improve self-esteem and well being • Improve health and mental health needs • Engage in gang exit activity There are three key aspects to the LGE approach: • Inclusive cohort – it is the first pan London service for anyone aged 16-24, who are at risk from gangs and/or who pose a risk of harm to others. It is unique in its approach by recognising the complexities of gang involvement and the blurred lines between offender and victim. • Tailor made support - this could include mental health support, employment support and specialist support for young women. • Holistic approach – The allocation of a trained caseworker to the young person, to provide intensive support and advocate on their behalf by building an effective and consistent relationship. 1 Tackling London’s Gangs. February 2012 2 As indicated during the consultation process for the Strategic Ambitions for London: Gangs and Serious Youth Violence 4 Additionally, the LGE service has an innovative commissioning approach, in that it transcends borough boundaries, to create a consistent service across London. The service will integrate with local delivery and provide additionality to local services. Evaluation overview Evidence and Insight (E&I), MOPAC’s in-house social research team, are conducting the evaluation of the LGE. A holistic plan has been designed with the intention to measure: 1) Performance-monitoring data 2) The process of implementing LGE, 3) How the service impacts on the user and the specified outcomes, and 4) The cost of the programme (i.e., cost benefit analysis). Using a mixed methods approach, the evaluation will draw on an array of data sources from MPS and Partner agencies across thematic areas as outlined in Appendix A. Scoping for suitable data is ongoing; access problems and/or data quality may limit inclusion in the final evaluation, and/or the ability to measure the impact against the programmes aims. Depending on sample sizes, there is a potential to use a matched controlled design for the more robust measures, such as offending data; it may only be possible to access indicative information for other outcome measures. To date, six monthly spreadsheets have been submitted by CRC and SL to understand who is accessing the service; a staff survey has been distributed; and nine interviews have been conducted to understand the process of implementation to date (see Appendix C for more details). Timeframes for deliverables Further staff surveys are planned for; November 2016, April 2017, and September 2017. A 15-month progress report, which will include year one impact (proxies of offending analysis), is scheduled for July 2017. The final impact assessment is due in May 2018 (see Appendix D for a more detailed timeline). 5 2. Performance Monitoring Process model The LGE consists of a referral process known as ‘Phase 1’, and subsequent intervention, known as ‘Phase 2’. Referrals are received from local authorities, police, health, as well as a plethora of other organisations into CRC3. CRC hold the decision-making responsibility as to whether clients are accepted on to the programme or not. Cases are assessed using information on the referral forms, risk assessments and motivational tools. Once accepted, the case is passed to LGE (known as Phase 2) where a client can be referred for: • ‘Component 1’ - they are allocated a support worker to act as their case manager, • ‘Component 2’ - they are allocated a relevant specialist worker to act as case manager. • If necessary, clients can be submitted for both component 1 and 2. LGE provides in-house services as a means of intervention and have the ability to ‘spot purchase’ if there is a specific service that they cannot offer (Appendix F shows this process in further detail). Current cohort demographics Safer London holds responsibility for the case management function for LGE, with individual level client data sent from both CRC and Safer London to E&I on a monthly basis to help monitor if the process is working as intended. There are currently 138 active cases on the LGE programme4, consisting of 12 females and 126 males. Since March 2016, the programme has allowed flexibility with the age of the client – with 14 individuals under the age of 165 and 16 over the age of 256. Where data on ethnicity is available (n=121), the majority of clients are classified as BME (n=107, 87%), this is in-line MPS Gangs Matrix ethnicity. A small number report a disability, sometimes related to injuries sustained because of a gang related incident i.e. shooting or stabbing7. Offending data It is the intention of the E&I evaluation to assess the offending behaviour of those individuals on the LGE programme who have been referred and accepted. The most robust way to analyse this is through proven reoffending. This can be defined as any offence being committed in a one-year follow-up period and receiving a court conviction, caution, reprimand or warning in the one year follow-up or a further six month waiting period. As such, this analysis is less timely – the full process requires an 18-month period 3 i.e. youth offending teams, MPS, social services 4 As of 16th September 2016 5 The youngest client is aged 12. 6 The oldest client is aged 29. 7 LGE – Quality thematic report Q1 June 2016 Figure 1. Cohort demographics 6 of waiting. It is the intention for this analysis to feature in the final evaluation report (due around May 2018) along with the possibility of including a comparison group. However, other less robust types of offending analysis are possible and for the second progress report (due around July 2017) - proxies of offending will be assessed e.g. arrest and police charge data in order to develop early insights. For this progress report, only the data provided from LGE with regards to the MPS Gangs Matrix8 have been assessed, however many inconsistencies have been highlighted e.g. client information is often unknown (38 cases). 37 cases have been identified as not being on the Matrix, and eight cases are identified as being on the Matrix but their scoring is unknown.9 Additionally, two of the 37 are classified as Red, two as Amber, and two as Green. However, there are caveats to using Matrix data and RAG ratings10 to assess general offending, as the scores are comprised of recent intelligence and known offending activity across a specific crime type. Referrals & Interventions The table below indicates the available data regarding the intervention strands on the LGE. Many clients referred into the service for gang involvement, are also experiencing harm from gangs which is reflected in the figures. Housing also appears to be a priority, along with education, training and employment (ETE). Whilst it is useful to monitor the number of people receiving interventions, there is little robust data provided on whether the client successfully engaged, with an intervention or whether the services received are being tailored to all the client’s original needs. Later reviews will look to supplement this data with coding of case notes, to obtain a greater understanding of the services provided. This will be 8 Trident Matrix - identified the most harmful gang members, scoring individuals identified by police and partners who are thought to be associated with gang violence. Intelligence on an individual’s previous 3 year violent history is used to score and assign a risk rating of red, amber, or green based on their level of harm they pose, in addition to the likelihood of them becoming a victim of gang related crime. 9 For impact analysis, MOPAC E&I will retrieve relevant Matrix scores from MPS. 10 Three possible RAG ratings (green, amber, red) therefore this may not be a sensitive enough measure. 11 LGE – Quality thematic report Q1 June 2016 Relevant Strand Total no. of Live Clients Gang Exit Strand 26 Harm From Gangs Strand 36 Offending Behaviour Strand 20 Housing Option 1 Strand 16 Housing Option 2 Strand 26 Housing Option 3 Strand 12 Health and Wellbeing Strand 19 Family and Relationships Strand 17 Harmful Sexual Behaviour Strand 0 ETE Strand 2711 Table 1. Total number of live clients and intervention strands 7 important analysis to understand whether certain individuals, having certain needs met end up committing less offences. Rejections and re-referrals So far, eight people have been rejected by CRC for the following reasons; - The application form has not been fully completed. - Lack of evidence of motivation for young person to change. - Individual issues (immigration, court, which has resulted in the referrer withdrawing). The majority of these (six) have been re-referred, and five of these have now been accepted due to other agencies being unable to take any new clients (or seeing no progress), and outstanding court cases have been dealt with. The need for wider support It has become apparent there is a greater need from individuals who do not fit the ‘regular’ MPS profile of gang-related offending. For example, the broader age groups (outside 16-24 range), the inclusion of females (who make up 9% of current cohort) and the lower offending/risk level (25% of current cohort are not on the Matrix), produces a different cohort of individuals than might be expected from MPS gangs data. As anticipated, data suggests around a third more individuals (n=115) are referred to the LGE for their risk of experiencing harm from gangs, rather than a need to exit the gang itself. It maybe those individuals scoring Red or Amber on the Matrix already have good access to services, in which case there is a greater need to fill the gaps in service provision for slightly different cohort who are affected by gangs but may not themselves be highlighted as high risk for gang offending. This is important to note for two reasons: 1) This cohort may have a different set of needs, in terms of practical implementation the services provided may need to adapt accordingly; and 2) The different make-up of this cohort will affect the ability to assess impact – not only who they are matched with, but whether we would expect to see a reduction in offending behaviour which is already at a lower level. 8 3. Findings to date Findings from the first LGE practitioner survey and the staff interviews have been split into eight emerging themes; understanding of the programme, training and resourcing, the separation of CRC, cross agency working, managing expectations, data quality and missing information, specific intervention strands and the cost of LGE. The first practitioner survey ran for two weeks from 7th July 2016 (see Appendix B). Findings should be interpreted with caution due to the low survey response rate (n=19/26) however the results below also draw from the nine interviews with staff members from both CRC and SL. Understanding of the LGE programme There is an appetite amongst professionals for a gang exit programme in London; all respondents from the survey agreed there is a significant gang problem in London, and the majority believe LGE will reduce gang violence. Positively, there appears to be a widespread understanding amongst the practitioners of the process involved in the LGE programme, such as who makes the referrals and the wider aims of the programme. Survey and interview feedback suggests practitioners are confident in their understanding of their roles. As more clients are currently accessing the LGE service for their risk of experiencing harm from gangs rather than a need to exit the gang itself, the name of the programme – ‘London Gang Exit’ is thought to be misleading for potential referrers, ‘it implies that certain things might not be offered’ (staff interviewee). In addition, practitioner interviews suggested there is further confusion with the programme still being associated with a previous SL project ‘Safe and Secure12’, particularly around the housing intervention (discussed below). Initial implementation: Training and Resourcing Despite an initial understanding of the programme, practitioners appear to be concerned they do not have the right tools to make referrals and a lack of resourcing was a universal concern. Only a third of survey respondents think LGE has enough practitioners to ensure clients receive an effective intervention, with staff interviews corroborating this. It is thought the current structure of Component One (support worker) and Component Two (specialist worker) (see Appendix E) could also be improved, as ‘currently everyone is integrating’ (staff interviewee), with specialist workers taking on more general roles and assisting in the absence of enough support workers. The need for them to hold cases themselves therefore limits their capacity to deliver specialist support. “There has always been a gap between staffing levels we have and staffing levels we should have, but we are still expected to deliver a high quality service and meet the needs of those on the programme” (staff interviewee) 12 Safer London programme to re house those at high risk of gang related violence and end individuals’ involvement in gang lifestyle http://saferlondon.org.uk/wp-content/uploads/2014/11/Executive-Summary-of-Safe-Secure-Progress-Review-Oct-2013.pdf 9 Some concerns were raised around staffing levels at the beginning of implementation13. Positively, interviews indicated staff feel workloads are constantly being reassessed and contingency plans are in place should there be an influx of referrals (e.g. the use of volunteer mentors). Feedback regarding training was positive, but there are distinct points for improvement. Feedback suggested that staff received training before they began their LGE role and whilst most thought this prepared them enough, it was highlighted a broader understanding of the processes used at both the referral centre and the intervention level was required to better clarify roles. Specialist staff training (e.g. harmful sexual behaviour) was highlighted as something required to up-skill the whole team, ensuring a universal understanding of the intervention pathways available. Practitioners thought training was of good quality, although some suggested this could be better tailored to account for individual experience, that those who are new might require more support. “I wish I had this quality of training when I started in the field” (staff interviewee) “Workshop attended by front line staff and operation management to amalgamate the processes involved in referral and delivery, and clarify what they are” (staff survey respondent) This indicates scope for continued work to provide training refreshers; LGE seeks to address this through the introduction of monthly case review panels, where support workers will present a case they hold to specialist workers for feedback. Separation of the referral centre (CRC) Barriers around the separateness of the CRC referral centre are apparent, creating difficulties with engagement and the timeliness to provide an intervention. Practitioners highlight how the referrers are possibly unaware of the involvement of other agencies in the process and their responsibilities (i.e. the interventions available), which can compromise the level of support provided. It is felt this disconnect between the initial referrer and the final service provider limits the credibility of the programme, creating an ‘ineffective feedback loop’ (staff interviewee). There are additional cultural differences between CRC and LGE which impact on how they work with clients; because LGE is voluntary the clients may need more motivating to engage in the intervention. However, it was also acknowledged there are benefits to having a separate referral centre; for example, by eliminating any bias of who is accepted or rejected from the programme. Positively, both CRC and LGE have been reactive to this potential barrier, offering cross-team shadowing experiences, weekly service manager meetings, and quarterly workshops for CRC so referrers can get to know the process. “CRC should integrate into delivery arm so it becomes a more seamless service, but should retain separation in order to retain integrity” (staff interviewee) Cross-agency referrals Respondents felt more initial consultation with partners and referring agencies would have been useful to manage expectations around the programme when it went live. As the main sources of referrals come 13 Lack of support worker at LGE 10 from statutory bodies rather than the voluntary and community sector (VCS, 25%), it suggests engagement work still needs to be done with these organisations. This is reinforced by the fact that although there is a good geographical spread of referrals14, some staff believed certain boroughs are suspicious about using the service because they have their own gangs unit. There is therefore a need to promote clarity around the additionality the service is aiming to provide. Survey findings indicate the majority of staff also remained neutral on whether the right partners were involved, but more positively practitioner interviews revealed a shift in opinion with many stating there is a good working relationship with the right agencies. “Until people know it is successful they aren’t as willing to refer” (staff interviewee) “There was a slow uptake as many referrers were initially unsure of the service and if it was a threat to other agencies” (staff interviewee) Managing expectations Some practitioners have expressed concern that without any visible outcomes of whether the programme is working in gang desistance or not, there will be a reluctance to refer people. Furthermore, there is uncertainty around what happens after the six-month intervention, and who will hold the responsibility for clients who fall out of the LGE remit. “What is the end product for the user? This needs to be clear else it won’t work” (staff interviewee) Poor data quality & missing referral information Another service assessment commissioned by CRCxxi highlighted the apparent lack of understanding regarding the importance and standardisation of risk assessments on the referral forms, with little consideration that elements of risk can affect the type of intervention used. This has been corroborated by feedback from staff interviews, which highlight problems regarding inconsistencies on the form, and the different ‘versions’ in circulation. It was found that 70 of the cases initially passed to LGE had incomplete referral forms, causing a delay in the process. As a result, there is now a preference by CRC to make phone referrals in the first instance, to prevent wasting time filling out extensive referral forms, only for the case to be turned down by the referral centre. Staff interviews highlighted further concern that the timeliness of the referral process could affect the willingness of the client to engage. This was supported by LGE performance data, which indicates incomplete referral forms cause delays (e.g. back and forth process). The main reasons for this were missing or incomplete information including risk assessments, motivational tool questionnaires and other 14 See ‘performance data overview’ CRC Service Assessment: Main Findings • No standardisation of risk assessments on referral forms. • Lack of understanding about the importance of the ‘risk of harm’ section of the form. • Lack of consideration of the background and risk elements and how they affect the intervention applied. • No value in the Motivational Assessment Tool as there is little understanding of how variations in answers are considered. 11 supporting documents. Some cases had missing information from Trident and the prison service, although this was less common. Specific intervention strands Housing is a priority for the clients - with around two thirds requiring this strand of support. From the 73 clients in the current cohort who need housing advocacy from LGE, 10 of these were deemed at ‘immediate risk’, needing relocation to a different borough. Staff fed back that few clients understand their housing rights, therefore a lot of work is required to manage expectations. “Some young people think it is a route to housing” (staff interviewee) The number of referrals for women is low (less than 10% of current cohort), suggesting referrers do not appear to be making the connection between the services on offer and the potential to support individuals affected by gang harm – in particular victims of gang related Child Sexual Exploitation (CSE)15. When practitioners were asked if there were any clients they would like to see on the programme, many cited more young females. LGE qualitative reports outline plans for future stakeholder engagement, to focus on key areas and raise awareness of the support on offer to gang affected young women. The Cost of LGE E&I are currently assessing the feasibility of a cost benefit analysis (CBA) of the LGE. CBA is a method used to evaluate the quantitative impact of an intervention by the use of robust economic analysis. A CBA of LGE would determine if the expenditure of public money is providing a net benefit to society, namely, in terms of reduction in harm caused by gangs. Furthermore, it would determine whether the intervention could make a case for future funding. To determine feasibility E&I have requested data from LGE concerning all costs associated with delivering the project. E&I have begun to baseline these costs as the service has commenced. Should CBA not be feasible, alternative economic analysis will be considered.16 15 LGE – Quality thematic report Q1 June 2016 16 Cost analysis, break-even analysis or cost effective analysis. 12 4. Conclusions and next steps The nature of gang involvement and desistance is multi-faceted and complex, and it is recognised that this is reflected in the difficulties of implementing a successful gang exit intervention. Although there has been some promising practice seen to date (e.g. the partnership working with CRC), there have also been several implementation challenges, which have initially hampered the programme. • The branding of ‘gang exit’ implies a service designed to solely help gang members leave, yet the programme has a much broader remit to help those affected by gangs, or who are at risk of harm from gangs. This has the potential to exclude certain groups. Looking forward, a consistent level of engagement will ensure the young people most in need are getting access to the LGE. • A shortage of LGE staff caused initial problems, with specialist caseworkers taking on clients in the absence of support workers. However, this was identified as an initial mobilisation issue, and staff are now confident there are sufficient contingency plans in place should caseloads reach full capacity (e.g. the use of volunteers). • There are barriers associated with the separateness of the referral centre and LGE, which affects the ability to retain the integrity of the programme. Greater knowledge and experience sharing is recommended to ensure a more streamlined feedback loop. Training in specialist areas to upskill all members of the team would also help to improve this. Moving forward: changes to the LGE referral process Positively, both CRC and LGE have been reactive to problems that have arisen. Since March 2016 referrals have been accepted for those younger than 16, indicating an understanding of the need to provide support to a younger cohort. Referral numbers continue to increase, demonstrating a good relationship with the boroughs, and the message of LGE is reaching the appropriate people. Based on initial experience with implementation, interventions and the current cohort the box below indicates proposed changes to the LGE referral process. The progress of these changes, uptake from the specific referral routes and any impact on the client cohort will be assessed at the next evaluation review. • Redthread ‘live’ referrals - Staff engage with young people who present at Trauma Centres at the time when the young person involved in the risky behaviours is more open to change – known as the ‘teachable moment’. This aim is to help the young person stop, review and make changes to their lives by agreeing to being referred to LGE. • Prison referrals - Literature has suggested gang members are most open to intervention when they have been convicted and sanctioned i.e. in custody. LGE will be piloting an ‘in reach’ service in HMP Thameside for those who are approaching their release date. Data from CRC has indicated that referrals have been turned down because the client has been in prison. • Self-referrals – LGE has plans to accept self-referrals. 13 Evaluation: next steps There will be a second wave of the staff survey distributed in November 2016 containing some questions from the first survey (in order to produce a ‘trend tracker’), plus additional aspects to cover topics relevant to the programme as it approaches 12 months. E&I plans to conduct further staff interviews; coding case studies to add context; collecting client satisfaction data and feedback; and the possibility of client interviews.. In order to make robust conclusions about the impact of the programme against the aims (particularly offending data), the most feasible methodology identified is to scope a matched control sample17 . For this to be successful, clients will be matched on a number of demographics as well as risk/offending history. A second interim report available in July 2017 will contain a performance overview, survey and interview feedback, offending proxies and conviction analysis for the year 1 cohort. A final report reflecting on the full two-year programme and a final impact assessment containing the conviction analysis for the year 2 cohort, plus cost analysis, and will be available in May 2018. 17 For example a technique such as Propensity score matching 14 Appendices Appendix A: Proposed data source Appendix B: Proposed methodology Appendix C: Current methodology Outcome Data Housing Self-reported data; MPS Merlin. Possibly local authority data Well-being and self-esteem Distance Travel Assessment Tool; self- reported; psychometric scales; MPS data on stop and search; MPS merlin - vulnerability; MPS victim data. Possibility for health data Training, education and employment Self-reported; MPS Merlin. Possibility for Department of Education; National Insurance; Department of workplace and pensions data Offending MPS data on proxies of offending, such as arrest or police charge – CRIS, CRIMINT, MATRIX, Custody, Gang flagged (individual & borough level); Official conviction analysis done at 15 months from PNC Method Participants Date Staff survey 28 people across Safer London and CRC received the survey, with 19 respondents - a 68% response rate. Launched 7th July 2016 for 2 weeks Staff Interviews E&I conducted 8 face-to-face interviews in August 2016, with staff members across a range of both strategic and practitioner roles within Safer London and CRC. August 2016 Process of Implementation - Staff surveys - Staff interviews - Performance information- Individual level case data to monitor the number of clients enrolled in the service. This includes - completed referrals; length of time between referral and service; demographic information; needs; referral matched to need; collaboration with existing services. - Client information - Existing or E&I surveys; Distance Travel Assessment Tool; coding case notes; assessment of documented needs; ethnography; field work Cost of LGE - Baseline (Set up costs and predicted spend for year 1 and year 2) - Actual spend (End of year spends) 15 Appendix D: Evaluation timeline Appendix E: CRC process model Performance data Excel spread sheet sent to E&I from Safer London 1st of each month 16 Appendix F: Safer London process model 17 References i MetMis ii Strategic Ambitions for London: Gangs and Serious Youth Violence. 2014. MOPAC iiiA randomised control trial of Mental Health Awareness and Safeguarding Training (MAST). Southall, E., Grossmith, L., and Dawson, P. 2016. iv http://www.sadiq.london/a_manifesto_for_all_londoners v Criminology and public policy. Esbenson et al, 2013 viDawson, P. (2008) Monitoring data from the Tackling Gangs Action Programme. Home Office Crime Reduction Website. London: Home Office. viiMarshall, B., Webb, B. and Tilley, N. (2005) Rationalisation of current research on guns, gangs and other weapons: Phase 1.London: University College London Jill Dando Institute of Crime Science. viiihttps://www.gov.uk/government/uploads/system/uploads/attachment_data/file/217379/research-gang- violence.pdf ix EXITING GANGS: TOWARD AN UNDERSTANDING OF PROCESSES AND BEST PRACTICE x Hastings et al (2011) xi Crime Prevention Ottowa xiihttps://www.gov.uk/government/uploads/system/uploads/attachment_data/file/217379/research-gang- violence.pdf xiiiFarrell, S. and Calverley, A.(2006) Understanding desistance from crime: Theoretical directions in resettlement and rehabilitation. Berkshire: Open University Press. xivhttps://www.gov.uk/government/uploads/system/uploads/attachment_data/file/217379/research-gang- violence.pdf xv https://www.nationalgangcenter.gov/Content/Documents/Getting-Out-Staying-Out.pdf xvihttp://www.centreforsocialjustice.org.uk/UserStorage/pdf/Pdf%20reports/DyingtoBelongFullReport.pdf xvii https://www.ncjrs.gov/pdffiles1/nij/188741.pdf xviii https://www.gov.uk/government/news/repeat-knife-offenders-face-new-minimum-prison-term xixhttps://www.gov.uk/government/uploads/system/uploads/attachment_data/file/491699/Ending_gang_viole nce_and_Exploitation_FINAL.pdf xxhttps://www.gov.uk/government/uploads/system/uploads/attachment_data/file/491699/Ending_gang_viole nce_and_Exploitation_FINAL.pdf xxi J., TOY (2016) PAN LONDON GANG EXIT PROGRAMME – QUALITY ASSURANCE BRIEFING MARCH 2016, COMMISSIONED AUDIT BY CRC. 1 Alcohol Abstinence Monitoring Requirement The pan London roll out: A review of process and performance from year 1. Month and Year Authors MOPAC Evidence and Insight July 2017 Z. Hobson, B. Dangerfield & A. Harrison MOPAC Evidence & Insight 2 Contents Executive Summary 3 1. Introduction 6 Sobriety Programmes and Interventions 6 Alcohol Abstinence Monitoring Requirement 7 The 2014/2015 Pilot 8 Pan London Roll Out 9 2. Methodology 10 3. Results 12 Using the AAMR: Performance Learning 12 Imposing the requirement 12 The AAMR Tag 13 Offender Demographics 14 Key Learning 14 Using the AAMR: Process Learning from Stakeholders 15 Training and Awareness 15 Working Practices 16 Using the AAMR 17 Potential Effects on Offenders 19 Domestic Abuse Perpetrators 19 Looking Forward 20 Key Learning 21 Using the AAMR: Process Learning from Offenders 22 Entry Survey 22 Exiting the AAMR Order 23 Key Learning 24 Using the AAMR: Cost Implications 25 4. Discussion 26 References 28 Appendices 30 3 Executive Summary In 2011 the Mayor’s office secured legislation to allow for the introduction of a new sentencing power, the Alcohol Abstinence Monitoring Requirement (AAMR) to tackle the significant problem of alcohol related offending in London. The AAMR gives the Judiciary the statutory power to stop an offender drinking alcohol (Compulsory Sobriety), where their offence is alcohol related. The AAMR involves fitting a tag to the offender’s ankle and monitoring their alcohol consumption for up to 120 days. When this is not complied with, the offender will be breached and punished further. Following the positive learning from the pilot rollout of the innovation, during 16-17 the AAMR has now been rolled out across London. This programme continues to monitor the technology and generate learning. This report provides learning from the first year of the pan London AAMR focussing upon process and performance evaluation. Key Findings: Performance learning  A total of 3671 AAMRs have been imposed between April 2016 and March 2017, with an average length of 67 days tagged.  The vast majority of offenders who completed the AAMR did so successfully, indicating a 92% (n=235) compliance rate with this requirement2.  AAMRs were usually given for a Community Order (74%, n=270), and standalone AAMRs accounted for 28% (n=100) of all Orders. Multiple requirement Orders usually consisted of AAMR and Unpaid Work (19%, n=69).  AAMRs were given for a variety of crime types, most commonly in relation to violence (31%, n=115) or drink driving offences (22%, n=82), which is similar to the pilot (63%, n=80 – both offences).  AAMRs have been imposed across London, most frequently from Magistrates Courts (93%, n=342), specifically Croydon Magistrates (17%, n=58) and Uxbridge Magistrates (14%, n=49). However, other courts such as Thames and Barkingside who were last to roll-out AAMR in January 2017 have also performed well and have awarded 16 and 17 AAMRS respectively.  The technology continues to work as intended, taking on average 48 readings per day, providing effectiveness and certainty to stakeholders.  Findings are generally comparable with the AAMR pilot in terms of usage and offences the AAMR was imposed for, although some differences were observed around the use of AAMR as a standalone requirement (28% vs 40% standalone requirements in pilot). It is encouraging to see that the compliance rate of 92% has remained consistent with the pilot. Process learning 1 A total of 378 orders were issued but due to data quality issues on NPS systems MOPAC was not able to include all of the cases in the analysis. 2 For the purpose of measuring compliance we have recorded an unsuccessful completion when alerts about violations on the tag led to enforcement action being taken by the Offender Manager that led to a breach conviction at Court. 4  Overall the AAMR has been accepted well across London by all delivery stakeholders as an additional tool that is tailored and specific to addressing alcohol related offending; a tool that has not been available to them previously.  Training was well attended and received. Stakeholders who attended the training were more supportive and aware of the programme than those who did not.  Stakeholders are trusting of the AAMR technology and recognised it’s capacity to facilitate engagement with other services, to target underlying issues, and allow offenders a ‘break’ in their drinking so to reflect.  Stakeholders reported that they expected more cases to be suitable however individuals had to be excluded largely due to alcohol dependency.  Stakeholders remain positive about the potential impact of the AAMR on offenders; however practitioners expressed a desire to see the longer term evidence around potential impacts.  While most offenders were relatively optimistic about the AAMR and had confidence in completing it successfully, as reported in the pilot research, the size and weight of the tag emerged as issues. A number of offenders (23/56) reported some discomfort; that it influenced their general well-being (7/56); some raised the issue of stigmatisation and negative influence on financial situations wearing the AAMR.  The strongest theme to emerge around learning related to the scale up challenges: o Delays in tagging offenders (52%, n=160, of offenders were tagged within 24 hours during the full roll-out compared to 82%, n=83, in the pilot) possibly reflecting issues with the taggers covering wider geographical areas. o More awareness raising and improved communication around the AAMR: Less respondents in the pan-London roll-out felt satisfied with the awareness raising information provided by MOPAC (69%, n=37/54 vs 78%, n=31/40) and many of the promotional guidance items had not been seen by many stakeholders in the Pan London cohort compared to the pilot (for example, the leaflet was seen by 57%, n=30/53, in the roll-out cohort compared to 80%, n=35/44, in the pilot). If further training or information is rolled-out it may be preferable to ensure all relevant organisations are represented. Additionally, stakeholders would like to be kept informed of performance of the programme on a regular basis. o The majority of stakeholders reported positively about the usefulness of the AAMR, however this was 15% less than the pilot cohort (78%, n=40/51 vs 93%, n=38/41). Interestingly little difference was seen between these cohorts when stakeholders attended training (92%, n=480/524). This highlights the positive impact that training by MOPAC and AMS has had but also the importance of continual awareness raising outside of formal training events by MOPAC. o Significantly fewer stakeholders from the pan-London cohort said that the AAMR has allowed them to develop better working practices with partners (49% decrease compared to the pilot cohort). o Some of the above differences in stakeholder perceptions about the AAMR could be due, at least in part, to the different respondent characteristics between the pilot and the pan-London roll-out stakeholder survey. For example in the pilot stakeholder survey 58 respondents answered the survey and were mainly probation 5 officers (58%) whilst in the pan-London survey 60 respondents answered the survey and over half were from the courts (magistrates, legal advisors and court staff). Much of the learning encountered in the current pan London evaluation report reflects that of the initial pilot research, indicating the importance of effective implementation - something that was always going to be a challenge in scaling up to a London level. It is also important to remember that it is always challenging trying to embed an innovative programme of work into a complex landscape with changing external pressures both politically and within individual organisations. Despite this, uptake and use of the AAMR has been positive to date. This report sits as part of a wider, holistic evaluation around the roll out of the AAMR across London. As we enter year 2, additional research is currently being conducted around performance and process analysis, proven reoffending analysis to explore the impact of AAMR on offending behaviour and a full cost benefit analysis. These elements will be reported on in the second interim report in late spring 2018 (at the end of the two year roll out) and the final report in winter 2018. 6 Introduction It is widely understood that alcohol use contributes to criminal behaviour, particularly around violent crime and public disorder in the UK. It was last estimated that the total cost of alcohol related harm was £21 billion, with alcohol recognised as a major cause of attendance at Accident and Emergency departments and hospital admissions (Public Health England, 2014a). Historically, as measured by the CSEW approximately half of violent incidents are related to the influence of alcohol. Whilst there has been a slight decline (from 53% in 13/14 to 40% 15/16) - it remains clear that there is a longstanding resistant association between alcohol and violence. Recent trends suggest that alcohol is present among violent incidents that are likely to occur in a pub or club (83%), at the weekend (49%), during the evening/night time (48%) (ONS, 2017), with victims also more likely to receive greater injuries (ONS, 2015). Wider data also contributes to the picture - around a fifth (21%) of Londoners think that people being drunk or rowdy in public places is a problem (MOPAC Public Attitude Survey (PAS), Financial Year 2016/17), a trend that has remained stable over the previous year3. Similar views are held by London businesses who perceive people being drunk / rowdy in their local area to a problem (26%) (MOPAC Business Attitude Survey, 2014 – 2016). Aside from violence, it is clear that other crime types (e.g., see McSweeney, 2015) such as driving offences, can also contribute to cause considerable harm. In 2015 for example, 37,578 defendants were convicted ‘Driving / attempting to drive with excess alcohol’, the majority of whom received a fine (76%) and a further 16% received a Community Order. Offences of this type have the potential to cause considerable harm; the total number of casualties of all severities in drink drive accidents in Great Britain in 2014 was 8,270. Of these, 1,310 were killed or seriously injured (Department for Transport, 2015a). The links between criminal behaviour and alcohol are intrinsically linked, with large costs to both the public purse and public health and wellbeing. Many attempts have been made to address alcohol use over the years from both a criminal and health related perspective. This report will review relevant interventions and further discuss the introduction of a compulsory sobriety programme introduced across London to address alcohol related offending. Sobriety Programmes and Interventions: Following a number of popular drug and alcohol sobriety programmes (e.g. 12-step Programme, Alcoholics Anonymous), the focus within sobriety interventions has shifted from addressing addiction, towards specific behavioural implications such as offending. The primary approach has been the monitoring of alcohol-use through various methods; from random sobriety check-points and ignition interlocks in addressing driving under the influence (DUI) (Roth, Marques & Voas, 2009; Bergen, Pitan, Shults & Sleet, 2012; Blais & Dupont, 2005) to continuous monitoring via transdermal tags, urine and blood testing to reduce alcohol-fuelled offending (Dougherty, Charles, Acheson, 3 The PAS explores the views of the residents across London around crime, ASB and policing issues via a face to face interview with over 12,800 respondents per year. In quarter 3 2015/16 20% of Londoners thought that people being drunk or rowdy in public places was a problem. 7 John, Furr & Hill-Kapturczak, 2012). The focus of many of these programmes has been around assessing the efficacy of the equipment and compliance with the programmes’ ethos. While few studies have explored the impact of such interventions, those that have, have published promising results; most notably, the South Dakota 24/7 Sobriety Programme (Kilmer, Nicosia, Heaton & Midgette, 2013). This programme sought to reduce DUI offences using transdermal tags and/or twice-daily breathalyser tests to encourage complete abstinence from alcohol. Primarily targeting repeat offenders, the programme combines constant alcohol monitoring with ‘swift’ and ‘modest’ sanctions - those who breach are immediately taken into custody or court (Kilmer, Nicosia, Heaton & Midgette, 2013). Compared to counties where a 24/7 sobriety programme was not implemented, results suggested a 12% reduction in DUI repeat-arrests and a 9% reduction in arrests related to domestic abuse across the 5 years following the intervention. A similar project assessing the use of Secure Continuous Remote Alcohol Monitor (SCRAM) tags (Flango & Cheesman, 2009) had mixed results but provided vital learning. Within this study a small difference, albeit not significant, (2.8%) in recidivism across two years was seen between those who wore a SCRAM tag and matched controls, increasing by over 10% when restricted to prolific offenders with at least two prior convictions (12.9%). While unable to infer impact from these results, the data provided learning in terms of the timeliness of the order. Those who wore the tag reoffended more quickly than controls and often to a greater extent following tag-removal. Investigating this trend further, Flango and Cheesman (2009) found this effect was mediated by the length of the order; those who wore the tag for at least 90 days reoffended at around half the rate of those who did not wear the tag (10.3% vs. 21.2% respectively) whereas recidivism for those who wore the tag for less than 90 days being almost equal to controls. It is unclear at this time what the longer term behaviour effects are from participating in a programme of enforced sobriety and Axdahl (2013) suggested that behavioural effects beyond tag removal may be short-lived. Comprehensive evidence around the efficacy of enforced sobriety is both lacking and mixed. Despite this, an intervention of this kind has been well-implemented across a number of US counties and states (Kilmer et al, 2013) and the Alcohol Abstinence Monitoring Requirement (AAMR), inspired by the approach in South Dakota, provides an opportunity to address alcohol related offending under UK criminal justice legislation. Alcohol Abstinence Monitoring Requirement: In 2012 a new sentencing power was introduced as part of the then Mayor’s manifesto pledge to address the significant problem of alcohol related offending in London. Under the Legal Aid, Sentencing and Punishment of Offenders (LASPO) Act 2012, the Judiciary are allowed to impose the punitive Alcohol Abstinence Monitoring Requirement (AAMR); a requirement that necessitates offenders abstain from alcohol for a fixed time period of up to 120 days. Compulsory sobriety is measured via regular testing via a transdermal alcohol monitoring devise (a tag around the offenders ankle) as part of a Community or Suspended Sentence Order4, and when this is not complied with, the offender will be prosecuted under breach proceedings and punished further. 4 Providing the offender is deemed both suitable and eligible (see Appendix A for criteria). 8 The 2014/2015 Pilot: An initial 12 month pilot of the AAMR was commissioned in south London by the Mayor’s Office for Policing And Crime (MOPAC) commencing in July 2014. This was subject to a process and performance evaluation5. Over this time, 113 AAMR Orders were imposed by the Courts. The AAMR pilot had a final compliance rate6 of 92%7 which compares favourably with other orders - analysis by the NPS in 2014 estimated a compliance rate of 61% for other community based Orders it managed with the Community Rehabilitation Company (CRC)8. Variation was also seen with different requirements - 82% of offenders completed Unpaid Work Orders successfully9 in London compared to Alcohol Treatment requirements (80%) and Drug Rehabilitation requirements (67%) (Ministry of Justice, 2015). The research indicated the AAMR was received well, particularly by the judiciary and professionals, who recognised the AAMR as an important ‘tool in the box’. This can in part be attributed to the strength of the design and implementation of the programme. There were clear toolkits and training provided, effective partnership working and a project management team in place with relevant experience in this area. The effectiveness and certainty provided by the technology, as well as a strong understanding of the aims of the pilot and how the AAMR works in practice amongst both offenders receiving the order and stakeholders involved in its delivery also helped. In addition, there were a number of associated positive consequences of the pilot, including but not limited to; the period of abstinence gave offenders a ‘pause’ in their drinking; it also provided time for reflection of their alcohol consumption and the impact it has on offending behaviour, work and relationships; and an opportunity was provided for offenders to break their cycle of routine drinking. The AAMR was also used as a ‘teachable moment’ in some instances, with products such as tailored advice and relevant literature supplied by the service providers to support offenders further. The report also stressed the potential challenge in sustainability and the scale up challenge in moving from a small scale pilot to a pan London approach. This initial pilot research was not able to explore robust impact on reoffending due to sample size and follow-up time limitations. 5 https://www.london.gov.uk/sites/default/files/aamr_final.pdf 6 Caution needs to be applied when interpreting the completion and compliance rate of AAMR – this was a pilot study with a small sample size enabling the project manager to provide some assurances that the large majority of the AAMRs were enforced when failures to comply arose. This may not necessarily be the case with other Orders and requirements. 7 The AAMR pilot had a compliance rate of 92%, based on the number of cases (n=9) who were returned to court and convicted on breaching their AAMR as a proportion of all cases imposed. Of these nine, five had their AAMR revoked and failed to complete, and the remaining 4 completed their AAMR following their return to court. This gives a final completion/compliance rate of 95% (Pepper, & Dawson, 2016). 8 However there are caveats to be considered such as a direct ‘like for like’ comparison is not possible due to different offence types, offender characteristics, breach processes and the length of the orders themselves. 9 These figures should be caveated however as the AAMR project manager recently reviewed other requirements on Orders (i.e. UPW), finding there were numerous occasions when breaches were not enforced and cases were simply closed, which may distort the actual compliance rate. 9 Pan London Roll Out: The pilot was considered to be a success and following the Conservative’s government’s manifesto commitment to make sobriety tags available across England and Wales, MOPAC and the Ministry of Justice agreed to joint fund the roll out of the AAMR pan London from the 1 April 2016 until the 31 March 2017. The pan London Roll out provides the opportunity to test the effectiveness of the South London pilot on a much larger, more complex scale. To achieve this we will ensure that the core elements of the South London proof of concept pilot are maintained, including the use of transdermal tags enabling a full and comparative evaluation to be completed. The London Sobriety Project will aim to test: - Learning from the original pilot; - Take up the AAMR requirement by the Judiciary; - Compliance with the AAMR; - Completion rates of the AAMR; - Re-offending rates Building upon the findings from the MOPAC pilot study, this report outlines initial findings from the first year of the Pan London roll out of the AAMR. It details the implementation process of the AAMR through the views and experiences of stakeholders involved in delivery and offenders sentenced to wear the alcohol tag and analysis of performance metrics. These findings sit as part of a wider holistic evaluation around the roll out of the AAMR across London, which also includes continued performance and process analysis, proven reoffending analysis to explore the impact of AAMR on offending behaviour and a full cost benefit analysis. These elements will be reported on in the second interim report in late spring 2018 (at the end of the two year roll out) and the final report in winter 2018. 10 Methodology Both quantitative and qualitative research methods were employed to triangulate learning and gain an understanding of the AAMR working processes, how the AAMR is performing and experiences of offenders and stakeholders. A variety of methodologies were used to collect data10, including:  Training feedback surveys: Training sessions were held throughout London for AAMR practitioners such as probation staff and the judiciary. At each training event surveys were distributed to gather information around training delivery, understanding the AAMR technology, understanding eligibility criteria, levels of confidence in using the technology and motivation for using the technology. 546 stakeholders completed this brief survey after attending a training event. No differences across training feedback findings from different LJA’s were identified so findings reported here are across all LJA’s.  Stakeholder surveys: An online survey was completed by 64 stakeholders to explore their understanding and experiences of AAMR (see Appendix B). Distribution was completed in waves, to reflect the staggered approach of AAMR roll out across the 9 Local Justice Areas (LJAs)11 in London, with survey’s ‘going live’ approximately one month after each LJA started to use AAMR’s. A follow up email was sent four weeks later to encourage stakeholders to respond. Respondents were largely based at Magistrates courts (46%, n=31/67), within the National Probation Service (NPS) (19%, n=13/67) and Her Majesty Courts and Tribunal Service (HMCTS) (13%, n=9/67) more widely, with Magistrates, Legal Advisors and Probation Service Officers being the most common responders.  Stakeholder interviews: To gain a more in-depth understanding of stakeholder’s views, face-to-face and telephone interviews were conducted. A total of 19 semi-structured interviews were conducted with a variety of stakeholders including practitioner and strategic levels across London12 (see Appendix C). Topics included: understanding the rationale, partnership working, training, implementation, usage, time taken for different aspects of the AAMR, decision making, suitability, perceived impacts, good practice, lessons learned, challenges, and broader attitudes to the equipment.  Offender surveys: Offenders who received an AAMR were asked to complete a survey at the time the electronic tag was fitted, and when it was subsequently removed. The two surveys sought to understand their first impressions of the tag, perceptions of what life maybe like whilst wearing the tag and once it has been removed. Surveys were given to the offender by the 10 Given the size of the research cohort (e.g. the number of respondents to the stakeholder survey/offender surveys), caution should be used when considering the results. Response base size is provided, however this varies as not all respondents answered every question. 11 London Local Justice Areas (LJAs) include: Central, South, South East, South West, West, North West, North, North East and East. 12 Potential interviewees were identified with the AAMR project manager and contacted via email by the researchers. There was no obligation to participate, therefore participants were self-selecting. Interviews were both face-to-face and conducted over the telephone, depending on the wishes of the interviewee. Where possible, interviews were audio-recorded and detailed notes were taken in all interviews, and analysed to draw out themes. 11 Electronic Monitoring Services (EMS) tag fitter13. Completing the survey was not compulsory, and some offenders chose not to participate. In total, 169 (out of a possible 356, 47%) completed the survey at the time when the tag was initially fitted, and 115 (out of a possible 29714, 39%) completed it during tag removal15.  Performance monitoring data: A range of performance data was gathered from both the NPS/CRC and EMS - the company which conduct the field delivery and assist in data collection of performance metrics about the tag. Performance metrics included: numbers of AAMR given, types of offences, court details, demographics on who received the tag, number of breaches and compliance with tag. 13 Whilst this method of distribution has its limitations, this was the most practical approach available for obtaining insightful data on offender perceptions and experiences. 14 This accounts for offenders who have completed their AAMR requirement and are no longer an ‘active’ case. 15 Due to the way the data was anonymously collected, it is not possible to link survey responses to know whether offenders who completed the initial survey also completed the removal survey 12 Results Using the AAMR: Performance Learning. Imposing the Requirement: A total of 367 AAMR’s were imposed during the initial year of the pan London rollout (April 2016 and March 2017)16 and 17% of these (n=59) remain active cases, in that the offender is still subject to the requirement. Following the initial pilot, the AAMR was introduced across London in a phased approach and therefore some LJAs have had the opportunity to impose AAMR for considerably longer than others. In fact, the AAMR has been used five times as much from the first quarter (FY 2016/17 Q1: average AAMR sentences/month = 13) to the last quarter (FY 2016/17 Q4: average AAMR sentences/month = 52), with AAMRs being imposed most frequently from Magistrates Courts (93%, n=342/356) (see full breakdown of AAMRs issued by Courts in Appendix D). Use of the AAMR has also varied across court and LJA. Following the South London LJA hosting the yearlong pilot (and continuation immediately after pilot period), it is unsurprising this area had already gained momentum and had the confidence to use the AAMR in comparison to other LJAs where roll out has been phased throughout the past year. Whilst Croydon Magistrates Court (17%, n=58/342) continues to be the court that imposes the most AAMRs, it would appear that without the ‘focus’ of a specific pilot project, there has been a slight decline in the use of AAMR in what was initially classified as the most ‘active’ borough in the pilot – with 82 sentences including an AAMR in the pilot compared to 63 AAMRs in the past year in Croydon17. However, other courts such as Thames and Barkingside which were last to roll-out AAMR in January 2017 have performed well and have awarded 16 and 17 AAMRS over three months. To be considered for an AAMR, the offence must be alcohol related, however this leaves a broad scope of the type of offences committed. In total, the 367 AAMRs were ordered in relation to 63 differently classified offences (see breakdown of offences in Appendix E). In light of the literature (for example McSweeney, 2015) it is unsurprising that the majority of AAMR’s were imposed for offences of violence (31%, n=115) or drink drive related offences (22%, n=82). Although the eligibility criteria meant domestic abuse offenders were excluded from an AAMR recommendation, data suggests that 10% (n=38) of the offenders had been highlighted as such18. Overall, compulsory sobriety was usually imposed as a requirement of Community Orders (74%, n=270), with nearly a third being given as a standalone requirement (28%, n=100/357)19. This is significantly fewer (p<.05) standalone AAMR’s than used during the pilot programme where nearly 40% of Orders used AAMR alone. This may indicate that those sentencing in court may feel that the AAMR requires the support of additional requirements, or that the AAMR only addresses a specific element of their offending behaviour (alcohol use). When the AAMR was used in conjunction with other requirements there appears to be a large variation in the types of requirements it was paired with. Table 1 details the most frequently combined requirements with a AAMR. 16 In fact NPS data records indicate that 376 AAMR’s were imposed in this time period, however data from the NPS and EMS was only available for 95% (n=356) of cases that will be used as the total base size throughout this report. 17 Magistrates and Crown court combined. 18 This domestic abuse flag indicates that at some point in their lives they have been a domestic abuse perpetrator, however it is not necessarily relevant to the current offence. NPS advised that caution needs to be applied around the reliability and robustness of this flag. 19 Requirements data not provided by NPS for 11 offenders. 13 Table 1: Additional requirements paired with AAMR. Requirement Number of Order’s Percentage AAMR Standalone 100 28% AAMR + UPW 69 19% AAMR + Rehabilitation Activity Requirement (RAR) 57 16% AAMR + Prohibited Activity 24 7% AAMR + UPW + Prohibited Activity 19 5% AAMR + RAR + UPW 16 5% AAMR + RAR + Accredited Programme 10 3% Other requirements20 62 17% The AAMR Tag: A requirement dictating compulsory abstinence from alcohol is initiated immediately when the Order is given in Court. Where possible, EMS should aim to fit the tag within 24 hours of an offender receiving the sentence, and in 90% (n=332) of cases EMS received notification from the court on the same day (n=279) or the following day (n=53). Despite speedy notification to the tagging company, only 50% (n=167/332) of offenders (where EMS were notified the same or next day) were tagged on the same day (n=65) or within one day of notification (n=102)21. This is a significant reduction (p<.05) from the pilot study (82%), which may reflect a resource issue deriving from the scaling up of the project and the EMS tag fitters ability to cover a much larger geographical area. This finding supports the potential changes to the operational model that will be piloted in year 2 in two courts (Westminster and Bromley). There are plans to introduce ‘tagging at source’ where offenders are tagged immediately after sentence either at the court house or nearby Probation office by a probation officer, with AMS supplying the equipment and EMS monitoring performance metrics only. This may prevent the delay in offenders receiving the tag and thus also ensure that the Order can be monitored and enforced from the earliest opportunity. In total over this reporting period 367 Orders were imposed. Of those offenders who have completed the AAMR, 92% (n=235) did so successfully22. During this period there were 75 offenders who were issued warning letters for non-compliance and discretion was applied in 32 of cases This provides a compliance rate that matches the pilot (92%). Of the 270 successfully completed Orders, offenders were subject to the tag for an average of 67 days (range 7 days – 120 days). This reflects an 8 day reduction in average length the AAMR compared to this requirement in the pilot (average = 75 days). In total, over 825,000 transdermal readings were taken by the tags in the initial year, an average of 48 readings a day, indicating that the technology underpinning the tag is working as intended. 20 This includes cases where there was no match n=15 21 EMS must attend on the day for all orders issued before 3pm, or within 24 hours for orders received after 3pm 22 For the purpose of measuring compliance we have recorded an unsuccessful completion when alerts about violations on the tag led to enforcement action being taken by the Offender Manager that led to a breach conviction at Court. 14 Offender Demographics: In terms of basic demographics, the majority of offenders who were sentenced to an AAMR were male (86%, n=314) and white (47%, n=171)23. Whilst this follows the trend of the pilot’s findings, there is a significant decrease (p<.05) in the proportion of white offenders receiving this Order (pilot: 66%). Additionally, across London the average age of offenders who received the AAMR matched the pilot cohort, averaging 33 years old (range 18 years – 74 years), with nearly two thirds (63%, n=226) aged between 18 and 34 years. The CRC and NPS conduct risk of harm assessments on all offenders who receive either a Community Order or Suspended Sentence Order. Nearly two thirds of those sentenced to an AAMR were assessed as medium risk of harm (61%, n=225). A mere four offenders were considered to be high risk of harm. On the whole this reflects that the AAMR is being used correctly. The Offender Group Reconviction Scale version 3 (OGRS3) scores for the AAMR pan London cohort were calculated at the point of receiving a Court Order by the NPS or CRC. OGRS uses static factors such as age at sentence, gender, offence committed and criminal history to predict the likelihood of proven reoffending within either one or two years after starting their Court Order. Offenders with a high OGRS score are at greater risk of reoffending. As a group (n=31724), the OGRS 2 year score was 37% (ranging from 5 to 97) – indicating that just over one third of offenders would be predicted to reoffend within two years (see Appendix F). This is comparable to the OGRS score of the pilot AAMR cohort (OGRS Year 2 = 35%). This shows that those receiving the AAMR are broadly identified as low and align to the general offending population in the UK, particularly those who receive community sentences (Farrington, 2005, Ministry of Justice, 2015b). 23 Including: ‘White: British/English/Welsh/Scottish/Northern’, ‘White: Irish’, ‘White: Gypsy or Irish Traveller’, and ‘White: Other’. 24 OGRS data not provided by NPS/CRC for 50 offenders. Key Learning: 367 AAMRs were imposed over the first year of the pan London roll out programme. The majority of AAMRs were given as part of a Community Order, with an average length of 62 days. The tag technology appears to be working as intended, taking on average 48 readings per day. Standalone AAMRs accounted for 28% of all Orders and multiple requirement Orders usually consisted of AAMR and Unpaid Work (19%). Offenders receiving the AAMR were largely white males with an average age of 33 years. Tags were given for a variety of crime types, most commonly in relation to violence or drink driving offences. As expected the AAMR cohort did not have an extensive criminal background, with an average two year OGRS score of 37%. In total over this reporting period 367 Orders were imposed. Of those offenders who have completed the AAMR, 92% (n=235) did so successfully1. During this period there were 75 offenders who were issued warning letters for non-compliance and discretion was applied in 32 of cases This provides a compliance rate that matches the pilot (92%). Findings are generally comparable with the AAMR pilot in terms of usage and offences the AAMR was imposed for and the continued high compliance rate, although some differences were observed around the use of AAMR as a standalone requirement (28% vs 40% standalone requirements in pilot). 15 “Even though we [probation officers] try our best to make sure that the magistrates, benches, justices etc. know about the AAMR, they just aren’t on board; they really do take some persuading...they do not have enough training to make them feel more open to it.” Using the AAMR: Process Learning from Stakeholders. Throughout the AAMR pan-London roll-out, stakeholders were surveyed to gain an understanding of the wider London perspective around the Order and to detect any changes in understanding or use, given its expansion from the initial pilot. This section details the views and experiences of stakeholders (including staff at the NPS, Magistrates Courts and HMCTS) around the training and awareness raising, working practices, using the AAMR and their views on the impact the AAMR has had on offenders. Training and Awareness: MOPAC and AMS delivered training to more than 1,250 London Magistrates, 50 District Judges, over 130 Legal Advisors and to all National Probation Service Court Teams. Overall, the training sessions have been well attended and received, with attendees’ finding it a positive and useful experience (97%, n=238/24525) (see Appendix G for a breakdown of all responses) and willing to use the AAMR in future (95%, n=499/526). In particular the training helped them to understand how to use the AAMR in their job role (93%, n=509/545) and understand the aims and objectives of the project (97%, n=528/545); a finding reflected in the wider stakeholder survey (93%, n=53/57). Whilst those who attended the training strongly indicated that they knew how to apply the eligibility criteria (97%, n=516/545), this was not so clear from respondents to the wider survey (who may not have attended the training course). Compared to the pilot, here stakeholders felt they had less understanding of the criteria (77%, n=44/57, p<.05) than those in the pilot cohort (95%26, n=38/40) despite more respondents using AAMR in their role (61% roll-out n=31/51 vs. 48% pilot n=19/40). It was felt that training would have benefitted from more ‘case study’ type information to demonstrate instances where a AAMR had been given and how it had worked in practice, as well as the opportunity to see the tag working ‘in action’ to really understand its implications. Interviewees also raised an issue around inconsistency in understanding of the AAMR across different organisations. A variety of awareness raising paraphernalia (i.e. AAMR poster, website and toolkit) around the AAMR have been produced, with the AAMR leaflet noted by respondent’s as the most useful method of information and communication, however this document has still been seen by less stakeholders than in the pilot (57%, n=30/53 vs. 80%, n=35/44). Indeed, many of the promotional/guidance items had not been viewed by stakeholders who responded to the survey27. In comparison, interviewee’s detailed their use of the professional documentation/toolkit, which they have found to be “very easy to read and straightforward”. The reasons for the decline in 25 Note that not all attendee’s to the training completed every question of the survey feedback; therefore base sizes vary for each question. 26 Pilot results are based on the collation of data from 3 surveys administered throughout the pilot project, due to low response rate (n=58). Eighteen respondents completed more than one survey throughout the pilot period, where possible, base sizes have been adjusted to reflect this. Survey respondents who had responded to previous surveys were not required to answer all questions in later surveys therefore response numbers differ depending on question. 27 AAMR documentation not seen by respondents: poster = 72%, website = 83% and toolkit = 59%. 16 “[AAMR is] the cherry on the cake” “I didn’t even know that MOPAC were involved with the AAMR!” awareness are complex and could relate to the larger size and scope of the pilot, i.e., more staff, more areas, the challenges in implementing and overseeing a pan London innovation. As noted in the pilot evaluation report, the level of involvement from the project manager was not necessarily sustainable as the project has been expanded more widely. This has impacted the evaluation in particular, especially around accessing relevant stakeholders and data, thus impeding the robustness of the conclusions that can be derived. This has been reflected in the struggle to recruit stakeholders to participate in research interviews and complete the survey. There was a desire for additional awareness raising and improved communication around the AAMR gleaned from stakeholder interviews and also less stakeholders were satisfied with the awareness raising information provided by MOPAC (stakeholder survey: 69%, n=37/54); a figure that has declined from the pilot cohort as the project has expanded across London (pilot cohort: 78%, n=31/40). Specifically there was a lack of information available in the Courts to other staff members, solicitors, or offenders, with a number of interviewee’s noting that “more awareness raising is needed” outside of the immediate people involved. Awareness raising was also raised by those at training - only half of stakeholders who attended the training (53%, n=28/53) reported being satisfied28 with communication from MOPAC, the project commissioners, compared to 85% in the pilot study. However it is apparent that satisfaction with the awareness raising and communication is improved if stakeholders have attended the training, rather than relying on their own self learning or gathering information from a colleague or documentation. If further training or information is rolled out, it may be preferable to ensure all relevant organisations are represented or focus on information sessions specifically for court staff. Additionally, respondents from both the surveys and interviews with stakeholders raised that they would like to be kept informed of AAMR performance, in terms of compliance rates, and uptake of the requirement on a regular basis so they could understand the trends and usage across London. Working practices: The AAMR directly addresses low level and non-dependant alcohol related offending behaviour - a requirement that, until now, has not been at the courts disposal. This has previously been an area of frustration to both sentencers and report writers, who felt there was very little they could do to directly affect specific underlying factors (e.g. non-dependant drinking) of someone’s offending. The AAMR has also been welcomed due to its “easy to understand and self-explanatory nature”, with an appreciation that “it’s useful to have another requirement” than the long-standing options previously available. In general the AAMR has been accepted positively by all stakeholders. However, the survey data suggests the roll out of the AAMR programme pan-London may have dampened appreciation or acceptance of the requirement. While the majority of respondents felt the AAMR was a useful additional tool for their role (78%, n=40/51), this did see a significant decrease (-14%, p<.05) compared to the pilot cohort (93%, n=38). Though, a different picture is seen in those who have attended training, where support for the order is comparable to that seen in the pilot (92%, 28 Categorised from where respondents have responded with ‘very satisfied’ or ‘fairly satisfied’. 17 “the AAMR is…ideal for people who have maybe had too much to drink on a Saturday night and become a nuisance and committed criminal damage etc.” “if it [AAMR] can open the door to other things such as counselling, talking to their GP, taking medication, talking to family and friends I am more likely to recommend it”. n=480/524), highlighting one of the benefits of attending training and the importance of continued awareness raising of the programme as it moves forward. One of the aims of the evaluation was to explore the effect that AAMR has on offender management processes and possibility of reducing staff workloads. Although the pilot report only briefly explored this theme, there was no firm evidence of this from the pan-London stakeholder surveys, where less than 10% of stakeholders agreed the AAMR speeds up their workloads29. However, AAMR is not considered to be particularly onerous. Significantly fewer stakeholders (17%, p<0.05) felt the AAMR had increased their workload, compared to the pilot cohort. However, this may be explained by the demographics of the pilot survey respondents who tended to be probation staff who would be more engaged in the delivery process (both at the pre-sentence report stage and managing offenders who receive the requirement) than magistrates, more of whom responded to the pan-London survey. A key element of AAMR, particularly around implementation of the new initiative, is the need for good partnership working between agencies such as CRC, NPS and the judiciary. However since rolling out the programme across London, significantly fewer stakeholders felt that the AAMR allowed them to develop new relationships with partners (49% decrease, p<.05) or improve their relationships with existing partners (55% decrease, p<.05). The stark differences observed in the pilot may be due to the smaller scale of the scheme, located in a small geographical area with high political and media interest. The dedicated pilot AAMR project manager pro-actively encouraged team working and delivered cross-agency training which likely facilitated communication between partners, however this may not have been possible as the programme rolled out across London. In particular MOPAC had to stop the Local Implementation Group sue to the size of roll-out and resources available. Using the AAMR: There are many factors that can influence the decision making process of whether to recommend/sentence an offender to a period of compulsory sobriety. Findings indicate that for many delivery stakeholders this process most commonly focuses on the facts of the case and a pattern of behaviour that emerges from alcohol use. However, stakeholders also suggested that ‘preventing people from committing alcohol-related crime’ was the most important goal of the AAMR, as well as ‘improving public confidence in the ability of the criminal justice system to tackle alcohol-related crime’. Stakeholders stated that they also took these factors into consideration when considering an AAMR. 29 74% disagreed with the statement “Offenders on the AAMR are dealt with more quickly than those on other orders”. 18 “AAMR is unrealistic for the majority of alcohol-related offending… we see a lot of dependent drinkers, and even the 'binge' drinkers often require more in depth treatment to address the psychological reasons underlying their alcohol abuse”. “the tide might be turning but we’ve still got a long way to go in terms of it being something that the courts think of really quickly and take on as a sentence”. “it also has the parallel benefit of giving people the chance to stop drinking and have some time to reflect on the impact that alcohol has on their life”, “assess whether … it will help them identify that alcohol is causing them problems”. The importance of identifying whether there are additional underlying issues that could also be addressed in a rehabilitative way, in addition to its punitive impetus was also highlighted by stakeholders. In particular, many stakeholders reported that they would consider the AAMR if they felt it would help offenders with their alcohol use as: and a period of compulsory sobriety provides an opportunity to: However, challenges have arisen around the usability of the requirement which have affected decision making. It was identified that the strict eligibility criteria for the requirement meant it was often difficult to recommend, as many offenders did not meet the criteria set out using the AUDIT screening tool or had perpetrated domestic abuse, which was something that was also found in the pilot evaluation. Additionally, issues of practicality and suitability such as nickel allergies, having no fixed abode or the nature of employment, were cited. Interviewees also referred to the issue of alcohol dependency, which meant they were unable to recommend the AAMR even if they wanted to. It is clear that there are many factors to consider in the decision making process that have influenced the usage and uptake of AAMR, but it is important to remember that it is often difficult to introduce new initiatives into already complex situations involving a multitude of agencies (Stanko & Dawson, 2015). Interviewees commented that the courts, at the beginning, did not have confidence in the requirement and were reluctant to impose it. However: 19 “the only thing we’re told anecdotally during training was that offenders had addressed their drinking because of being on the AAMR” and “more updates on how successful the AAMR cases have been would be really useful...my inspiration for recommending the AAMR is that it will make a difference, and it would be good to have the evidence to back it up.” This has been compounded further by the pan-London roll out coming at a time when London CRC is experiencing wider disruptions and transformations. Challenges have arisen due to internal changes in terms of policies, new ICT and staffing structures, which have meant that the “new senior team … don’t have as much knowledge about the AAMR”. This hampers progress as there is not the senior level support to champion new initiatives, and one interviewee described that such disruptions “do not support the imbedding of good practice”. Potential effects on offenders: When complied with – the AAMR requires abstinence from alcohol, which could have an impact upon the individuals offending as well as other areas (i.e., health, education, wellbeing etc.). The current report does not cover offending at this stage; however, some insights around impact can be sought from other sources, such as staff and the offenders themselves. With the roll-out of AAMR across London, stakeholders’ and offenders perceptions around its potential impact have been sought. Overall, interviewees were generally positive about the potential effect this requirement may have in terms of reducing offending behaviour and alcohol intake, although the majority of those surveyed were unsure about the longer term effects; a change from the pilot cohort. Now we see slight reductions across the board in the perceptions on impact from the pilot, but small sample sizes preclude firm statements here. Stakeholders stated they did not know if AAMR will stop people committing crime in the long term (63%, n=32/51; pilot: 55%, n=22/40), help people to drink less alcohol in the long term (55%, n=28/51; pilot: 40%, n=16/40) or help people to play a more positive role in society (53%, n=27/51, pilot: 40%, n=16/40). The wider applications of the AAMR in terms of rehabilitation have been acknowledged, as it provides opportunities for reflection and further intervention. When explored through interviews, a frequent suggestion was additional health benefits may also stem from this period of compulsory abstinence, providing a pause in their drinking to “evaluate whether they think they need to cut back or stop completely going forward”; a message strongly identified during the pilot. The learning, around the wider rehabilitative application of AAMR, has been reiterated through the pan London AAMR training, which may explain the frequency at which it has been reported during the current evaluation. Domestic abuse perpetrators: One theme to emerge from the interviews was the use of AAMR with DA perpetrators. This was excluded in the pilot but, within the Pan London approach, there has been a plan to conduct a 20 “If AAMR is to be used in DA cases it should only be after very careful and thorough assessment…Prohibiting alcohol consumption when it is a coping strategy for someone could actually raise the risk unless other protective factors and supports are put in place…” “I wanted to recommend someone for the AAMR…but they were from Middlesex and committed the offence in London. This is a common issue” feasibility study exploring the use of AAMR with DA perpetrators. Overall, practitioners broadly supported the concept of AAMR with DA, 70% (n=35/50) would support rollout to DA perpetrators and two thirds of respondents (66%, n=33/55) felt AAMR could be a useful and effective tool for cases of domestic abuse. Although it should be noted that some concern around this issue was raised, with the need for additional measures to be put in place for these cases emphasised: Such concerns should be explored within the feasibility study and associated evaluation and will be reported on in year two. Looking forward: This programme has another year to run and given its popularity, the majority of stakeholders would support the roll out of AAMR nationally (86%, n=43/50), with hopes that a wider roll-out would overcome difficulties where offences were committed in London by those residing elsewhere, which are linked to the night time economy. In terms of learning, there have been suggestions for using the AAMR in a more creative manner. For example, respondents recognise that compulsory abstinence may not be the most suitable requirement for an offender immediately, but an AAMR could be used in a staggered approached with other requirements, thus providing the offender with treatment and support to reduce their alcohol intake and ensure they are better able to cope with sobriety. A period of enforced monitored abstinence may then as motivation to continue, especially considering that producing negative drugs tests can boosts people motivation and self-esteem. Despite some minor scale up challenges, it is an exciting time for the AAMR programme, with new innovative ways of using AAMR being tested as the project moves forward into its second year pan London, such as the DA feasibility study and tagging at source. Overall, this requirement has been well received by delivery stakeholders, who appear to have sufficient knowledge and understanding of the AAMR to conduct their work effectively, but would appreciate more performance information around the AAMR and evidence of long term impact. The offenders, who are subjected to wearing the tag, have also been given the opportunity to express their views about this innovative programme - their views are heard next. 21 Key Learning: The AAMR has generally been received well across London by all delivery stakeholders, and training has been well attended, although there is still a need for more awareness raising and communication around AAMR, especially from MOPAC. Better facilitation of partnership working at these events and outside of training would also prove advantageous. Stakeholders have welcomed the additional tool that is tailored and specific to addressing alcohol related offending – a tool not previously available. But there have been barriers to uptake during the pan-London roll-out namely, reorganisation of the CRC and cases not meeting the eligibility criteria, which has led to fewer cases being recommended than expected, largely due to issues around alcohol dependency and offences of domestic abuse. Also issues of practicality and suitability have reduced the use of the AAMR. Whilst the AAMR is thought to have wider application than just to reduce offending and alcohol intake, such as having a rehabilitative effect, providing the offender the opportunity for reflection and further intervention, it is worth noting that significantly fewer stakeholders thought this compared to the pilot cohort. 22 “the tag is too big and is uncomfortable to wear” “I am unable to exercise, which I love doing, because the tag hurts my ankle” Using the AAMR: Process learning from offenders. To gain a better understanding of the impact of the AAMR tag on offenders, EMS tag fitters delivered surveys to 169 offenders at tag fitting and a further 115 offenders at tag removal (see Appendix H for full breakdown of responses). This is a far larger number than was able to be delivered in the pilot research and brings a unique voice to the discussion. Entry Survey - Expectations and concerns: In general, the majority of those surveyed appeared to have relatively good relationships with family (90%30, n=120/133) and friends (92%, n=122/133), somewhere suitable to live (86%, n=111/129) and a relatively good sense of well-being31 (75%, n=95/127). However, over a third of respondents suggested they were currently experiencing financial difficulty (38%, n=46/122) despite the majority being in an enjoyable job (77%, n=83/108), possibly reflecting the experiences of many in the current economic climate. At the start of their Orders however, offenders were broadly optimistic, although with clear subsets that were more mixed, with many feeling the AAMR would improve all aspects of their lives (‘life in general’: 65%, n=79/122) (see Appendix H) including family relationships (50%, n=60/119), current financial situation (62%, n=75/121), offending behaviour (67%, n=77/115), physical health (70%, n=87/127) and mental wellbeing (60%, n=71/119). This may be indicative of the recognition of AAMR’s ability to infiltrate other aspects of people’s lives, beyond their offending behaviour. Despite having received a AAMR order, when asked to reflect on their alcohol use, just under half felt that drinking alcohol had a negative effect on their lives (43%, n= 54/126) and 60% (n= 76/127) felt that going out and socialising with their friends did not cause them any problems. However, nearly all offenders (93% n=116/125) felt they understood why they had received the tag and felt confident32 they would successfully complete the order (99%, n= 113/132). Which is reflective of the high compliance rate with this requirement (92%) While the potential benefits of the requirement were understood by offenders, as demonstrated in the pilot findings, practical concerns were raised about the tag itself. Offenders commented on the size and weight of the tag. Additionally, concerns were raised around health and well-being implications, such as bathing restrictions, exercise and sleep. All of these are consistent with views from the pilot research. The pilot also highlighted that the size of the tag could lead to issues around stigmatisation (“People assume you are dangerous”33). This was reiterated through the roll-out entry surveys, where 41% (n=46/112) of offenders were worried about what their friends and families would think, with some expressing concern that the tag would make these relationships worse (“Relationships with family” 5%, n=6/50, “Relationships with friends” 7%, n=8/60). Concerns around stigmatisation also extended to work and education; while more respondents felt the tag would have no impact on their 30 Based on collated ‘Strongly Agree’ and ‘Agree’ responses. 31 Measured by those who stated they ‘are happy most of the time’. 32 Based on collated ‘Fairly Confident’ and ‘Very Confident’ responses. 33 Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC 23 “This has put pressure on my already tight finical situation; I work in a corporate environment and have had to buy all new clothes to cover the Tag”. “I work as a project manager consultant for various blue chip companies, having to be dressed smartly all the time, this tag is very big and can easily be noticed… I'm very worried about this…it will be very hard to hide.” employment or education, (44% (n=52/117) and 49% (n=56/114), respectively), a small proportion (14%, n=16/117) felt the tag would make their employment situation worse: Exiting the AAMR order - were concerns realised? Once an offender had completed the AAMR Order, the tag was removed and offenders were asked to complete another survey. Overall, offenders continued to report living a stable life with good relationships with friends (89%, n=93) and family (86%, n=90), health (83%, n=87), happiness (75%, n=78) and accommodation (80%, n=84). Small changes were observed in some of the feedback, for example, some evidence of perceived stigmatisation being realised; 4% more respondents were worried about what friends and family thought about the tag, after finishing the order (45%, n=47) than before it had commenced (41%, n=46), however, these are very small scale changes. Overall, when asked whether they felt the AAMR had made things better, worse or had no impact, the trend in the data suggests that by and large, offenders felt the AAMR had made their lives better. (over 40% of offenders across 6 of the areas) or had no impact on them (over 30% of offenders across 7 of the areas) (see Figure 1). When considering these results, caution needs to be applied due to the small percentage changes and it should be noted that the survey methodology relied on a between-subjects design. At the time of writing, it was not possible to link responses from the entry survey to those at the exit survey; this means that these changes cannot be directly attributed to the tag itself. Furthermore, participants responded to this question on a 5-point Likert scale from Strongly Agree - Strongly Disagree, with little opportunity to expand on why they have given the answers they have. 24 Key Learning: Offenders reported generally being satisfied with their lives when they initially received the Order, maintain good relationships with friends and family, accommodation and well-being, which continued post tag removal. However, offenders did raise practical concerns about the tag itself citing concerns about financial implications due to having to buy new clothes to cover the tag, health and wellbeing concerns, as well as concerns about the size and weight of wearing the tag that may lead to stigmatisation. Figure 1: Offender Entry and Exit survey responses 25 Using the AAMR: Cost implications Implementing a new innovative programme comes with associated costs, however it is important to understand whether these costs have been beneficial, whether that is through a reduction in reoffending or a speedier course through the criminal justice system, requiring fewer resources. Throughout the evaluation attempts were made through surveys to collect data around the time taken to assess, prepare, induct, enforce and manage AAMRs in comparisons to other requirements. However these attempts were not successful either because respondents found it difficult to quantify due to the varying nature of cases and workload, or more commonly, because their job role did not provide the opportunity to work directly with AAMRs ‘on the front-line’. For example, a number of Magistrates completed the survey, and whilst they are the individuals who impose the requirement and hear the breach cases, the AAMR would not be a factor that would affect the time they spent dealing with the case. By contrast, Probation Officers or Probation Service Officers, who had responsibility for managing the cases, and thus would have a good indication of the time required to deal with an AAMR case, were not well represented in the survey responses. Therefore limited conclusions can be drawn around whether AAMR affects the amount of time required to deal with a case, however the general feeling was that in most cases, the AAMR was largely similar to other requirements. In order to fully understand the benefits that may be gained from the AAMR, throughout year 2 a full cost benefit analysis will be conducted, providing the opportunity to consider implications for reoffending and wider impact on health and other partners. 26 Discussion The AAMR was introduced as a pilot concept in 2014 to address the significant problem of alcohol related offending in London. This new sentencing power enabled courts to impose, as part of a Community Order or Suspended Sentence Order, a requirement that compelled an offender to abstain from alcohol for a fixed time period and be regularly tested via electronic monitoring. The AAMR is now a pan London programme that has been generally welcomed by delivery stakeholders as an additional option to specifically address alcohol related offending behaviour. Following the success of the initial south London pilot in 2014, this programme provided the opportunity to test the effectiveness of AAMR on a larger, more complex scale. This report helps to build the evidence base around compulsory sobriety electronic monitoring, detailing a process evaluation generating learning through the views and experiences of stakeholders involved and offenders who were sentenced to this requirement. In many respects, the roll out of AAMR has been viewed positively by both the judiciary and Probation staff, often reflecting the views and experiences highlighted in the pilot evaluation report (Pepper & Dawson, 2015). There appears to be continued confidence in the electronic monitoring equipment which is reflected in the steady uptake of the AAMR and now that all LJAs are ‘live’, year 2 will provide a clearer landscape of the overall uptake. Whilst there has been overall acceptance of the project across London, it appears that the impact of this requirement has been minimal for those that work in the field. It is now another “tool in their toolbox” that equips them to do their job better. It has been incorporated easily into everyday working, which may in fact be a testament to the ease of application of this innovative programme. However, there have been some concerns about cases not meeting the eligibility criteria due to alcohol dependency, allergies and other issues, which has meant that the judiciary have not been able to recommend the AAMR as much as they would have liked to. This would indicate that although the eligibility criteria is restrictive, the AAMR is actually being used appropriately by the judiciary. The main goal of the AAMR was, stakeholders suggested, to prevent people from committing further alcohol related crime. Although a greater number of stakeholders in the pan London roll-out did not know if the AAMR would have long lasting effects in stopping people committing crime, or drink less alcohol,. Indeed, previous research would suggest that although offending may reduce whilst subject to this type of requirement, behavioural effects beyond tag removal may be short-lived (Axdahl, 2013). Despite this there was still a general perception that the AAMR would provide offenders with a pause in their drinking behaviour and the opportunity to reflect on their lives. When implementing a large scale programme of work across London, it is expected that challenges and difficulties may arise (Dawson & Stanko, 2013). Indeed, the AAMR pilot was considered a success due to the positive implementation, mainly down to clear communication and consistent engagement with stakeholders from the outset, and a dedicated programme manager with ‘on the ground’ experience and established contacts. As noted in the pilot evaluation report, this level of involvement from the project manager was not necessarily sustainable as the project has been expanded more widely. This has impacted the evaluation in particular, especially around accessing relevant stakeholders and data, thus impeding the robustness of the conclusions that can be derived. This is reflected in the struggle to recruit stakeholders to participate in research interviews and complete the survey. Despite many efforts to engage stakeholders in the research, there has 27 been a definite reluctance, evidenced by the considerable decline in participation since the pilot study. Reports are, particularly from the CRC, that staff members do not have the time or inclination to participate, due to substantial restructuring and organisational changes that are currently taking place. It appears that they are in unstable times and understandably not prioritising the AAMR. Further challenges have arisen since the scale up around both the practicalities of tagging the offenders, and stakeholders and staff working with the AAMR. For example, since the pilot there has been delays in getting offenders fitted with the tag within one day of receiving their court Order, which may reflect a resource issue from the tagging company now that the ‘taggers’ are expected to travel across London. For those working with AAMR, the new initiative has not provided as many opportunities to build partnership relations that had previously been available during the pilot. Again, this may not be reflective of the implementation of AAMR itself, but more a reflection of the current offender management landscape. Looking forward, since the start of the pan London roll out, MOPAC were successful in a bid to the Home Office’s Police Innovation Fund to undertake development work. This includes the testing of AAMR for domestic abuse perpetrators and designing efficiencies within the current model to enable sustainability. One way to do this is through ‘Tagging at Source’ which aims to administer the electronic monitoring tag to the offender at the Court or probation office rather than their home address. These elements are still in development, however as they progress forward, MOPAC’s Evidence and Insight team will include these in the overall evaluation. Finally, this report is just one element of the evaluation taking place around the roll out of AAMR across London. A holistic approach is being taken overall, including continued process evaluation (focusing on large scale implementation, stakeholder views, offender experience), cost benefit analysis, performance metrics and the impact of AAMR on proven reoffending. In late spring 2018, at the end of the second year, an update report will be produced focusing upon performance and process, and early impact on the pan London cohort (e.g. 6 + 3 months reoffending). A final impact report, focusing on proven reoffending and cost analysis will be produced in winter 2018. 28 References Axdahl, L. (2013). Analysis of 24/7 Sobriety Program SCRAM Participant DUI Offense Recidivism. Mountain Plains Evaluation, LLC. USA: Salem Bergen, G., Pitan, A., Shults, R, QU, S. & Sleet, D. (2012). Current evidence on publicised sobriety checkpoint programmes: are they still effective? Injury Prevention, 18, A45-A46. Blais, E. & Dupont, B. (2005). Assessing the capability of intensive police programmes to prevent severe road accidents. British Journal of Criminology, 45,6, 914-937. Cattell, J., Kenny, T., Lord, C. and Wood, M. (2014b). Community Orders with Punitive Requirements: Results from the Offender Management Community Cohort Study. London: Ministry of Justice Department for Transport (2015a). Reported Road Casualties in Great Britain: Estimates for Accidents Involving Illegal Alcohol Levels: 2013 (final) and 2014 (provisional). London: Department for Transport Department for Transport (2015b). Self-reported drink and drug driving: Findings from the Crime Survey for England and Wales 2014/15. London: Department for Transport Department for Transport (2016). Reported road casualties in Great Britain: Estimates for accidents involving illegal alcohol levels: 2014 (final) and 2015 (provisional). London: Department for Transport. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/543627/rrcgb- drink-drive-final.pdf Dougherty, D., Charles, N., Acheson, A., John, S., Furr, R. & Hill-Kapturczak, N. (2012). Comparing the Detection of Transdermal and Breath Alcohol Concentrations during Periods of Alcohol Consumption Ranging from Moderate Drinking to Binge Drinking. Experimental and Clinical Psychopharmacology, 20, 5, 373 – 381. Farrington, D. P. (2005). Integrated Developmental and Life-Course Theories of Offending. New Brunswick, NJ: Transaction Flango, G. & Cheesman, F. (2009). The effectiveness of the SCRAM alcohol monitoring device. Drug Court Review, 6, 2, 109-134. Kilmer, B., Nicosia, N., Heaton, P. & Midgette, G. (2013). Efficacy of frequent monitoring with swift, certain, and modest sanctions for violations: Insights from South Dakota’s 24/7 Sobriety Project. American Journal of Public Health, 103, 1, e37 – e43. McSweeney, T. (2015). Calling time on alcohol-related crime? Examining the impact of court- mandated alcohol treatment on offending using propensity score matching. Criminology and Criminal Justice, 15, 4, 464 - 48. Ministry of Justice (2015). Prison and Probation Performance Statistics 2014 to 2015. London: Ministry of Justice. 29 Ministry of Justice (2015b). Criminal Justice System Statistics Quarterly: December 2014. London: Ministry of Justice Nicosia, N., Kilmer, B. and Heaton, P. (2016). Can a criminal justice alcohol abstinence programme with swift, certain and modest sanctions (24/7 Sobriety) reduce population mortality? A retrospective observational study. The Lancet Psychiatry. Published online February 9th 2016. Office for National Statistics (ONS) (February 2016). Overview of violent crime and sexual offences, Nature of Crime Table 3.2 Pepper, M. & Dawson, P. (2016). Alcohol Abstinence Monitoring Requirement: A process review of the proof of concept pilot. MOPAC Roth, R., Marques, P. & Voas, R. (2009). A note on the effectiveness of the house-arrest alternative for motivating DWI offenders to install ignition interlocks. Journal of Safety Research, 40, 6, 437–441. 30 Appendix A: Eligibility and Suitability Criteria In order to be eligible for an AAMR, the following conditions must be met:  Consumption of alcohol must be an element of the offence or an associated offence, or the court must be satisfied that consumption of alcohol was a factor that contributed to the offender committing the offence or an associated offence;  The offender must not be dependent on alcohol;  The court must not include an alcohol treatment requirement (ATR) in the order (ATRs are for dependent drinkers only);  The offender must live in London;  The offence must not have involved domestic abuse (at the present time). The Probation Court Team officer will assess “suitability” with alcohol AUDIT Tool. This assessment is used to determine the offender’s alcohol dependency levels. 31 Appendix B: Survey Respondents Organisation Job Role Local Justice Areas represented No. of respondents NPS Probation Officer / Probation Service Officer /Court Based Officers Central, South West, South, South East, East 10 CRC Probation Officer / Probation Service Officer North West, South, South East 3 Strategic Leads / Administrators 4 HMCTS / Courts Legal Advisors North West, South, South East, South West, West 10 Court Staff 6 Magistrates 21 MPS Police Officer South East 1 EMS/AMS Project Team South, South East, West 5 Overall Total 60 Figure 2: London Justice Areas represented in the stakeholder survey. 32 Figure 3: Job titles of stakeholders who responded to the stakeholder survey. Figure 4: The organisation breakdown of all stakeholders who responded to the stakeholder survey. 33 Appendix C: Interviewee Respondents Full breakdown of those who agreed to be interviewed. Organisation Job Role Local Justice Area No. of interviewees NPS Probation Officer - Court Based Central 3 North 3 South 1 South East 3 Probation Officer / Probation Service Officer North 2 North West 1 Probation Prosecutor South 1 CRC Probation Officer / Substance Misuse South 1 West 2 Courts Deputy Justice Clark South West & West 1 Legal Advisor South East 1 Overall Total 19 34 Appendix D: Breakdown of Courts that have imposed AAMR. Magistrates Court No. of AAMR imposed Roll-out date Croydon MC 58 Apr-16 Camberwell Green MC 29 Apr-16 Wimbledon MC 27 Apr-16 Bromley MC 20 Apr-16 Bexley MC 11 Apr-16 City of London MC 3 Jul-16 Ealing MC 8 Jul-16 Hammersmith MC 19 Jul-16 Feltham MC 3 Jul-16 Uxbridge MC 49 Jul-16 Westminster MC 29 Jul-16 Willesden MC 4 Oct-16 Hendon MC 21 Oct-16 Highbury Corner MC 28 Oct-16 Thames MC 16 Jan-17 Barkingside MC 17 Jan-17 Total 342 Crown Court No. of AAMR imposed Roll-out date Woolwich CC 4 Apr-16 Inner London CC 2 Apr-16 Croydon CC 5 Apr-16 Central Criminal Court 6 Jul-16 Isleworth CC 3 Jul-16 Kingston Crown Court 2 Jul-16 Harrow CC 2 Oct-16 Wood Green CC 1 Oct-16 Total 25 35 Appendix E: Offence type This table describes the offence type for which an AAMR was ordered, 01 April 2016 to 31 March 2017. Offence Type No. of offences % Damage/Theft Burglary 4 1% Criminal Damage 22 6% Theft 4 1% Unauthorised taking of a motor vehicle 4 1% Driving Driving or attempting to drive whilst unfit through drink/drugs 82 22% Failing to provide specimen 12 3% Other driving 8 2% Drugs Failure to cooperate (drugs) 1 0% Supply/Possession of drugs 5 1% Harassment Harassment 7 2% Racially aggravated harassment 14 3% Threatening words or behaviours 21 6% Public Order Drunk and disorderly conduct 7 2% Offences against Public Order 1 0% Other offences 3 1% Sexual Exposure 5 1% Sexual Assault 9 3% Violence Assault (beating, common assault, ABH, GBH) 115 31% Assault on Police Officer 29 8% Possession of a weapon 12 3% Resisting/Obstructing a Police Officer 2 1% Total 367 100% 36 Appendix F: OGRS3 scores for offenders who received an AAMR. OGRS score No. of AAMR offenders % Very low (0 - 24%) 105 33% Low (25% - 49%) 107 34% Medium (50% - 74%) 77 24% High (75% - 89%) 20 6% Very high (90% +) 7 2% Total 316 100% 37 Appendix G: Stakeholder training survey The table below indicates both the feedback questions and responses34 provided by stakeholders who attended the AAMR training. 34 Responses were on a 7 point Likert Scale from 1 = Strongly Agree to 7 = Strongly Disagree. Scores ‘1’, ‘2’, and ‘3’ have been combined into one output ‘Agree’ and scores ‘5’, ‘6’, and ‘7’ have been combined into ‘Disagree’ category. Question Agree Neutral Disagree Total The trainers did a good job delivering the training event 525 96% 2 0.4% 18 3% 545 I was given enough time to ask questions 464 85% 4 1% 78 14% 546 The training was delivered at the right pace 481 88% 16 3% 48 9% 545 I understand the aims and objectives of the AAMR 528 97% 3 1% 14 3% 545 I understand how to use the AAMR in my role 509 93% 1 0.2% 35 6% 545 I understand the offender eligibility criteria for the AAMR 516 95% 4 1% 22 4% 542 I understand how the AAMR equipment works 512 94% 3 1% 31 6% 546 I know where to get more information/support around the AAMR 415 82% 36 7% 55 11% 506 I see the AAMR is a useful way to tackle alcohol related offending 480 92% 24 5% 20 4% 524 I will use the AAMR 499 95% 6 1% 21 4% 526 The first session was useful 238 97% 2 1% 5 2% 245 The second session was useful 164 96% 4 2% 3 2% 171 The third session was useful 119 94% 3 2% 5 4% 127 Overall this training has provided me with enough information to confidently use the AAMR 495 96% 4 1% 18 3% 517 Overall I was satisfied with this training event 505 97% 0 0% 15 3% 520 38 Appendix H: Offender Survey Questions Entry Survey: Total number of respondents = 169 Question Strongly Agree/ Agree Neither agree nor disagree Disagree/Strongly Disagree Total Responses35 I have a good relationship with my close family 90% 120 5% 7 5% 6 133 I have a good relationship with my friends 92% 122 5% 7 2% 3 133 I have good physical health 86% 114 8% 11 5% 7 132 I have a nice place to live 86% 111 8% 10 6% 8 129 I have a job which I enjoy 77% 83 9% 10 12% 13 108 Going out and socialising with my friends causes me problems 18% 23 21% 26 60% 76 127 My current financial situation is difficult 38% 46 30% 37 30% 37 122 My offending behaviour causes me problems 46% 57 18% 23 32% 40 125 Drinking alcohol has a negative effect on my life in general 43% 54 21% 27 34% 43 126 I am happy most of the time 75% 95 17% 21 9% 11 127 I understand why I received the Alcohol tag 93% 116 0% 0 7% 9 125 I understand what I must do to comply with the Alcohol tag order 99% 124 0% 0 1% 1 125 I am worried about what my friends and family will think of the Alcohol tag 41% 46 0% 0 59% 66 112 35 Total Responses includes ‘Don’t Know’ responses and missing data, therefore data shown does not necessarily sum to the data shown in Total Responses. Not every respondent answered every question so there is variation to the total number of respondents overall (n=169). 39 Entry Survey 36 Offending behaviour - 'Better' may refer to less offending and 'Worse' may refer to more offending. Question Better Worse No impact Total Life in general 65% 79 12% 15 19% 23 122 Relationship with family 50% 60 5% 6 40% 48 119 Relationships with friends 45% 54 7% 8 45% 54 121 Money/current financial situation 62% 75 7% 8 29% 35 121 Offending behaviour36 67% 77 0% 0 24% 28 115 Physical health 70% 87 4% 5 22% 27 124 Mental wellbeing 60% 71 8% 10 29% 34 119 Housing situation 39% 47 5% 6 52% 62 120 Employment situation 35% 41 14% 16 44% 52 117 Ability to ‘go out’/socialising 33% 39 28% 33 35% 42 119 Education situation 28% 32 1% 1 19% 56 114 40 Exit Survey: Total number of respondents = 115 37 Total Responses includes ‘Don’t Know’ responses and missing data, therefore data shown does not necessarily sum to the data shown in Total Responses. Not every respondent answered every question so there is variation to the total number of respondents overall (n=115). Questions Strongly Agree - Agree Neither agree nor disagree Disagree-Strongly Disagree Total Responses37 I have a good relationship with my close family 86% 90 2% 2 4% 4 105 I have a good relationship with my friends 89% 93 1% 1 2% 2 105 I have good physical health 83% 87 5% 5 4% 4 105 I have a nice place to live 80% 84 5% 5 5% 5 105 I have a job which I enjoy 66% 69 12% 13 7% 7 105 Going out and socialising with my friends causes me problems 10% 11 27% 28 52% 55 105 My current financial situation is difficult 31% 33 22% 23 37% 39 105 My offending behaviour causes me problems 33% 35 18% 19 38% 40 105 Drinking alcohol has a negative effect on my life in general 35% 37 17% 18 36% 38 105 I am happy most of the time 74% 78 11% 12 7% 7 105 I was worried what my friends and family thought of the alcohol tag 45% 47 0 0 55% 58 105 The alcohol tag felt comfortable to wear 27% 28 0 0 73% 77 105 The AAMR guidance document I received was useful 71% 75 0 0 29% 30 105 41 Exit Survey 38 Offending behaviour - 'Better' may refer to less offending and 'Worse' may refer to more offending. Questions Better Worse No impact Total Life in general 49% 51 11% 12 27% 28 105 Relationship with family 41% 43 5% 5 40% 42 105 Relationships with friends 37% 39 7% 7 43% 45 105 Money/current financial situation 46% 48 6% 6 33% 35 105 Offending behaviour38 53% 56 4% 4 23% 24 105 Physical health 58% 61 6% 6 26% 27 105 Mental wellbeing 49% 51 9% 9 31% 33 105 Housing situation 30% 31 3% 3 55% 58 105 Employment situation 25% 26 15% 16 49% 51 105 Ability to ‘go out’/socialising 30% 32 27% 28 31% 33 105 Education situation 25% 26 4% 4 50% 53 105 Briefing What: LGE: Completed Cases Offending Analysis to inform PBR (Year 2) Date Provided: Wednesday 12th December 2018 To: Julia Mlambo & Caroline Tredwell From: Tom Davies & Lynne Grossmith (Evidence & Insight) Overview London Gangs Exit (LGE) provided data on 126 completed cases up to and including September 2018. 69 cases were closed between October 2016 and September 2017 (Completion Year 1); these cases featured in the previous PBR analysis in December 2017. A further 57 cases were closed between October 2017 and September 2018 (Completion Year 2). This briefing note provides a preliminary recorded offending and victimisation analysis on this cohort, as well as some initial exploration of longer term impact based on all completed cases to date. Methodology Analysis presented below is indicative only, therefore caution should be exercised when interpreting findings. The low case numbers, restricted period of analysis, relatively low prevalence of offending, and before and after methodology mean that robust causal statements as to LGE’s impact on offending behaviour are not possible at this stage of the evaluation. Additionally, it should be noted that only completed cases are included in the analysis at this point. The final evaluation report seeks to utilise a matched control methodology, allowing for a more robust exploration of impact. Ongoing, closed and completed cases will also be included in the analysis. Analysis compares all police recorded offending and victimisation six months before the date of referral to LGE, with the six months following. This time period allowed all but one of the completed cases to be included (n=56). Recorded crime incidents where the individual appears as EITHER Suspect OR Accused are included to maximise the offending sample. Only broad categories are included to maximise aggregated data1. Completed Cases: Descriptive analysis There is little difference in the age or ethnic distribution of completed cases between year one and year two. There has been an increased proportion of females completing LGE in Year Two. • Average age of 19.3 (Youngest 13, Oldest 28) • 70% males (n=40); 30% females (n=17) • 89% BAME (n=109)2; 11% White (n=6). 1 These categories are: all offending, Violence Against the Person Offending (VAP), all victimisation, VAP victimisation. Due to matching procedure victimisation data is only available for individuals who have also appeared as suspects/accused in the data period retrieved, therefore individuals with no offending data are excluded from victimisation analysis (n=23) 2 Majority Black African or Caribbean heritage (63%; n=36). Case related information: • All cases (n=57) were opened3 between January 2016 and June 2018 and closed between October 2017 and September 2018. • A PNC ID was identified for all but nine of the individuals. • 16% (n=9) of individuals appear on a recent version of the MPS Gang Matrix (October 2018).4 Preliminary Offending Analysis • Average offending marginally increased and violent offending showed no change when comparing the six-month period before and after case opened date (see Table 1). • Average victimisation (both overall and violent*) decreased in the six-month period following case opening. Caution should be exercised in interpreting these results in line with the caveats in methodology. • Individual offending and victimisation was very low prevalence in the time-period analysed: o All offending (Max=5; Violent Offending (Max=3) o All victimisation (Max=5); Violent victimisation (Max=4) • 36 of the 57 individuals had no recorded offending in the period of analysis (that is, in the six months either side of their LGE referral). o Of the 21 with recorded offending in the period of analysis, 9 committed a lower number of offences after being referred to LGE, whilst 10 committed more. • 23 individuals had no recorded victimisation during the period of analysis.5 o Of the 11 individuals with recorded victimisation, 9 experienced a decrease and 2 experienced an increase in victimisation. Table 1: Offending and Victimisation for 57 completed Cases (6 months before/after LGE referral)6 3 Due to data extraction issues from Safer London Lamplight database, a mixture of ‘case referred’ and ‘case opened’ dates were provided; this will be standardised for the final report. 4 Final report will include those who have ever appeared on Matrix in addition to currently appearing. 5 Out of 34 eligible * statistically significant (Wilcoxon Signed Rank) Before After Before After 56 56 33 33 All Incidents 23 24 16 6 Individuals 13 15 9 5 Proportion 23.2% 26.8% 27.3% 15.2% Cohort Average 0.41 0.43 0.48 0.18 Violence Incidents 7 7 14 4 Individuals 5 5 8 4 Proportion 8.9% 8.9% 24.2% 12.1% Cohort Average 0.13 0.13 0.42 0.12* Cohort Offending Victimisation • For all offending, 13 individuals were responsible for the 23 offences in the 6 months before LGE referral, compared to 15 individuals recording a total of 24 offences in the 6 months after referral. • For violence against the person, 5 individuals were responsible for the 7 offences before referral; 5 individuals recorded a total of 7 offences after referral. • 9 individuals were victims of a total of 16 offences before referral; 5 individuals were victims of a total of 6 offences after referral. • 8 Individuals were victims of a total of 14 violent offences before LGE referral; 4 individuals were victims of a total of 4 violent offences after referral. The decrease is statistically significant. Longer Term analysis of Year 2 Cohort Analysis also explored longer term impact using a 12 month period either side of case opening; this reduced the cohort by nine (n=45). Results are similar, likely to be indicative of the different types of referrals accepted in Year 2: • For all offending, 20 individuals were responsible for the 48 offences in the 12 months before LGE case opening, compared to 21 individuals recording a total of 44 offences in the 6 months after opening. • For violence against the person, 11 individuals7 were responsible for the 19 offences before case opening; 9 individuals recorded a total of 18 offences after case opening. Victimisation analysis provides more positive indications, although the same caveats apply: • 13 individuals were victims of a total of 21 offences before case opening; 4 individuals were victims of a total of 5 offences after case opening. The decrease is statistically significant. • 16 Individuals were victims of a total of 12 violent offences before LGE case opening; 3 individuals were victims of a total of 3 violent offences after case opening. The decrease is statistically significant. Longer Term Analysis: All Completed Cases Further exploratory analysis was conducted on longer time periods before and after referral date, with findings similar to those reported in the October 2017 interim evaluation report. This will be developed in the version of this paper for publication by MOPAC in early 2019. 12(3) Months Analysis explored offending and victimisation within the entire completed cases cohort (n=126) for 12 months before/after with a three month exclusion period from the referral date to allow time for the 7 Out of an available cohort of 29 intervention to take effect. This reduces the cohort in scope for analysis (Offending, n=101; Victimisation, n=71)): • Overall offending decreased, though not significantly. This is driven by the Year 1 cohort, with offending in Year 2 increasing slightly (not significant). • Violent offending marginally increased amongst the cohort; the increase was not statistically significant. o In the 12 months before case opening, 27 individuals were responsible for 49 Violent offences; in the 12 months after the 3 months exclusion period 26 individuals were suspect or accused in 51 violent offences. • Consistent with other periods of analysis, victimisation and violent victimisation decreased both overall, and within Year 1 and Year 2. All decreases for the 12(3) time-period are statistically significant (*). Table 2: Offending and victimisation for completed cases by year and overall The preliminary analysis described above indicates some key changes in the offending and victimisation patterns of those completing in Year 2 compared to Year 1. A greater proportion of Year 2 completions had no offending history before or after case opening, across all time-periods explored. Similarly, where offending is recorded, the Year 2 cohort includes less prolific individuals. The Year 2 cohort were more likely to be victims, and particularly victims of violence. This may in part be explained by changes in the types of individuals referred and accepted (i.e. more females at risk of violence or exploitation from gangs). This will be explored fully in the final evaluation report. Next Steps These findings are preliminary and data and findings are subject to change. The final evaluation report, which explores both the process of implementation and any impact of LGE, is scheduled for Early 2020. Analysis will utilise a matched control methodology, allowing for a more robust exploration of impact and factoring in needs and interventions received to understand ‘what works’. Before After Before After Before After Offending All Cohort 69 69 32 32 101 101 Offences 118 96 30 35 148 131 Offenders 34 29 15 17 49 46 Proportion Off 49.3% 42.0% 46.9% 53.1% 48.5% 45.5% Cohort Average 1.71 1.39 0.94 1.09 1.47 1.30 Violence Offences 38 37 11 14 49 51 Offenders 20 18 7 8 27 26 Proportion Off 29.0% 26.1% 21.9% 25.0% 26.7% 25.7% Cohort Average 0.55 0.54 0.34 0.44 0.49 0.50 Victimisation All Cohort 49 49 22 22 71 71 Victim Incidents 30 14 15 3 45 17 Victims 17 11 10 3 27 14 Proportion Vic 34.7% 22.4% 45.5% 13.6% 38.0% 19.7% Cohort Average 0.61 0.29* 0.68 0.14* 0.63 0.24* Violence Victim Incidents 21 8 14 2 35 10 Victims 15 7 10 2 25 9 Proportion Vic 30.6% 14.3% 45.5% 9.1% 35.2% 12.7% Cohort Average 0.43 0.16* 0.64 0.09* 0.49 0.14* All Completed Cases: 12(3) Months Year 1 Year 2 Overall Police, Camera, Evidence: London’s cluster randomised controlled trial of Body Worn Video October 2015 November 2015 Lynne Grossmith Catherine Owens Will Finn David Mann Tom Davies Laura Baika © Jointly owned by the College of Policing Limited and the Mayor’s Office for Policing And Crime (MOPAC) – 2015 This publication (excluding logo’s) is licensed under the terms of the Non-Commercial Government Licence v1.0. To view this licence visit http://www.nationalarchives.gov.uk/doc/non-commercial- government-licence/non-commercial-government-licence.htm or email psi@nationalarchives.gsi.gov.uk For a Welsh version visit http://www.nationalarchives.gov.uk/doc/non-commercial-government- licence-cymraeg/non-commercial-government-licence-cymraeg.htm Where information featured in this publication is being republished or copied to others, the College of Policing and the Mayor’s Office for Policing And Crime must be identified as the source of the information and their copyright ownership acknowledged with the following attribution statement: Contains copyright jointly owned by the College of Policing Ltd and the Mayor’s Office for Policing And Crime (MOPAC) and licensed under the Non-Commercial Government Licence v1.0. Where we have identified any third-party copyright information, you will need to obtain permission from the copyright holders concerned. This publication is available for download at: http://whatworks.college.police.uk/Research/Pages/Published.aspx Any enquiries regarding this publication or to request copies in accessible formats please contact research@college.pnn.police.uk or evidenceandinsight@mopac.london.gov.uk 1 Executive Summary Overall the findings suggest there are potential benefits of Body Worn Video (BWV), although those related to criminal justice outcomes were not fully realised during the timescales of the trial and need the support of criminal justice partners to be achieved.  BWV can reduce the number of allegations against officers, particularly of oppressive behaviour. Complaints related to interactions with the public also reduced and, although it did not reach statistical significance, the trend in overall complaints was consistent with these findings.  There was no overall impact of BWV on the number or type of stop and searches conducted. In addition, there were no differences in officers’ self-reported behaviour relating to how they conducted stops.  No effect was found on the proportion of arrests for violent crime. When an arrest had occurred, there was a slightly lower proportion of charges by officers in a BWV team.  There was no evidence that BWV changed the way police officers dealt with victims or suspects.  The Public Attitude Survey found, in general, London residents are supportive of BWV, with their opinions of the technology positively associated with their views of how ‘procedurally just’ the police are, and their confidence in the MPS.  Officers reported a range of innovative uses of BWV, including professional development; use of intelligence; and sharing information with partners and the public. Background Some early evaluation work undertaken in the United Kingdom and the USA, has shown promising evidence that BWV can increase the proportion of violent incidents which end in a criminal justice (CJ) outcome and reduce complaints against officers. The cluster randomised controlled trial (RCT) reported in this paper tested the impact of BWV on complaints against the police, frequency of stop and search and CJ outcomes for violent incidents in ten Metropolitan Police Service (MPS) boroughs between May 2014 and April 2015. The BWV intervention involved: training officers to use BWV; allocation of a personal issue BWV; and on-going supervision and guidance on use. In each of the ten MPS boroughs included in the trial, five Emergency Response Teams (ERTs) were randomly assigned to either the treatment group, who wore BWV cameras (two teams), or the control group (three teams). Each of the ten boroughs started the trial at different points in the year-long evaluation, depending on when they received their cameras. Once more than 50% of officers in a treatment team uploaded an operational clip1, the borough was deemed ‘live’ for evaluation purposes. In total throughout the trial, 814 officers in 19 teams were assigned to wear cameras and 1,246, in 29 teams were assigned to not receive cameras2. Randomly assigning teams of officers to treatment and control groups provides a strong basis on which to draw inferences regarding the effects of BWV, establishing ‘cause and effect’ relationships, ruling out other explanations. The focus on ERT officers in 10 boroughs means findings cannot be generalised to officers in other roles, boroughs across the MPS or other force areas. During the trial BWV captured 48,281 recordings, totalling 12,212 hours of video (average 15 minutes per clip), of which 4,678 hours (28% of clips) were tagged as ‘evidential’ for potential use within the criminal justice system (CJS). Across both the control and treatment groups, approximately 11,300 stop and searches, 261 complaints and 64,355 notifiable crime reports, 16,191 classified as violent incidents3 were included. Officer surveys, observations and interviews were conducted to understand why any changes in outcomes may have occurred, as well as to capture information on context and implementation of the cameras. Victim views were explored using the existing victim User Satisfaction 1 Operational refers to any clip uploaded that is not categorised or flagged as for training purposes. 2 The wrong allocation of a team to the treatment condition in one borough led to the exclusion of two teams from the analysis. 3 Defined as: Serious wounding; assault with injury; common assault; attempted murder; assault on a constable; affray; & violent disorder. 2 Survey (USS), in addition involvement of victims in the CJ process was monitored using MPS data. Public attitudes were captured through MOPAC’s routine Public Attitude Survey (PAS). Findings from a survey of those stopped and searched and sentiment analysis of BWV related tweets could not be included due to extremely low numbers of responses and tweets. Primary Outcomes Complaints BWV can reduce the number of allegations against officers4, particularly of oppressive behaviour5. Complaints related to interactions with the public also reduced6 and, although it did not reach statistical significance, the trend in overall complaints was consistent with these findings. Whilst there is no evidence of a change in the quality7 and type of interactions between officers and the public, during interviews officers reported particular instances where BWV changed behaviour. Officers also gave anecdotal evidence of using BWV recording to achieve early resolution of potential complaints, where they felt the footage demonstrated limited grounds, which is consistent with BWV officers reporting a feeling of greater protection from complaints8 in the survey. Stop and Search There was no overall impact of BWV on the number or type of stop and searches conducted. In addition, there were no differences in officers’ self-reported behaviour relating to how they conducted stops, but arrests as a result of a search are slightly less likely when officers are in a BWV team. Officer surveys showed no difference between officers with and without BWV in their reported use of discretion; compliance with procedures; and self-reported behaviour9 during a search. However, officers with BWV were less likely than those without BWV to agree they needed stronger justification for their actions10. Therefore, rather than affecting officer decision making, BWV may enable officers to feel more confident if challenged. Analysis of interviews suggests officers feel BWV footage can provide support for their justification to search which they did not have previously. Criminal justice outcomes BWV did not affect the proportion of arrests for violent crime. When an arrest had occurred, there was a slightly lower proportion of charges for incidents reported by officers in a BWV team. It is not known if BWV leads to the preparation of fewer, but stronger cases without further exploration of later CJ outcomes at court. When only those cases where BWV officers specifically flagged footage were analysed, a higher arrest rate was found for officers in a BWV team, but there was no difference to the charge rate. This finding could indicate officers were using cameras during arrest to strengthen existing evidence, or officers chose particular types of situation to flag BWV evidence. Although the statistical analysis suggests BWV had a limited impact on CJ outcomes, officer surveys indicate a belief that BWV helps to collect better quality evidence. However, officers perceive there were blockages in the system that prevented BWV evidence being used in later stages of the CJ process. In particular, officers interviewed felt that BWV could be valuable for incidents of domestic abuse, where footage could be used to show a level of detail and/or emotion not possible in written statements, an impact which may be seen later in the CJ process (for example at court), but has not been possible to capture during the current RCT. 4 Each complaint can comprise of a number of different allegations (for example, an officer could be alleged of unlawful arrest and discriminatory behaviour during the same incident – this would equate to one complaint, with two allegations). 5 MPS categorise the following types of allegations as Oppressive Behaviour: Oppressive conduct or harassment, Other assault, Other sexual conduct, Serious non-sexual assault, Sexual assault, Unlawful/unnecessary arrest or detention. 6 Breach Code A PACE/ Breach Code B PACE/ Discriminatory Behaviour/ Incivility, impoliteness and intolerance/ Lack of fairness and impartiality/ Oppressive conduct or harassment/ Other assault/ Other sexual conduct/ Serious non-sexual assault/ Sexual assault/ Unlawful or unnecessary arrest or detention. 7 As measured by officer self-reported behaviour, injury and assault data, and data from the victim user satisfaction survey. 8 See Annex B for statements combined to create this factor. 9 Including questions in the officer survey specifically constructed around procedural justice principles -the public’s perception of fair decision making and respectful treatment by the police. 10 See Annex B for statements combined to create this factor. 3 Contents Executive Summary ................................................................................................................... 1 Contents ..................................................................................................................................... 3 Acknowledgements ................................................................................................................... 4 1. Setting the Scene ................................................................................................................... 5 2. Methodology.......................................................................................................................... 8 3. Implementation Learning .................................................................................................... 10 4. Results .................................................................................................................................. 13 5. Conclusions and Implications .............................................................................................. 27 Annex A: Methodology ............................................................................................................ 29 Annex B: Secondary Data Sources and Analysis ..................................................................... 34 Annex C: Body Worn Video Usage .......................................................................................... 41 Annex D: Criminal Justice Case Attrition................................................................................. 46 References ................................................................................................................................ 46 Table and Figures Figure 1: Borough “Go-live” for each month of the trial ........................................................... 9 Figure 2 Average number of clips recorded by each officer per one active month ................ 11 Figure 3 Percentage of overall clips by category..................................................................... 11 Figure 4: Difference between the average number of complaints per officer in the control and treatment teams in each borough .................................................................................... 15 Figure 5: Difference between the average number of stops per officer in the control and treatment teams in each borough ........................................................................................... 18 Figure 6: Difference in the proportion of crimes with an arrest at any time, between the treatment and control teams in each borough........................................................................ 21 Figure 7: Difference in the proportion of crimes with charge/summons, between the treatment and control teams in each borough........................................................................ 22 Figure 8: Trial CONSORT diagram ............................................................................................. 31 Figure 9: Flow of Officers in the Trial ....................................................................................... 33 Figure 10: The total number of clips recorded per month by all boroughs: ........................... 41 Figure 11: The total number and total length of clips recorded per hour of the week........... 43 Figure 12: Number of clips recorded depending on the length of a clip in minutes ............... 45 Table 1: Average officer saturation, uploaded clips and evidentially marked clips ................ 32 Table 2: Borough level clips recorded during the trial ............................................................. 42 Table 3: Evidential Footage recorded per borough ................................................................. 42 Table 4: Differences in clips recorded and total length of clips via the day of the week ........ 43 Table 5: Rate of clips recorded per officer per month ............................................................. 44 4 Acknowledgements The authors from the College of Policing and MOPAC were responsible for overall design of the trial and the evaluation. Superintendent Adrian Hutchinson led the strategic delivery of the BWV programme, Inspector Ben Clark led policy, training development and the delivery management of cameras to officers, technical and back office support for the trial and Sergeant Stuart Murrell led the liaison work with boroughs. This study would not have been possible without the advice and support of a wide range of people. The authors would particularly like to thank the following for their invaluable contributions:  The police officers of the MPS who attended training, wore the cameras and gave their time to participate in the evaluation surveys and interviews.  The command team of the Metropolitan Police Service (MPS) for their support for the trial and commitment to building the evidence base in policing – especially the ACPO team: DAC Mark Simmons, DCC Steve Watson, DCC Adrian Hanstock and Commander Jeremy Burton.  The MPS staff – Nick Kettle, Dominic Quinn, Amy Howells, Tony Gallagher and the MPS BWV team, Mark Bolingbroke, Ian Cunningham, Yomi Thomas and Justina Brown for implementing the cameras, supplying data, supporting the evaluation, and embracing a different way of working.  MOPAC analyst Michael Keenan and in the College of Policing, David Brown, for supporting the trial and analysis and in the College of Policing, Rory McKenna and Shayan Moftizadeh for helping with the officer interviews and analysis.  Opinion Research Services for their work on the stop and search survey of the public.  Professor Martin Bland for his review of the analysis and advice, Dr Paul Quinton, Nerys Thomas, Dr Paul Dawson and Professor Betsy Stanko for their advice, guidance and comments, and Professor David B. Wilson and Dr Katrin Hohl who peer reviewed the report. 5 1. Setting the Scene The Mayor’s Office for Policing and Crime (MOPAC), the Metropolitan Police Service (MPS) and the College of Policing (College) have worked in collaboration to test the impact of Body Worn Video (BWV), on complaints against the police, frequency of stop and search, and criminal justice outcomes. The introduction of BWV was evaluated through a cluster randomised controlled trial which allows strong statements to be made about the impact of cameras because it can test ‘cause and effect’ relationships. This report provides an overview of the trial, summarises its main findings, and discusses implications for policing policy and practice. Background In both the UK and the US, police use of BWV is increasing. The technology is being used in a variety of ways, most often to capture police operational activity first hand, using helmet or vest mounted camerasi. The uptake of the technology may be explained by its perceived potential to assist with a range of policing problems; notably, cameras have been proposed as key to increasing the transparency, efficiency and effectiveness of police conductii. Despite the growing popularity of BWV there is much to be learnt regarding its effectiveness, particularly its role in reducing attrition through the Criminal Justice System (CJS), as well as understanding any impact it has on the nature of police/public encounters. The MPS has previously trialled BWV technology, most notably a yearlong 2008 Home Office funded pilot in four London boroughsiii. There have been further local initiatives in eight boroughs, as well as the MPS team at Heathrowiv. However, all previous MPS pilots have been locally managed and not designed or supported in a way that allowed the impact of BWV to be established. This research addresses many of the evidence gaps around the impact of BWV, and provides valuable insight for future pan-London roll out. Given growing interest in BWV across England and Wales this trial starting in May 2014 sought to test a consistent approach to the distribution of approximately 500 cameras across Emergency Response Teams (ERTs) in ten London boroughs11. The basic premise of introducing BWV was that the presence of a camera and the captured footage would improve CJ outcomes because the quantity and quality of available evidence would increase, thereby supporting victims and witnesses. In addition, it would introduce a layer of accountability for the police and public, which would impact on the quality and nature of interactions - reducing complaints and the number of stops and searches. London’s ‘Global City’ status, with around 31,000 officersv, means this trial will address an evidence gap on the impact of BWV in a larger UK force. 11 Barnet, Brent, Bexley, Bromley, Camden, Croydon, Ealing, Havering, Hillingdon, Lewisham. 6 There are three primary outcome measures in the trial, each with their own theories for change and evidence base. 1) Complaints Outcome: BWV cameras will reduce the number of public complaints made against response officers. Theory for change: • Officers and members of the public are likely to be aware of their behaviour during contact with each other; and • Unsubstantiated complaints and complaints with little grounds are likely to reduce because of availability of independent evidence. Evidence Base: One prior study has explored the impact of BWV on complaints in a test and comparison site. Operation Hyperion, which introduced BWV to all officers on the Isle of Wight (IoW), found ‘lower level’ complaints on the IoW reduced by 15% in the period after personal BWV cameras were issued, while the equivalent figure for the rest of Hampshire was a 5% reduction. Data on more serious complaints showed a reduction of 11.5% in complaints on IoW compared to an equivalent increase of 6.9% in complaints for the rest of Hampshirevi. BWV has also been tested in American trials and pilots to assess its impact on the quality of police- public interaction, use of force, and complaints. The U.S. study in Rialto (California)vii is often cited to demonstrate accountability benefits, with a reduction in complaints (from 24 to 3) and officer use of force (from 61 to 25) when comparing before and after the introduction of BWV. However this study should be understood in its context, of a Californian force with a relatively small number of officers (n=54). Several other studies found similar reductions in complaints, including Mesa (Arizona)viii and the 2007 Home Officeix research in Plymouth. There are a number of limitations with all of these studies which affects the strength of the findings. 2) Stop and search Outcome: BWV cameras will reduce response officers’ use of stop and search. Theory for change: • Officers are likely to stop and search only when they are very confident of the grounds; and • Officers are likely to carry out fewer, speculative searches. Evidence Base: There is no existing evidence on the impact of BWV on stop and search practices, with the MPS trial filling this evidence gap. In the UK and globally the decision to search, the compliance with regulations and subsequent interaction between police and the public is a contentious topic, with much research examining the impact on various communitiesx. However the effect of BWV, especially within groups who have historically had challenging relationships with the police and who may also experience high rates of stop and search, has not been addressed. Prior general researchxi shows there is a link between fairer interactions, an increase in police legitimacy and an increase in public confidence. This could be an aspirational benefit of BWV, as the greater transparency could make interactions fairer, in turn improving legitimacy and confidence. Exploring this aspect of BWV use and public relations is important, as legitimacy and the trust of communities is a key component of crime reductionxii. In addition, changes in officers’ behaviour and the public experience of contact will be measured, it is thought:  Officers are likely to feel more supported, less vulnerable to complaints, and more confident;  Officers are more likely to report better interactions and follow expected process;  Victims are more likely to feel supported and satisfied;  The public will feel more confident in the police; and 7  Interactions with the police will be reported to be better, particularly for specific groups, e.g. those being stopped and searched. 3) Criminal justice outcomes for violent crime Outcome: BWV cameras will increase the proportion of violent12 incidents attended by response officers that result in a criminal justice outcome. Theory for change: • Officers are likely to capture more and better evidence at the scene; • Officers are likely to follow expected process; • Victims and witnesses are likely to participate in the CJ process because of availability of independent evidence; • Offenders are likely to admit guilt because of availability of independent evidence; and • Officers and the public are likely to be confident that a conviction will be secured. Evidence Base: The quantity and quality of evidence added to a case is thought to be the most direct benefit to using BWV, especially when capturing emotion, recording injuries and victim statements, (particularly to pursue ‘victimless’ evidence led prosecutions13). There is strong evidence of the impact of BWV on CJ outcomes for DA incidents. A randomised controlled trial (RCT) undertaken by the College of Policing with Essex police showed that issuing officers with BWV could be effective in increasing the proportion of detections that resulted in a criminal charge (81% of the sanction detections were charges in the treatment group compared to 72% in the control group). This finding was consistent across all domestic abuse incidents regardless of initial assessment of risk by the control room. There were no differences in incidents being recorded as crimes, or rates of arrest, and too few cases to identify impact on guilty pleas and sentencing at the timexiii. There is supporting evidence about the impact of BWV in incidents of violence. A 2007 Home Office pilotxiv in one Basic Command Unit in Devon and Cornwall evaluated the use of BWV on improving the CJ outcomes in violence related incidents, including cases of domestic violence. The pilot involved training 300 police officers and Police Community Support Officers (PCSOs) to use 50 head cameras, which were available from a pool to any trained officer on any shift. The use of the cameras was dependent on the officer choosing to wear it and so the results could be affected by a bias. For violent incidents/crime only the Home Office pilot in Devon and Cornwall found an:  Increase in converting an incident into a crime (71.8% to 81.7%);  Increase in Penalty Notices for Disorder (and administration detections) (2.4% to 3.9%);  Increase in charge/summons (10.2% to 15%); and  Increase in sanction detections (29% to 36.8%). Since these two studies, BWV has been frequently used in pilots for CJ outcomes, but not all of these pilots have been evaluated, or the evaluations have been limited by the design or size of the sample. 12 Whilst all crime was monitored as a secondary outcome, it was decided to focus on instances where BWV would have a greater likelihood of impact – namely violence. 13 Prosecutions where the victim is unwilling to take part. 8 2. Methodology The Randomised Controlled Trial After successfully receiving Home Office funding in 2014 of £1.1m jointly with BTP14 to pilot the use of BWV, the MPS together with the College and MOPAC identified a low cost opportunity to evaluate the impact of BWV technology to generate significant learning for policing, prior to any wider implementation. It was agreed the evaluation of BWV would be run as an RCT, so causal statements could be made about impact. 15. In May 2015 the Mayor, Boris Johnson, and the Commissioner of the Metropolitan Police, Sir Bernard Hogan-Howe, announced the MPS would look to buy around 20,000 cameras for officers across the capitalxv. As the MPS intends to use cameras in other areas of policing and possibly for future full roll out, investing in the development of a trial to evaluate both the use and impacts of BWV offers valuable learning for potential future use of BWV. Body Worn Video Delivery The intervention for treatment officers included 1) training officers to use BWV, 2) allocating personal issue BWV and 3) on-going supervision and guidance on use of BWV. Cameras were allocated solely to emergency response team (ERT) officers at constable rank; as they had sufficient frequency of initial attendance at incidents, complaints, and stop and searches. BWV cameras were not given to police officers in other roles (e.g. neighbourhood officers), PCSOs, or other members of staff who have a wide range of contact experiences, perform different duties, and sometimes have different powers. See Annex A for more detail on the BWV cameras and the support provided by the BWV central team. The trial was managed jointly between the MPS, the College and MOPAC with an on-going feedback loop to assist implementation, provide performance data and ensure the fidelity of the design. A number of challenges which arose during the trial are discussed in the next sections of this report. The Study Design Borough Selection The number of cameras allocated for the trial was fixed (500), as was the maximum length of time the trial could run for (one year), so the choice of location within London was one of the only flexible elements of the design. Boroughs were selected based on the following inclusion and exclusion criteria: Exclusion criteria  Boroughs with a consistently low complaint rate16;  Boroughs that were already using BWV cameras; and  Boroughs whose inclusion would otherwise adversely affect the analysis due to the relatively large team sizes or number of officers on bespoke shift patterns. Inclusion criteria  Boroughs with a relatively high rate of stop and search; and  Boroughs with a relatively high crime rate. 14 MPS actual spend for 2014/15 was £165,041 – of which 50% was match funded by the Home Office 15 Barnet, Brent, Bexley, Bromley, Camden, Croydon, Ealing, Havering, Hillingdon, Lewisham. 16 Set as an exclusion criterion as there is a high degree of variation in the boroughs with higher rates of complaint. 9 All boroughs were ranked accordingly to the above criteria. Some were also excluded based on practical considerations which may have affected their ability to implement the BWV cameras successfully17. Team Selection Teams of officers within boroughs were randomly assigned to the intervention in order to provide a strong basis on which to draw inferences regarding the effects of BWV – a ‘cluster’ RCT design. A ‘cluster’ RCT randomly allocates whole groups of individuals to different conditions being tested. In this instance the cluster or group were ERTs, so rather than individual officers, entire emergency response teams were randomly allocated to receive BWV or not. Once a team was assigned to a condition it applied to all officers in that team – so in a BWV team all officers were allocated to wearing BWV and in a control team no officers were assigned to wear cameras. All five ERTs in each of the ten boroughs were randomly assigned to either:  the treatment group – to receive a body worn camera; or  the control group – to not receive a body worn camera18. Each of the ten boroughs had two treatment teams and three control teams. Contrary to the exclusion criteria on the previous page, after selection ‘Borough 3’ was subsequently found to have cameras and so had them initially removed for some months, then randomly re-allocated. However, due to the incorrect allocation of a team to the treatment condition, it was necessary to exclude two teams from the analysis – leaving ‘Borough 3’ with one treatment and two control teams. Randomisation occurred at the beginning of the trial under the assumption all boroughs would go live at the same time and all officers in the teams would be trained and receive cameras. However, when the trial launched in April 2014, the MPS decided a staggered roll out would ensure sufficient centralised resources to train ERT officers and install technical hardware across the 10 borough sites. In practice, the roll out took over 6 months to complete, which reduced the length of the trial in some boroughs (see Figure 1). The length of the roll-out reduced the chances that less frequent outcome measures such as complaints would occur in large enough numbers to detect a significant change between the treatment and control groups (see Annex A for more details on the trial design). Figure 1: Borough “Go-live” for each month of the trial Boroughs were deemed to go live for evaluation purposes once more than 50% of officers in a treatment team had uploaded an operational clip19. 17 For example, some boroughs were involved in major site relocation and others were participating in other pilots. 18 These procedures were carried out by College researchers to prevent any selection bias. Each team was given a random number and the two teams with the smallest numbers were selected into the treatment conditions. 19 Operational refers to any clip uploaded that is not categorised or flagged as for training purposes. In the event that an officer has never categorised or flagged a clip as training purposes, the 4th clip they uploaded is deemed to be their first operational clip, as the majority of officers who had correctly labelled their uploaded clips upload between 1 and 3 clips during initial training. 10 3. Implementation Learning This trial tests the impact of BWV in its complex ‘real world’ setting rather than how BWV could work in ideal settings. This is important because interventions that lead to significant improvements in ‘ideal’ settings do not necessarily deliver the same results in the ‘real world’xvi. The results of this trial therefore offer a more accurate picture of the potential impact of implementation of BWV in other forces. As previously identifiedxvii documenting learning from implementation is beneficial, so that replication in similar operational settings can better identify and mitigate potential challenges earlier. This trial intended to apply learning from previous studies, which only identified implementation problems at the point of assessing impactxviii. In order to identify any implementation issues an early feedback loop was set up between the evaluation team and the MPS BWV central team. By the evaluation team monitoring staff survey feedback and limited performance data, timely feedback was provided to the MPS BWV central team. Attempts were then made to rectify emerging issues giving the evaluation the best chance of detecting any measureable change. Annex A provides more detail on the trial design and challenges. In particular, while the aim was for all officers in the treatment teams to receive the intervention (BWV), in practice not all ended up being sent on training and issued with a camera as intended. In addition, there were implementation challenges; as the frequent movement of staff in and out of the response teams caused an ongoing training gap; the central MPS BWV implementation team having to address potential resistance from officers; and ensure they had the ability to use the technology and access equipment. A shortfall early on of around 180 cameras and officer churn meant no ERT ever reached 100% of officers fully trained and active in using the equipment. When the in-house training was provided before the roll out in May 2014 officers were positive about the support provided. Although satisfaction regarding support was sustained in subsequent surveys, feedback throughout the trial suggested some frustration, notably around a lack of on-going or ‘refresher’ training and technical support. Three main issues were highlighted with training: a lack of accessible practical and technical support once back on borough (i.e. for effective use of back office functions); that training had focused too heavily on how to use the cameras (as opposed to issues such as when to use the cameras); and there was too long between the training and receiving a camera for operational use. This issue was echoed by the BWV central team who identified delays as a result of challenges installing infrastructure. “I feel that the time between training for the cameras and the borough actually going live was too long. Perhaps a condensed refresher at a briefing would be beneficial for all officers” By far the largest implementation risk was the lack of BWV usage by officers, therefore monitoring compliance was critical to maintain the fidelity of the trial. Over the 12 month period of the trial (May 2014 to April 2015) BWV cameras were used to generate 48,281 recordings, totalling 12,212 hours of 11 video (an average of 15 minutes per recording). Of the recordings submitted, 13,678 (28%), totalling 4,678 hours, were tagged as ‘evidential’ for potential use within the criminal justice system (CJS). This is marginally higher than findings from previous research regarding evidential clips – 22% were found during Operation Hyperionxix. Figure 2 Average number of clips recorded by each officer per one active month During the course of the trial, 649 response officers were given a BWV. These officers have been ranked20 by upload rate (see Figure 2). Officer usage varies considerably, with a small portion officers account for a large portion of clips. The highest average number of clips recorded per month was 58 (230 clips in total) by an officer who was only active for 4 months. Overall, 42% of officers with a camera recorded 10 or more clips per active month, while 26% recorded less than 5 clips per active month, and 6% uploaded nothing during the trial. Figure 3 Percentage of overall clips by category MPS policy makes the use of BWV in a number of defined areas mandatory. Figure 3 displays the percentages of all clips flagged as in one of these mandatory categories21. It is important to note that categories are not mutually exclusive, for example a recording could be of officers attending a domestic violence incident, making an arrest and deploying use of force. Of the 48,281 recordings during the trial, footage recorded from mandatory situations varied considerably. Domestic Violence was the most frequently recorded from all the mandatory situations in all boroughs. It was not possible to understand the proportion of mandatory situations that were recorded, or how representative the number of clips recorded are of an officer’s overall activity (see Annex C for a full usage breakdown). Lack of camera usage may also affect the size of the effect seen in the trial, potentially underestimating the true impact BWV could have if there was total compliance with policy. Making the most of BWV is not just a matter of testing the technology, but influencing the culture in which the technology is usedxx. To explore lack of use, officer surveys and interviews were used to understand practical or 20 The rate of clips recorded by each officer per active month, calculated by dividing the total number of each officer’s recorded videos with the number of months they were in the treatment group. 21 The numbers are likely to be an underestimate due to the reliance on officer labelling. 12 personal barriers officers faced when deciding to activate their cameras. The findings suggest there are four broad areas impacting on the use of cameras: Practical Problems – although officers believed cameras were easy to operate and footage was of sufficient quality, barriers included the bulkiness of the equipment; poor battery life; temperamental connectivity; and the camera positioning on their Met vest, which was particularly awkward when driving. Technical problems – were in the main a lack of docking stations at local custody suites, with uneven distribution causing logistical difficulties and increased travel times. These are problems that will reduce with full roll out. Lack of access to footage on the street, the need to improve sound quality and having headphones to review footage were also identified. Lack of organisational support –shortfalls in timely initial and refresher training may have caused gaps and misunderstandings over how to embed the use of videos into current organisational processes. For example, officers believed cameras increased time spent on paperwork, particularly witness and victim statements. However, there was a misconception that written notes should mirror footage – something which although addressed during training, was not consistently understood by officers. “Having to write word for word what you have said at a scene… [when footage] is already recorded and available evidentially, it seems a duplication and waste of time to then sit and write it all down … makes case progression slower” Resistance – Perhaps unsurprisingly, the majority of officers believe they should have more discretion over when to activate BWV. Research has previously identified lack of discretion as a barrier for officers, who see mandatory recording as a lack of organisational trustxxi. Officer interviews in this trial highlight concerns including the perceived increased scrutiny from management and outside agencies (notably courts), with worry that footage would not accurately reflect dynamics relayed at the time. An officer’s ability to exercise discretion during incidents was also frequently mentioned, with officers feeling overly self-conscious and obliged to apply the letter of the law in order to protect themselves from any future reviews. Concerns that BWV affected interactions with the public and colleagues were emphasised in officer surveys, with some suggesting that they affect team morale by increasing suspicion that footage could somehow be used against them. “…less comradeship at calls as everyone is wary that cameras are on and recording, thus a drop in morale.” 13 4. Results Headline Findings: Summary The cluster randomised controlled trial (RCT), issuing officers with Body Worn Video (BWV) showed:  BWV can reduce the number of allegations22 against officers, particularly of oppressive behaviour23. Complaints related to interactions with the public24 also reduced and, although it did not reach statistical significance, the trend in overall complaints was consistent with these findings.  There was no overall impact of BWV on the number or type of stop and searches conducted. In addition, there were no differences in officers’ self-reported behaviour relating to how they conducted stops.  Arrests as a result of a search are slightly less likely when officers are in a BWV team  No effect was found on the proportion of arrests for violent crime. When an arrest had occurred, there was a slightly lower proportion of charges for reporting officers in a BWV team. It is not known if BWV leads to the preparation of fewer, but stronger cases without further exploration of later CJ outcomes at court. Subsequent analysis isolating the effects in cases where BWV officers specifically flagged a case with footage, showed a higher arrest rate than cases reported by officers not in a BWV team, but no difference to the charge rate. This finding could indicate officers were using cameras during arrest to strengthen existing evidence, or officers chose particular types of situation to flag BWV evidence This section summarises the findings from the trial. It starts by examining the effect of the intervention on the principal outcome measures, and then discusses the officer survey and interviews, as well as the public attitude information to contextualise the findings. For all primary outcome measures, multivariate statistical models were created to compensate for the effect of the team clustering and to assess to what extent other explanatory factors had an impact on outcome measures when considered alongside the effect of BWV. Explanatory factors in the models included: the demographic characteristics of the officers; factors associated with level of camera use (e.g. time in the trial); and contextual factors (e.g. borough). Use of multivariate models is intended to eliminate the possibility that any potential effects seen in our initial analysis was actually due to another factor aside from BWV. In addition, exploratory analysis was conducted using chi-square tests to further interrogate data for the primary outcomes and compliment the main results drawn from the multivariate statistical models (e.g. analysis by type of complaint or proportion of searches where an arrest was made)25. The data for all the analysis came from a merging of data sets from MPS that held incident, complaint, stop and search and officer details and BWV usage information. To support the data analysis semi- structured interviews and surveys of officers were undertaken to help understand their experience of using the BWV cameras. The officer survey was designed to enable comparison of treatment and 22 Each complaint can comprise of a number of different allegations (for example, an officer could be alleged of unlawful arrest and discriminatory behaviour during the same incident – this would equate to one complaint, with two allegations). 23 MPS categorise the following types of allegations as Oppressive Behaviour: Oppressive conduct or harassment, Other assault, Other sexual conduct, Serious non-sexual assault, Sexual assault, Unlawful/unnecessary arrest or detention. 24 Breach Code A PACE/ Breach Code B PACE/ Discriminatory Behaviour/ Incivility, impoliteness and intolerance/ Lack of fairness and impartiality/ Oppressive conduct or harassment/ Other assault/ Other sexual conduct/ Serious non-sexual assault/ Sexual assault/ Unlawful or unnecessary arrest or detention. 25 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. 14 control officers’ attitudes and self-reported behaviour; respondents were asked the extent to which they agreed or disagreed with a series of statements about their interactions with the public, the organisation and their decision making. No reference to BWV was made in the statements allowing officers in the trial to respond to the same statements, with differences in response directly attributable to BWV. Both surveys and interviews were used to help identify the possible mechanisms of any change (or lack thereof) identified through the data analysis. As part of the trial a large amount of data was collected and a number of other measures monitored, including public attitudes and victim user satisfaction, which are discussed in the results section. A list of all secondary data sources and analysis can be found in Annex B. 15 Impact on complaints The trial tested whether BWV cameras would reduce the number of public complaints made against response officers. During the course of the trial 261 complaints were recorded on the MPS internal database, comprising 462 individual allegations26. This equates to one complaint per officer every 4 years. The difference between the officer complaint rates in the control and treatment teams in each borough is shown in Figure 4. Blue diamonds represent the individual borough average and the black diamond is the basic overall effect. When the arms of the diamonds cross the central axis (at 0) it indicates the difference between the control and treatment groups is not statistically significant27. Six of the ten boroughs had on average a lower rate of complaints in BWV teams than their control teams and this difference was statistically significant in two boroughs. Despite a lower overall rate of complaints in the BWV teams, basic analysis28 shows there was not a statistically significant combined effect from all ten boroughs. It should be noted that the small number of complaints recorded during the trial made it unlikely that a statistically significant difference between the two groups would be found. Figure 4: Difference between the average number of complaints per officer in the control and treatment teams in each borough Multivariate statistical modelling was performed to rule out other explanatory factors29, and to compensate for the effect of the clustering in the data. Modelling confirmed the difference between treatment and control groups was not significant. The difference only met the 10% significance level30 rather than the more usual 5% level, giving a higher chance that observed differences are a result of sampling error, but providing some evidence that BWV may reduce complaints. When the complaints 26 Each complaint can comprise of a number of different allegations (for example, an officer could be alleged of unlawful arrest and discriminatory behaviour during the same incident – this would equate to one complaint, with two allegations). 27 The black arms represent a 95% confidence interval. 28 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. Results from more comprehensive statistical modelling are shown below, and these provide the strongest analysis of the data. 29 Gender, number of months in the trial, officer age, number of stops, team and borough. 30 Hierarchical generalised linear model with a Poisson link function and teams within boroughs (BWV, β=-0.29, exp(β)=0.75, p=0.10). 16 are narrowed down to categories linked to police-public interaction31 where BWV is most likely to have had an impact, there is a statistically significant difference, with fewer complaints in the BWV teams compared to the control groups32. Although there was no effect when looking at overall complaints, this finding shows BWV may reduce particular types of complaints. This finding is supported when conducting a range of other modelling, undertaken to further explore the results. Data relating to the number and type of allegations that make up complaints were explored and it was found that there was a statistically significant difference between the two groups in the average number of allegations per officer33 (0.17 in the BWV group, and 0.26 in the control group), a 33% reduction. When allegations were broken down by type it was found that the greatest difference between BWV and non BWV teams appeared to be those relating to oppressive behaviour34 and incivility35. When statistically tested there was a significant association between the absence of BWV and officers being alleged of oppressive behaviour36. The odds of an officer in a non BWV team receiving an allegation of oppressive behaviour was 2.55 times higher than for an officer in a BWV team (2% of BWV officers vs. 5% of control officers). However the difference for allegations of incivility was not significant at the 5% level37. To understand if BWV had a particular impact on officers with a higher than average number of identified behavioural issues, data drawn from the MPS’s ‘Complaints Intervention Scheme’ (CIS)38 was analysed (Chi square). This showed there was no difference in the proportion of officers enrolled, newly joined, or removed from the CIS or had a ‘behaviour issue39’ between those wearing BWV or not. Understanding the results Survey analysis indicated treatment group officers were significantly40 more likely to feel protected against complaints41 than the control group, as their survey responses indicated they felt allegations with little evidence were not being upheld or were resolved quickly. However, exploring the complaints data there was no difference found in the time resolved and too few complaints were found to be substantiated during the trial for analysis to detect any difference between groups. The survey findings that officers feel protected against complaints may not be reflected as strongly in the results due to 31 Breach Code A PACE/ Breach Code B PACE/ Discriminatory Behaviour/ Incivility, impoliteness and intolerance/ Lack of fairness and impartiality/ Oppressive conduct or harassment/ Other assault/ Other sexual conduct/ Serious non-sexual assault/ Sexual assault/ Unlawful or unnecessary arrest or detention. 32 Hierarchical generalised linear model with a Poisson link function and teams within boroughs (BWV, β=-0.54, exp(β)=0.58, p=0.04). 33 Hierarchical generalised linear model with a Poisson link function and teams within boroughs (BWV, β = -0.51, p=0.03) 34 MPS categorise the following types of allegations as Oppressive Behaviour: Oppressive conduct or harassment, Other assault, Other sexual conduct, Serious non-sexual assault, Sexual assault, Unlawful/unnecessary arrest or detention. 35 MPS categories one allegation type as ‘incivility’ - ‘Incivility, impoliteness and intolerance.’ 36 X2 (1) = 12.97, p <0.001 37 X2 (1) = 3.067, p = 0.08 38 An officer is enrolled onto the CIS when they have had 3 or more behavioural issue cases (either a public complaint, conduct matter or civil action) in a 12 month period. Each case lapses after 12 months which results in the officer being removed from the scheme. 39 public complaint, conduct matter or civil action. 40 Second officer survey: U = 21678.5 , P < 0.001, Treatment Mdn = 4.00, Control Mdn = 3.25, r = 0.26 and third officer survey: U = 30875.5, p < 0.001 r = 0.21, Treatment Mdn = 4.50; Control Mdn = 3.75. Mann-Whitney Test. 41 As measured by responses to the statements, “If a member of the public made a false complaint against me, it would get sorted out quickly”, “I’m confident people would believe my side of the story if a complaint was made against me.”, “A complaint made against me with little evidence would be dropped quickly.” And “I feel protected from malicious complaints when dealing with the public.” Other Complaints analysis BWV and non-BWV wearing ERT teams were also compared over a variety of measures of complaints. No significant difference was found across: the proportion of officers in the teams who received a complaint or not during the trial; the proportion of allegations resolved within the trial period; and the time taken to resolve. From the 462 allegations, only three were found to be substantiated during the entire trial, it was not possible to explore any difference between the control and treatment groups. 17 the small number received and may also be explained through possible impacts of BWV earlier in the complaints process. Officers anecdotally reported that unfounded complaints were being quickly resolved locally using BWV footage, but there was no data available at the initial contact stage to be able to confirm these claims. BWV may lead to a reduction in recorded complaints due to early resolution of issues with little grounds. Officers were positive about the impact of BWV on complaints, with survey findings supported by the interviews. When asked to show interviewers footage that sums up their experience of using BWV four officers of the nineteen brought in video clips demonstrating protection from potential complaints, a seemingly high number given the relative rarity of complaints. During interviews some officers said a potential complainant would be aware that the footage would counteract false or exaggerated claims. “A lot of the time we have a lot of completely made up complaints, not exaggerated truths, completely ludicrous, made up, and I think at the point of that complainant being told that there is body worn footage of their entire dealing, I think a lot of complaints will be retracted or solved really quickly.” “I don’t know whether it’ll stop people complaining. I’d like to think it did but even if the level of complaints stays the same, it makes it easier to deal with if you’ve got the camera footage there that can make the investigation a whole lot shorter than it currently is.” There was no difference in survey findings of self-reported behaviour and perceptions of the public’s behaviour between officers with BWV and those without, which suggests that officers did not feel a difference from BWV or the difference was too small to alter survey responses about their interactions with the public. However, in interview officers reported mixed experiences about how the public communicate when BWV is present, with interactions appearing to be dependent upon the individual and the situation. Whilst a few officers during interview indicated BWV did not change their behaviour, others believed it impacted on their interactions with the public. No difference was found for officer survey responses in relation to: procedurally just policing42; use of force; compliance with protocols; or an officer’s belief that behaviour below expectations would be challenged. “I suppose it reinforces that you’ve got to give all your grounds and your interactions nicer, maybe, at the right times. It makes you do it more textbook like, I suppose.” “I think it just makes you think for that split second longer about what you’re going to do and what you’re going to say first…” 42 Procedural justice relates to the public’s perception of fair decision making and respectful treatment by the police. Case Study 1: Protection from malicious complaints Officers conducted a vehicle stop and were confronted with a hostile and violent suspect who exited the vehicle and threatened officers. As a result of his conduct officers deployed their Taser. Whilst being booked into custody the suspect stated that both officers had looked at each other and laughed before saying "let’s Taser him". Officer BWV footage was shown to professional standards who stated that officers had no case to answer. 18 Findings Summary: Complaints BWV can reduce the number of allegations against officers, particularly of oppressive behaviour. Complaints related to interactions with the public also reduced and, although it did not reach statistical significance, the trend in overall complaints was consistent with these findings. Whilst officer surveys did not indicate a change in the quality and nature of interactions between officers and the public, during interview officers report particular instances where BWV has changed behaviour. Officers also gave examples of using BWV recording to achieve early resolution of potential complaints with little grounds, preventing them from becoming formally progressed, findings which are supported through the officer survey, as BWV officers felt greater protection from complaints43. Impact on stop and search It was hypothesised that BWV cameras would reduce response officers’ use of stop and search. The MPS stops database recorded approximately 11,300 stops as being carried out in the ten boroughs during the course of the trial. This is an overall average of 5.54 stops per officer (5.62 BWV group, and 5.50 in the control). There was some variation in the volume of stop and search across the boroughs, perhaps reflecting the local mix of crime, hotspots, or local policies. The lowest borough-wide stop rate for ERTs was 2.2 per officer, while the highest was 13.4. Figure 5 shows the difference between the average number of stop and searches for officers in the control and BWV teams across boroughs44. With three boroughs, their BWV teams had a lower rate of stops than their control teams with one of these being statistically significant. Combining results from all ten boroughs indicates there is no significant difference. Figure 5: Difference between the average number of stops per officer in the control and treatment teams in each borough It might be that BWV affects boroughs in different ways depending on their stop and search activity, as the only borough with a statistically significant reduction had over twice the rate of stops prior to 43 See Annex B for statements combined to create this factor 44 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. 19 the trial than the next highest trial borough45. When completing multivariate statistical modelling to rule out other explanatory factors46, which includes the influence of individual borough, there was no evidence that BWV resulted in a reduction in the overall number of stop and searches carried out by each officer47. Understanding the results The finding that BWV does not decrease the rate of stop and search is supported through the officer surveys, with no difference in officers’ self-reported behaviour in relation to discretion or compliance with procedures. In addition there was no reported difference in behaviours linked to procedural justice, which included variables directly related to stop and search procedures (see Annex B). However, arrests as a result of a search are slightly less likely when officers are in a BWV team, indicating a potential difference in decision making48. Officer interviews revealed mixed views, with some reporting no change to their approach and other officers giving examples of change. “… you’re dealing with the situation that’s in front of you … I don’t think it changes the decision, it changes the execution of how you do it… …so I’m still going to stop… but how I go about the stop is going to change.” “I think you do take a second to think ‘well why am I potentially gonna stop this person?’ and you think in your head ‘right, I’m gonna search them for this’…... and sometimes you think ‘oh actually maybe there’s probably not enough [grounds]…” Officers with BWV were less likely than those without BWV to agree they needed stronger justification for their actions49. This finding may explain, to some extent, why BWV has not resulted in a reduction of searches or a change in the quality or nature of the search; BWV, rather than affecting officer decisions to make a stop, may enable officers to feel more confident if challenged. Analysis of interviews suggests officers feel BWV footage can provide support for their justification to search which they did not have previously. “…it also backs my justification for searching them so it’s got evidence again for me doing what I’m doing so I know that again if he complains or that person complains people can understand you know, why our actions are justified.” The need for stronger justification links to the feeling of protection from complaints discussed in the previous section – as officers feel they can support the action they have taken. However using BWV as supporting evidence led more than half (11 of 19) of the officers during interview to be concerned BWV would result in an erosion of the value of officer word, feeling that written statements alone would no longer suffice as acceptable evidence. 45 Bromley: 56 searches per 100 officers; Barnet: 22 searches per 100 officers. 46 Gender, ethnicity, number of months in the trial, officer age, team and borough. 47 Hierarchical generalised linear model with a Poisson link function and teams nested in boroughs. β=0.12, p=0.21. 48 Hierarchical generalised linear model with a binomial link function and teams nested in boroughs. B= -0.17, p=0.048 49 See Annex B for statements combined to create this factor Other Stop and Search analysis BWV and non-BWV wearing ERT teams were also compared over a variety of measures of stop and search: the proportion of stops of people recorded as being from a BME group; and the type of stop conducted. Although there were some small variations between the treatment and control groups these were not statistically significant when the effects of the clustering were taken into consideration. 20 Findings Summary: Stop and Search There was no overall impact of BWV on the number or type of stop and searches conducted. In addition, there were no differences in officers’ self-reported behaviour relating to how they conducted stops, but arrests as a result of a search are slightly less likely when officers are in a BWV team. Officer surveys showed no difference in discretion; compliance with procedures; and self-reported behaviour in relation to procedural justice during a search. However, officers with BWV were less likely than those without BWV to agree they needed stronger justification for their actions. This may explain to some extent why BWV has not resulted in a reduction of stop and searches, or changed their quality or nature. Rather than affecting officer decision making, BWV may enable officers to feel more confident if challenged. Analysis of interviews suggests officers feel BWV footage can provide support for their justification to search which they did not have previously. Impact on criminal justice outcomes It was hypothesised that BWV cameras would increase the proportion of violent incidents attended by response officers that resulted in a criminal justice outcome. Use of BWV at incidents of violence potentially offers evidence of injuries, the scene, and heightened emotional responses of suspects, victims or witnesses which could otherwise affect the assessment of evidential thresholds. Approximately 148,000 crime reports were extracted from the MPS CRIS (Crime Reporting Incident System) between 1st May 2014 and 30th April 201550. From this, a total of 64,355 (44%) were identified as viable for analysis (see Annex D). Data was analysed as cases progressed through different stages of the CJ process, with the main milestones being: 1) Where at least one suspect was arrested; and 2) Where at least one suspect was charged/received a summons. From the 64,355 case included for analysis:  Just under 50% (n=31,682) had at least one suspect;  Of these 68% (n=21,604) had recorded at least one arrest;  Of the cases with at least one arrest, 59% (n=12,717) featured a judicial outcome51; and  The majority of these judicial outcomes were a charge/summons (n=8,849, 70%). The officer52 who reported the case determined whether it was allocated to treatment or control conditions53. Arrests There was some variation in the difference between the treatment and control group rates of arrests across the boroughs54. Figure 6 shows the variation across boroughs55 with three boroughs displaying 50 Data downloaded on 08/06/2015 – any CJ outcomes up until this date included in analysis. 51 Judicial outcomes include: Charge/Summons, Cautioned, Warning or other outcome (e.g. Community Resolution, Cannabis Warning, Penalty Notice) from the police 52 To note, this officer may not have made an arrest, but will most likely be the first contact point with the case and therefore most likely to use BWV for evidential purposes. 53 To increase the likelihood of detecting impact, it was decided to focus on violent incidents where it was felt BWV was more likely to capture more and better quality evidence. For our analysis, violent incidents were defined as: Serious wounding; Assault with Injury; Common Assault; Attempted murder; Assault on a constable; Affray; and Violent disorder. 54 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. Results from more comprehensive statistical modelling are shown below, and these provide the strongest analysis of the data. 55 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. 21 their BWV teams had a higher rate of average number of arrests than their control teams and only one of these being statistically significant. Combining results from all ten boroughs indicates there is no significant difference56. A more robust multivariate statistical modelling was performed to rule out other explanatory factors57, and to compensate for the effect of the clustering in the data. This modelling confirmed there was no evidence that BWV increased the rate of arrest for violent incidents58.. Figure 6: Difference in the proportion of crimes with an arrest at any time, between the treatment and control teams in each borough Charge/Summons following arrest There was variation in the difference between the treatment and control group rates of charge across the boroughs59. Figure 7 shows the variation across boroughs60 with five boroughs displaying their BWV teams had a lower rate of average number of charges than their control teams and two of these being statistically significant. Combining results from all ten boroughs indicates there is a small difference61, with the cases where the reporting officer is in a team BWV less likely to result in at least one charge. This result was confirmed by a more robust multivariate statistical modelling62, which was performed to rule out other explanatory factors63. However, with very large sample sizes, small differences can be statistically significant and so the real impact should be treated with caution64. Without further exploration of later CJ outcomes at court it is also not known if a lower proportion of charges is a result of BWV leading to preparation of fewer, but stronger cases. 56 𝑥2 (1, N = 13,635) = 2.855, p = .091. Phi = 0.014 57 Borough, crime types, DA incidents, VR flags, number of suspects, number of victims, previous convictions, victim-suspect relationships. 58 Hierarchical generalised linear model with a binary link function and teams nested in boroughs. β=-0.08, p=0.25, (n=13,655) 59 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. Results from more comprehensive statistical modelling are shown below, and these provide the strongest analysis of the data. 60 Descriptive statistics, and basic statistical tests, do not compensate for any “inherent” complaints rates in the boroughs (‘clustering effects’), the time an officer spent in the BWV or control groups, or other demographic factors. Nevertheless, they are useful for understanding and describing the data. 61 x2 (1, N = 8222) = 4.229, p = .040. Phi = -0.023 62 Hierarchical generalised linear model with a binomial link function, and teams nested in boroughs. β=-0.097, p=0.048, (n= 8,222) 63 Borough, team, crime types, DA incidents, VR flags, number of suspects, number of victims, previous convictions, victim-suspect relationships. 64 Whilst the Borough 4 received additional resources from the MPS as it had a digital court; this appeared to have no effect on the charging outcomes, with no real difference seen between the control and treatment groups, seen in the basic analysis. 22 Figure 7: Difference in the proportion of crimes with charge/summons, between the treatment and control teams in each borough A range of other analysis was undertaken to further understand these results. To begin, the use of BWV in cases was explored, using the ‘VR flag’ (signifying when an officer has flagged video evidence)65. It was found that 28% of violent crimes in the treatment team had a VR flag compared to 0.2% in the control team, showing an increase in amount of available digital evidence for cases from BWV. In the original multivariate statistical modelling for arrest, which took into account other strong explanatory factors (such as number of suspects and victims, or the type of crime), the variable ‘VR flag’ had a significant positive relationship with arrest66. In other words, crimes where officers had flagged BWV evidence have an increased likelihood of an arrest. Subsequent analysis isolated the effects of footage. All cases reported by non-BWV officers were compared to only cases where video footage was specifically flagged as included by BWV officers. This was used to explore the effect of the presence of the BWV evidence itself on the decision to arrest and charge. This analysis supported the model, indicating a significant relationship with arrest67. This finding could indicate officers are using cameras during arrest to strengthen existing evidence, or be a result of a different type of situation where video recording occurs, in which BWV is more likely to be used to collect evidence. It may be that more serious incidents or more motivated officers explain the increased arrest rate. However this positive relationship between BWV evidence and CJ outcomes was not seen for charges or summons – suggesting any bias in recorded incidents is not a factor. As other possibilities not controlled for in our analysis cannot be ruled out, this needs further investigation to ensure the difference is a result of BWV. 65 In the majority of cases this will relate to BWV, but could signify the use of CCTV. 66 b = .589 p = <.000 (N=13,655) 67 x2 (1, N = 9832) = 129.481, p = <.000. Phi = 0.115. Borough_1 Borough_2 Borough_3 Borough_4 Borough_5 Borough_6 Borough_7 Borough_8 Borough_9 Borough_10 Overall -30.0% -20.0% -10.0% 0.0% 10.0% 20.0% 30.0% Charging Rate (difference between BWV and control teams) Greater proportion of Charge/Summons in the BWV teams are to the right . Other Criminal Justice Milestones BWV and non-BWV wearing ERT teams were also compared over a variety of secondary CJ milestones: whether or not at least one suspect was identified; whether or not at least one arrest at the scene occurred and whether there was any Judicial Outcome. No significant difference was found across all the above additional outcomes. 23 There is evidence from officer interviews and surveys, as well as management information from the BWV central team to suggest the lack of impact in the level of charges/summons may be due to barriers earlier in the CJ process, both internally with secondary investigators and with external partners such as the CPS not routinely using BWV evidence. This blockage in the flow of evidence has been frequently reported in officer surveys and perhaps exacerbated by low officer usage, meaning secondary investigators and the CPS reportedly never became proficient in the use of footage. “CPS do not know how to view the footage… they need to be shown how to use [the cloud based evidence repository]” Monthly compliance and audit checks by the BWV central team support this lack of use throughout the CJ process; however the data collection was not systematic, so can only be used as indicative evidence as it may not be representative of all cases. Understanding the results Although there is no increase in arrest or charge as a result of BWV, officer surveys indicate that as teams became more acquainted with the technology, treatment officers were significantly more likely than control officers to report being confident they were capturing good quality evidence68. On the whole, interviewed officers appreciated the evidential value of using BWV at incidents, specifically in relation to Domestic Abuse (DA). When asked to show interviewers footage that sums up their experience of using BWV, over half brought in video clips demonstrating their ability to capture key pieces of evidence - many from DA incidents. Specifically, the ability to accurately present evidence rather than depending solely on memory was felt to be a key benefit, as well as revealing a level of detail and emotion not possible to capture in written statements. “…you can be as eloquent as you can with your statement it still, it’s still very powerful to see that person stood there tears streaming down their eyes, you know covered in blood or whatever T-shirt torn, you know children screaming, you can put that in a statement but it’s still very powerful to see it, you know in the live film.” Even as a key mandatory recording situation DA flagged incidents (n=10,006) did not see any significant difference between the control and treatment groups in arrest or charge. This may be a reflection of the existing drive for positive action at DA incidents by the MPS. It may be that impacts will be felt further down the CJ process, through outcomes in court, no data was available at this time for analysis. Future papers plan to explore impact on early guilty pleas and sentencing decisions as a result of BWV for all crime types. 68 U = 32932.0, P < 0.00, r = 0.17: Treatment: Mdn = 5.50, Control: Mdn = 5.00) to feel they collect quality evidence at the scene. In the third survey treatment officers also disagreed significantly more than control officers with statement “The evidence I gather doesn’t convey what it’s really like at the scene” Treatment Mdn =XX, Control Mdn XX, U = 35131, P = 0.002 r = -0.13. These findings were not observed in surveys earlier in the trial. Case Study 2: Evidence led prosecutions Officers responded to a domestic incident where a suspect had punched his partner in the face and struck her several times with a frying pan in front of her children. Initially the victim provided a statement, however this was withdrawn. As a result of evidence captured on BWV, the suspect was arrested and interviewed and the case was put before CPS who authorised a charge. 24 Findings Summary: Criminal Justice Outcomes BWV did not affect the proportion of arrests for violent crime. When an arrest had occurred, there was a slightly lower proportion of charges for incidents reported by officers in a BWV team, is not known if BWV leads to the preparation of fewer, but stronger cases without further exploration of later CJ outcomes at court. Subsequent analysis isolating the effects in cases where BWV officers specifically flagged a case with footage showed a higher arrest rate than cases reported by officers not in a BWV team, but no difference to the charge rate. This could indicate officers are using cameras during arrest to strengthen existing evidence, or officers chose particular types of situation to flag BWV evidence. Although the statistical analysis suggests BWV has had a limited impact, officer surveys indicate a belief that BWV helps to collect better quality of evidence. However, there appears to be a blockage preventing BWV evidence being used in later stages of the CJ process. Exploring only incidents of Domestic Abuse (DA) (a mandatory recording situation), analysis found BWV does not affect the decision to arrest or charge. This may be a reflection of the existing drive for positive action at DA incidents by the MPS. Nevertheless officer interviews indicated BWV can provide evidential value particularly during incidents of DA, to show a level of detail/emotion not possible in written statements, an impact which may be seen later in the CJ process (for example at court), but has not been possible to capture during the current trial. Secondary Outcome Analysis The research explored a number of other areas where BWV may have an impact, specifically: police/public interaction, innovative use, and implementation learning. Public Attitude The Public Attitude Survey (PAS)xxii found London residents are generally supportive of BWV, agreeing positively with the below statements, perceiving BWV:  will make officers more accountable for their actions (92%, n=11,756);  reassure them the police will do the right thing (87%, n=11,035);  will make officers treat people fairly (87%, n=11,127);  will ensure officers act within the law (90%, n=11460); and  will ensure officers follow correct procedures (90%, n=11496). These statements link to the concept of procedural justice69 and whilst the direction of influence cannot be implied, across boroughs where the public had a higher level of awareness of BWV they were also more likely to agree with the statements. In addition, agreement with these statements was significantly linked to confidence in the police. These findings demonstrate potential benefits of the public awareness of BWV. MPS external communication, consultation and transparency should therefore be at the forefront of the wider roll out of BWV to raise awareness. During the trial, the majority of PAS respondents (51%, n=6,483) were not actually aware MPS officers wore BWVs. Of those who were aware (6,435) it appears the majority (69%, 4,416) resided in the non-trial boroughs. This may be a reflection of the historical pockets of BWV use; that residents encountered BWV outside of their home borough; or something about the boroughs selected to take part in the trial being different to the other boroughs (higher stop and search and compliant rates), or may not even be borough related, as most of the respondents awareness of BWV came from the media rather than direct contact. Of the respondents who were aware of BWV, the majority (84%, n=5,437) had heard about it from the media. Only 13% (n=806) had seen an officer wearing a camera and 5% (n=255) had contact with an officer wearing BWV. There is no difference 69 Procedural justice relates to the public’s perception of fair decision making and respectful treatment by the police. 25 between levels of confidence (police are doing a good job in this area) whether respondents were aware of BWV or not. As reflected in the officer survey, PAS respondents saw the potential evidential benefits of BWV, with 95% (n=12,193) believing they will help the police to gather evidence. They did not think BWV would negatively impact on police-public relations; disagreeing that BWV would invade people's privacy (62%, n=8,155); or make the police less approachable (61%, n=7,883) and agreeing they would stop people making false allegations against officers (89%, n=11,345). However, as only 5% of PAS respondents had contact with an officer wearing BWV, the general acceptance of BWV may therefore not reflect those members of the public who have had direct contact with the police. A number of officers reported in the final survey being asked to turn their camera off at least once by: individuals being stopped and searched (20%, n=66); suspects (34%, n=120); victims (36%, n=126); the general public (17%, n=61); non-BWV fellow officers (12%, n=62); and by senior supervisors (5%, n=19), highlighting contact may affect acceptance70. Dealing with victims It was hypothesised BWV may increase the chance of a case having a CJ outcome, in part because of the increase in cooperation of victims and witnesses to proceed with a case and, in parallel, the ability to proceed without a victims involvement. From all 64,355 cases included in analysis 60,368 had at least one victim. Analysis compared cases from the treatment and control teams on the suspect elimination reasons of ‘Victim Unwilling to Prosecute’, ‘Insufficient Evidence to Proceed’, as proxies for victim cooperation and the ability to proceed with a case, found no significant difference was between the groups. Therefore, there was no evidence of an impact on victim involvement from BWV at this stage, perhaps related to the block in the flow of evidence in the CJ process discussed in the section above. The MPS trial attempted to further explore BWV use, assessing the impact on victim satisfaction with the police. Although no difference has been found when analysing the User Satisfaction Survey (USS) using multivariate logistic modelling, it may be there has not been enough time for BWV to embed to see stronger benefits in victim experience, this is therefore another area which should be assessed on full roll out. There were indications from exploratory analysis that there could have been differences in satisfaction which, although they not sustained after more robust modelling, may be worth further exploration: the presence of BWV and one of the drivers of satisfaction (police action at the scene)71; satisfaction with the initial police action72; and overall experience73 for victims of Racially Motivated Crime. Officer Safety To understand more about officer safety and to further explore findings from officer surveys regarding self-reported and public behaviour, data from officer injury and assaults were analysed. 70 Each question has varying responses rates. 71 x2 (1, N =4,002) = 6.832, p = -.042 72 f = .151, p < .05.Phi-Coefficient; x2 (1, N =371) = 8.493, p = -.004 73 f = .118, p < .05.Phi-Coefficient; x2 (1, N =371) = 5.207, p = -.022 Case Study 3: Capturing victim impact Officers made use of BWV when responding to a domestic incident where a suspect’s parents had refused to give him money to buy heroin. The suspect smashed a T.V., a laptop and pushed family members before threatening them with a screwdriver. Officer’s BWV captured first hand the threatening demeanour of the suspect and emotional anguish the parents were suffering. The suspect was later charged and was remanded in custody. Case Study 4: Officer Protection An officer reported: “We stopped a moped in a back alleyway by a local drug dealer probably within seconds surrounded by about 15, 20 youths all quite hostile, just two of us….. within about seconds they all realised they were on camera and it was like the effect of having an extra 10 PCs here. Everybody realised I’m on camera, I better watch what I say here, I could get arrested either now or at a later date and its effect was really quite impressive.” 26 Incidents in which an officer was a victim of a violent assault were analysed (n=499), of particular interest as this crime type is the most likely to been recorded on BWV. There was no significant difference between officers wearing BWV or not in their levels of assault. This mirrors the results for arrests for all violent crime where no difference was seen. Conflict was further explored by analysing officer injuries data. Chi squared tests shown the proportion of officers that had incurred an injury whilst arresting or restraining a member of the public were no different between treatment and control officers. Taken together, the assault and injuries data suggests an officer is no less likely to be assaulted if wearing a BWV. As mentioned previously, officer surveys found no difference in the self-reported frequency of use of force between the two groups. The lack of impact on use of force was further evidenced in interviews, where officers emphasised their decision to use force tended to be made as an immediate responses to a situation, so BWV had no impact upon this. “No not using force ‘cause when you need to use force, you just have to do it. I’m not really one to use force without needing to do it, so the camera wouldn’t change that.” “that’s pretty much instantaneous. Once you think you need to use force, you use force. No [BWV] doesn’t affect me in any way.” These results combined appear to suggest BWV had no impact upon the likelihood of physical conflict with suspects during this trial. This mirrors previous stated findings that BWV does not change decision making or the quality of an interaction, but does provide officers with more confidence (Case study 4) and justification for their actions. Innovative Use Throughout the trial officers have highlighted how BWV has been used outside of mandated situations to secure additional benefits. These have included, but not been limited to: capturing the behaviour of individuals in mental health crisis to subsequently inform NHS partners (i.e. to demonstrate need and secure a place of safety); innovative evidence gathering at crime scenes and premise searches; providing quick time evidence to partners (e.g. the coroner); and using footage as an intelligence source for other policing purposes (Case study 5). Furthermore, officers describe using footage for their own professional development - both individually and as part of a team - highlighting the potential to use BWV as an aid for learning, by reviewing footage to provide practical examples of good practice (Case study 6). The MPS have recognised the opportunities that the recording of encounters, particularly in respect of stop and search, can bring in order to demonstrate accountability of routine policing actions to the public. In 2014 the MPS Central BWV team pioneered the first independent scrutiny of stop and search footage by allowing vetted members of Borough community monitoring groups to randomly select stop and search footage for review; providing immediate feedback on encounters to Borough Commanders. Case Study 5: Police Intelligence A BWV equipped team responded to a victim of a stabbing. Officers uploaded the footage, accurately recording the location field. It later transpired the victim and his friends were of interest to Trident, all linked to serious gang related activity including drug dealing. Due to BWV footage with accurate location details, Trident officers obtained gang association intelligence and were able to use high quality imagery from the footage to assist their on-going operation. Case Study 6: Learning Lessons Officers reported watching footage, during a team briefing session, from a Taser incident where lessons about improved practice could be learned. The incident had been captured on BWV and footage shared with the team, to talk through how they could have handled the situation differently. 27 5. Conclusions and Implications This trial was an opportunity for a large urban police service - the MPS - to test the impact of implementing BWV using a robust methodology. As the results are not entirely consistent with the size and scale of previous findings (see Setting the Scene) the dangers of making policy decisions, practice or recommendations from weak or non-existent evidence are highlighted. For example, a non- experimental design would have suggested BWV had a positive impact in reducing the prevalence of stop and search, as the trial boroughs saw a 49% reduction compared to the previous year. Use of comparison groups and the RCT design was able to demonstrate that this reduction occurred in the absence of BWV also, ensuring positive impacts have not incorrectly been attributed to BWV. However, to ensure this methodology was robust and implemented correctly took a proactive collaborative approach by MOPAC, the College of Policing and the MPS – as this was a complex multi-site cluster randomised controlled trial, with many lessons learned throughout. Findings from the trial challenge previous assumptions about the impact of BWV, especially on police- public interaction and increased accountability for stop and search. The trial has demonstrated no change in the number or nature of searches and there is little supporting evidence about the quality of the interaction improving. This suggests a gap between the high expectations around BWV and the actual impact it has in relation to stop and search and officer behaviour change. However, BWV can still provide transparent evidence about whether stops were warranted and whether the outcome of such stops (e.g. the discovery of lethal weapons for instance) was achieved. In addition, officers themselves welcomed the transparency BWV provides with respect to their decision making. This trial found evidence for a reduction in complaints and that BWV does reduce the number of allegations an officer receives, particularly for oppressive behaviour. Whilst there are strong case studies where BWV has deescalated situations, the more modest impact on complaints observed in our study suggests that positive interactions may already be occurring, making the number of confrontational situations which BWV might influence, relatively rare and impact difficult to detect. A common challenge to BWV as tool for behaviour change is that officers can chose to switch the camera off when they would not like their behaviour recorded. However, there is no evidence of this occurring and the team based allocation would mean these officers would require other BWV officers present to also switch their cameras off. In addition to improving the quality and nature of police-public interactions, there have been suggestions that BWV would improve the evidential process, increasing criminal justice outcomes. During the length of the trial it has only been possible to test the impact of BWV on arrest and charge rate. The trial found BWV had no impact on these measures, but future research is planned to explore the outcomes in court for convictions and sentencing. The research will also help to understand the facilitators and barriers throughout policing and the wider criminal justice system to the effective use of BWV footage for evidential purposes. Findings from the officer surveys suggest effective use of BWV is not something that is down to the police alone, it needs other partners and all parts of the CJ system to be on board and be ready to use footage, making the most of BWV. When new recording technology is introduced within policing, such as CCTV and ANPR, there can often be concerns around privacy and accountability. The trial found the balance between the perceived benefits of BWV and concerns about privacy at present lean towards public acceptance for the technology, things may tip the other way if, however, perceived benefits of BWV are not realised. Currently, London residents are supportive of BWV, with their opinions of the technology positively associated with their views of how ‘procedurally just’ the police are, and their confidence in the MPS. However, public attitudes to BWV should be monitored, to understand if the technology is viewed as increasing transparency, or as an unnecessary intrusion into privacy. 28 Despite the limited impact demonstrated in the trial, there is still an argument BWV is a key tool to increase police transparency. In order for transparency to be maximised, BWV footage needs to be accessible, and decisions regarding video access, dissemination; and review will need to be made. Forces will need to consider how transparent they want to be, and this will be critical in preparation for high profile incidents where BWV is available. Routinely sharing footage may not be desirable, and is currently prohibitive due to expense, and the lack of automated redaction softwarexxiii. However, further research is needed to explore the impact of increased public exposure to BWV footage. Lessons learned from the implementation of this trial, shows that any force considering rolling out BWV needs to consider the practical challenges. These challenges include, timely training, support for technical issues, monitoring compliance with recording and back office policies and internal and external cultural change. Further research is required to understand if the impacts seen, or not seen, in this trial hold for officers in other roles, boroughs in the MPS and other force areas. Finally, this trial has found a number of very positive and innovative uses of BWV that were not being directly tested, but that officers have been sharing with each other. Officers reported a range of innovative uses of BWV, including professional development for officers individually and organisationally (sharing good practice and training); use of intelligence; and sharing information with partners (e.g. mental health organisations, social services) and the public. There were also benefits for evidence gathering and searching. Future Research In addition to the research gaps mentioned above, it has not been possible to capture an understanding of a number of related areas including:  Impact on officers who have no policing experience before BWV - new police joiners, who have never known a different way, may become more public-interaction focused and more reflective about their behaviour.  Wider impact on the Criminal Justice System - evidence led prosecutions and sentencing.  Lessons learnt from assessing footage of police interactions, how this can be used in training environments to assist professional development.  Time savings or increased efficiency as a result of BWV.  The cost-benefits across the entire CJ process.  If BWV can be used during community resolutions – particularly the ability to view footage during complaint resolution and its impact on both parties.  The other innovative uses highlighted across this report e.g. other crime types and incidents involving mental health needs. 29 Annexes Annex A: Methodology The intervention for treatment officers included 1) training officers to use BWV, 2) allocating personal issue BWV and 3) on-going supervision and guidance on use of BWV. 1) The training The MPS were responsible for providing the cameras, and for preparing the policy and training material. The training given to officers covered the following:  aims of the trial;  process for operational use, such as when to switch the camera on and off and the notification to give to the public when using the cameras;  statement taking process and disclosure considerations;  practical use of the cameras themselves; and  uploading, retention, storage issues and the back office process for providing footage to CPS. 2) The cameras The BWV cameras were attached to the outer vest of the officers, at shoulder/chest height (one borough, Borough 3, had some officers who wore a headcam which could be worn on the head or shoulder). Cameras had fixed lenses, capable of capturing 125 degrees horizontally and 70 degrees vertically. The cameras had a 30 second video only pre-buffer, so that when an officer presses the record button, the camera captures the previous 30 seconds of visual (not audio) information. The cameras had fixed non removable memory – docking the camera charged it and uploaded the footage to a cloud storage facility, clearing the camera ready for the next deployment. Footage would be manually labelled by the officer indicating the type of incident, whether the footage should be retained for evidential purposes and then linked to a case/crime. CPS and secondary investigators had access to the cloud and could view the footage when necessary. All footage unless marked as evidential would be automatically deleted after 31 days. 3) On-going supervision and guidance All response officers assigned a BWV camera were required to wear it whilst on duty, switching it on as prescribed by force policyxxiv, i.e. as soon as practicable for all stop and search encounters, domestic abuse incidents, use of force incidents and other mandatory situations74, as well as for evidential purposes. All other use was discretionary. If an officer agreed to wear a camera their use was monitored for compliance with MPS policy by the MPS BWV team, who provided regular feedback to boroughs during the trial – as well as creating amended policies and supporting documents as aide memoirs for officers to address common challenges. The MPS BWV team also collated good news stories and lessons learned which were disseminated in a monthly BWV news email. The team ran monthly meetings for single points of contact from each of the ten boroughs to receive updates/support on common issues and feedback – this included discussing barriers to compliance and technical support. In addition there were bi-monthly project management meetings where senior representatives from relevant areas of the MPS met and discussed the evaluation and related issues. The MPS BWV team gave on-going training support and trained trainers for each borough, so they could prepare new members of the response team who joined during the life of the trial and beyond. In addition, the MPS BWV team ran training for secondary investigators to ensure they were comfortable accessing and using BWV footage. The training support was in reaction to a lack of usage 74 Mandatory situations were, motor vehicle stops, premises searches, arrest enquiries, stop and account/search, statutory directions, use of force, critical incidents and domestic abuse incidents 30 identified by the BWV central team (see design limitations). One of the boroughs, Borough 4, had a digital magistrate’s court, so special support was given to take advantage of the Wi-Fi enabled Clickshare process –a new system which allows streaming of evidence into courtrooms and digital sharing of case files with CPS/Defence/Court users. The MPS dedicated a member of staff to work with the CPS (for all ten boroughs) and the digital court to maximise the benefits of digital working across criminal justice agencies. Methodology – trial design When a camera assigned officer permanently left an ERT (e.g. to change roles or move to restricted duties), they returned their camera to the borough to be reallocated to the new officer replacing them. New officers were allocated to response teams by local posting panels based on corporate need (e.g. team strength/skill). The posting process was the same for the control and treatment teams and there were no records of BWV being mentioned in any posting process. One of the main threats to an RCT design is contamination; where the control group do not remain completely ‘clean’ of the treatment and so no longer provide a good comparison of business as usual. In this instance there was a risk of officers with and without cameras attending the same incidents, so although only one officer is wearing the camera, both officers at the incident are affected by its presence. Contamination can also work at different times with the same officer. In the Rialto studyxxv shifts were randomly allocated for officers to either wear BWV or not, however reductions in complaints and use of force happened during both treatment and control shifts – suggesting officer’s behaviour was modified during the times they were wearing the camera and that carried over to when they were not. For this reason when officers changed ERTs, moving from treatment to control conditions, their time in the control condition was excluded. If moving from control to treatment, both periods were included. To balance a design which gives the greatest chance of detecting (statistically) a real difference and reducing the likelihood of contamination, team based allocation was chosen ahead of shift based or individual allocation because: 1. The outcomes of interest are linked to officer behaviour, and so vulnerable to contamination effects. If whole teams of officers wear the cameras then they should only be attending incidents with other officers in the same conditions most of the time. As the shifts rotate different teams will be policing the same area and community, just at different points in the day. While there may be a small period of time at the beginning and end of a shift when treatment and control teams will both be on duty (for hand over, etc.) one will usually be back at the station being briefed or writing up notes from the incidents attended during that shift. 2. In previous studiesxxvi usage compliance (ensuring officers wear and use the camera in line with force policy) has been an issue. It was hoped team based allocation would increase compliance for two reasons. First it more accurately replicates ‘real world’ conditions as peers are all wearing the equipment, learning and sharing experiences. Hesitant officers would be unable to avoid BWV entirely as they would attend incidents with colleagues who are recording. Second, it supports enforcement of compliance as supervisors know everyone on their team should be wearing BWV and are able to encourage usage. In addition, messages from the central team about usage, good practice, and policy/guidance could be easier disseminated to an entire team by feeding into briefings. Team-based allocation was seen operationally as the most feasible design, which would ease the administrative burden of running the trial75. 75 The practicalities of alternative implementation could have been a barrier, for many reasons including the cost of IT infrastructure; monitoring officers in the trial; and reassignment of cameras to new officers. 31 Figure 8: Trial CONSORT diagram Before the final decision to allocate the intervention based on teams rather than shifts or officers was made, analysis was undertaken to examine how ERT teams across the ten boroughs compared on our three key outcome measures. The analysis found bigger differences in ERT outcomes between boroughs than within borough. Overall, this suggested that a design where teams in each borough could be compared would help to isolate the impact of the introduction of BWV on our key outcome measures. Whilst the teams are broadly equivalent allowing strong statements about impact to be made for the boroughs, the generalisability of results are limited, as the selection of the ten boroughs was designed to maximise the likelihood of detecting a difference. Therefore the results from this trial may not be directly translated into other boroughs in the MPS, which may vary in important ways to those chosen to be part of the trial. Figure 8 explains the inclusion, allocation, follow-up and analysis stages. After teams had been allocated to treatment and control groups, further analysis was undertaken to compare key demographic information. The random assignment appeared to have been successful as comparisons of key demographic information (proportion of females, BME and age ranges76) show the treatment and control groups to be broadly equivalent before the introduction of BWV cameras, which allows more confidence that comparison between the two groups would be fairxxvii. Thus any difference after can be directly attributed to the intervention. One possible explanation of any identified effects of BWV is that they are the product of officers knowing they are taking part in an experiment. Officers may be aware that the organisation is interested in the impact of BWV on CJ outcomes, reducing complaints, stop and search activity and that their activity will be monitored, so alter their behaviour as a result. Officers may also perceive an expectation that they (and colleagues) will perform better if they wear cameras. If this were the case, then this expectation in itself could potentially drive performance improvement. While the aim was for all officers in the treatment teams to receive the intervention (BWV), in practice, not all ended up being sent on training and issued with a camera as intended. This was for a variety of reasons (e.g. sickness), but most commonly because there was an underestimate of the number of officers who would need a camera for personal issue (as there were more officers than anticipated – see figure 3, there were 596 in the original allocation) in all ten boroughs77. In July 2014 it was clear there was a shortfall of around 180 cameras. To reduce the impact of officers without cameras, the decision was made to distribute remaining cameras to treatment teams across a smaller number of boroughs to enable them to reach an 80% saturation as quickly as possible (so that if no more cameras became available they could be a smaller trial with a greater fidelity to the intervention). To address this potential allocation bias all officers in both groups were included in the analysis. This saturation management was important as the order in which officers went on training and received cameras was not random, and so an element of bias could have influenced early use and boroughs with a low saturation (e.g. the most motivated officers could have volunteered for training 76 Both groups had 26% female officers. Age distribution and number of years in service were very similar too. Ethnicity was very slightly different, with 15% of the officers recorded as BME in the treatment group compared to 11% of the control group. 77 Camera were also needed for a range of reasons including for training, to replace those lost or broken, as well as those taken by officers who transferred teams. *The random allocation of two teams to treatment and control in one borough was not followed therefore these two teams were excluded from all analysis. 32 first and been the only officers with cameras). This decision in practice had little long term impact, as more cameras were found for the trial and as shown in table 1, the average saturation of camera wearing officers in treatment teams across boroughs is broadly equivalent. Table 1: Average officer saturation, uploaded clips and evidentially marked clips Despite some treatment group officers not being trained and/or given cameras, it nevertheless remained important to include them in the analysis – known as intention to treat analysis- (see figure 3 flow of officers). Exclusion could have biased results, as those who did not attend training or were not issued with cameras could have been different to those who did (e.g. in terms of motivation, their attitudes towards BWV or newness to the team). The inclusion of all officers randomly assigned to the treatment group in the analysis – regardless of whether they received the training, a camera, or neither – provides a better ‘real world’ assessment of the impact of the intervention. The same principles apply to officers who have been issued with a camera but do not follow the usage policy and chose not to record situations. This analytical approach would also seem best as officers attending an incident as part of a treatment team would likely be exposed to aspects of the BWV intervention when colleagues used BWV as instructed. Team dynamics and the presence of cameras at jointly attended incidents could impact on both officers with and without cameras, as well as all others who come into contact with them (e.g. members of the public, suspects, victims). To take account of varying usage rates between officers, statistical modelling undertaken in our analysis included total number of clips, total length of clips and length of time with a camera, for all key outcome measures, and whether a case had an associated video recording for CJ outcomes. Due to the different start dates for each borough, and the overall shortfall in available cameras, it would not be representative to analyse outcome measures in the months where boroughs had no cameras – as this would effectively compare two groups, both without cameras, and would not be a test of the impact of BWV. To make analysis a fair test of the intervention itself, each borough was given its own trial start date. Boroughs were deemed to go live for evaluation purposes once more than 50% of officers in a treatment team had uploaded an operational clip78. On two occasions a borough was deemed live prior to reaching 50% saturation, as the borough saturation level was marginally below 50%, equating to a shortfall of only a few officers79. The decision was based on four factors: having enough usage of the technology to test the true impact of BWV; counteracting any potential bias from boroughs selecting a particular type of officers to receive training/equipment early on; the trial being a ‘real world’ test of BWV; and having enough time for the outcome measures to be detectable. 78 Operational refers to any clip uploaded that is not categorised or flagged as for training purposes. In the event that an officer has never categorised or flagged a clip as training purposes, the 4th clip they uploaded is deemed to be their first operational clip, as the majority of officers who had correctly labelled their uploaded clips upload between 1 and 3 clips during initial training. 79 Borough 1 became live at 42% saturation in September 2014 as only 4 cameras were required to reach 50%. Borough 5 became live at the beginning of the trial period, May 2014, at 46% saturation as only 3 officers were required to make 50% and more than 3 officers uploaded operational footage within the first week of June 2014. These numbers were negligible, but significantly increased the boroughs’ time in the trial. Borough Officers Trained (%) Officers Uploaded Clip (%) Officers Uploaded Evidential Clip (%) BOROUGH 1 62% 55% 44% BOROUGH 2 82% 77% 66% BOROUGH 3 69% 55% 41% BOROUGH 4 77% 70% 63% BOROUGH 5 79% 70% 55% BOROUGH 6 92% 83% 68% BOROUGH 7 88% 80% 67% BOROUGH 8 94% 86% 61% BOROUGH 9 87% 78% 53% BOROUGH 10 92% 82% 66% 33 Figure 9: Flow of Officers in the Trial Workforce Churn Further potential limitations for the evaluation came from workforce churn, i.e. the movement of officers into and out of the ERT teams. Average workforce churn was monitored throughout the project and was higher than anticipated, with a monthly staff turnover of between 4% and 9% (approximately 20-50 officers leaving treatment teams and being replaced each month). This effectively caused a continuous training gap, with the central MPS BWV implementation team having to constantly address potential resistance from officers, and ensure they had the ability to use the technology and access equipment. This challenge confirms the ‘real world’ nature of the trial and whilst work was done with the central BWV implementation team to lessen the effect (e.g. monthly tracking of leavers and joiners) officer churn remained a threat to both the implementation of the trial and our ability to assess impact throughout the evaluation. * The number of officers entering the trial each month as a new borough goes live: May 14: 178(1 borough); July 14: 541 (4 boroughs); Aug 14: 539 (3 boroughs); Sep 14: 154 (1 borough); Feb 15: 98 (1 borough) **Where officers have moved from treatment to control, only the treatment period of their trial participation has been analysed to avoid contamination of any treatment effect in control conditions. Where officers have moved from control to treatment, officers have been split into two records and analysed as separate participants. ***Includes 12 officers who moved from control to treatment NB: 135 officers (61 treatment, 74 control) left the trial ERTs temporarily (e.g. extended annual leave, secondment) and then returned to a team in the same borough and condition. Data relating to officers during the time of absence has been removed from analysis however to avoid complication, this is not shown in the diagram above. This movement is not shown in the diagram to avoid complication. 34 Annex B: Secondary Data Sources and Analysis Officer Interviews There were 19 officers from four boroughs selected to take part80 in semi-structured face to face interviews with College of Policing researchers. For practical purposes, boroughs chose which specific officers were interviewed, which may add an element of bias; however they were provided with purposive sampling criteria in order to obtain a variety of experiences with BWV, which were a range of ages, genders, ethnic backgrounds, lengths of service and BWV usage patterns. Interviews lasted 30 minutes to an hour, covering how the cameras were used by officers; how BWV may have affected their behaviour and decision making; the quantity and quality of the evidence they were gathering; their confidence and accountability; the impact cameras had on discretion; and team dynamics. More general implementation issues were also covered. As part of the interviews officers were asked to show one clip that best demonstrated their experience of using the cameras. This could include any benefits or challenges they had encountered; the choice of video was entirely at the officers’ discretion. This provided participant led data on officers’ experience and acted as a point of reference to explore officers’ usage patterns during the interviews. The interviews were audio recorded, transcribed, and thematically analysed using dual coding to ensure consistency. User Satisfaction Survey (USS) MOPAC administer the MPS User Satisfaction Survey (USS)xxviii, a random telephone survey of crime victims. It is a mandated assessment of satisfaction with those who have experienced police service provision as a victim of crime81. Regular analysis looks at overall satisfaction as well as satisfaction with: initial police action; treatment; and follow up. The USS is a random survey of victims, and some officers in the trial would have had more contact experiences than others reflected in the data. To link with the impact of BWV, cases were given a BWV marker, so analysis could occur on anonymised records at borough level82. BWV marker cases were then compared to cases without a marker, which would include a limited number of non-ERT officers. A total of 4,042 cases were used for analysis (1,045 with BWV and 2,997 without), from interviews taking place between July 2014 and June 2015. Stop & Search survey A market research company, Opinion Research Services, were commissioned to conduct a survey of those members of the public stopped and searched by officers in the trial. The public were provided with a ‘survey card’ by the searching officer immediately following a stop with details of how to respond via an online, text/SMS or automated telephone survey. Using four short questions on scale ranging from 1 to 7, respondents were asked to indicate the extent to which they agreed the officer was respectful; professional; that a clear reason for the search was given; and whether they felt the search was justified. Officers in treatment teams were provided with different cards to those in control teams to allow comparison of responses. The survey ran from the 17th of March 2015 to the 31st of May 2015, however no results are reported here due to insufficient responses to make any meaningful conclusions (n = 14). 80 Reason for borough selection: Borough 3 - consistently low saturation rate and the shortest trial period; Borough 4 - a digital magistrates courts; Borough 5 - the longest active trial period; Borough 9 - consistently high saturation rate. 81 Approximately 16000 victims spoken to each year via a telephone interview, which asks about their contact, the response and how they were treated by police. Crime types include Domestic Burglary; Violent Crime; Vehicle Crime and Hate Crime. Exclusion criteria: Under 18; Domestic Violence; Police Officers on duty. 82 Due to its complications with team allocation the borough of Borough 3 was removed. 35 Public Attitude Survey (PAS) The general public were consulted around their views of BWV through MOPAC’s Public Attitude Survey (PAS)xxix. The PAS is well established83, consulting London residents on a host of policing related topics. For the purposes of the trial some additional questions relating to the general awareness and opinions of BWV were added. Due to the randomisation of ERTs within boroughs during the BWV trial, before and after or between borough comparisons of PAS outcomes was more difficult, so rudimentary comparisons were made between trial and non-trial boroughs. This is limited, as the trial boroughs only have relatively few officers with cameras, and some non-trial boroughs may have historic pockets of camera use. There may also be differences in the trial and non-trial boroughs, based on the selection criteria used. Sentiment Analysis Sentiment analysis of Twitter activity was attempted through the Cosmos (Collaborative Online Social Media Observatory) platform to try to discover public opinion and first hand experiences BWV cameras. Key word searches (e.g. Body Worn Video), were performed for five days, however due to a very small number of relevant tweets and volume of unrelated tweets picked up no conclusions could be made from the research. This approach was then discontinued. Officer attitudes All officers in the treatment and control groups were asked to complete an online questionnaire. Due to the staggered ‘go live’ dates they were administered approximately at the beginning, middle and end of the 12 month trial period. Surveys were intended to give some understanding of the impact on officer attitudes and self-reported behaviour, and provide implementation learning. Respondents were asked the extent to which they agreed or disagreed with a series of statements about their interactions with the public, the organisation and decision making using a scale ranging from 1 (strongly agree) to 7 (strongly disagree). The statements were designed to enable comparison of treatment and control officers’ attitudes and self-reported behaviour. No reference to BWV was made in the statements allowing officers in the trial to respond to the same statements. Statements are shown in the tables below. In addition, treatment officers were asked BWV specific questions to provide implementation learning. The questionnaire was designed so that single questions could be combined to give an overall understanding of an underlying concept. For example a score for the concept of ‘certainty of sanctions’ was created from three statements in the attitudes survey: • I would be held to account if I under-performed in my role. • I would be challenged if I didn’t do my job properly. • I would be disciplined if I broke the rules. Some questions did not fit into a concept and therefore were analysed as a single indicator. The difference between the treatment and control groups questionnaire responses were tested via a range of statistical approaches (e.g. T-tests, Mann Whitney and Chi-squared). Exploratory factor analysis84 was undertaken on the responses from the second and third surveys’ which included attitudinal statements separately. For an item to form part of a scaled variable it needed to have a factor score of 0.585 or more86. The reliability of each scale was then checked using Cronbach’s Alpha test. For a scale to be deemed suitably reliable, an Alpha score of at least 0.6 was required. This threshold was kept relatively low to allow the inclusion of some 3-item scales, as the 83 First conducted in 1983 to give the MPS an understanding or the views of London residents. 84 Extraction Method: Maximum Likelihood, Scree plot. Rotation Method: Direct Oblimin. 85 One item was included that had a factor score of 0.48 as it made sense as part of the theme. 86 Following guidance Field, A (2009), Discovering Statistics Using SPSS, third edition. 36 Alpha score is directly related to the number of items in a scale. The tables below show the results of the factor analysis for both rounds of the survey. Differences in the mean scores of the newly created factors between the treatment and control groups were compared initially using a t-test. Where scaled variables were not normally distributed transformation was performed. In some cases the data was unable to be normalised through transformation, so non-parametric Mann Whitney tests were also performed on all scaled variables. As the results of t-tests and Mann Whitney tests were very similar for the normally distributed data, the results of the Mann Whitney tests have been reported for convenience. Some questionnaire items did not load onto any scale so were treated as single indicators. In such instances Mann Whitney tests were performed on these items, as well as chi square tests (agree (1 – 3 on scale) vs. else (4 – 7 on scale)) to assure results. In these instances, only results with a p value of <0.01 have been reported to ensure reliability of results. 37 Factor Analysis Results (Officer Survey Round 3) Scaled Variable Contributory Variables Factor Loading Reliability Statistics Cronbach’s Alpha N of Items Protection Against Complaints If a member of the public made a false complaint against me, it would get sorted out quickly. I’m confident people would believe my side of the story if a complaint was made against me. A complaint made against me with little evidence would be dropped quickly. I feel protected from malicious complaints when dealing with the public. .92 .60 .83 .73 .86 4 Procedurally Just Policing I explain the reason for a stop without fail. I make certain the person I am searching understands why they are being searched. I take the time to explain my decisions to members of the public. I allow members of the public to voice their opinion when decisions are made that affect them. I am polite to everyone regardless of how they behave. .72 .74 .61 .53 .50 .73 5 Public Co- operation The public I deal with tell me about suspicious activity. The public I deal with willingly assist me when asked. The public I deal with willingly provide me with information about suspects. When dealing with an incident, the public will follow my instructions. People are usually polite to me when I’m dealing with an incident. .55 .77 .63 .66 .64 .80 5 Frequency of Force On average, how often in the past 3 months have you drawn your baton? On average, how often in the past 3 months have you restrained a suspect on the floor (e.g. ground pin)? On average, how often in the past 3 months have you used a strike (e.g. palm heel strike/ knee strike)? .53 .54 .91 .66 3 Quality of Evidence I can gather good quality evidence at the incidents I attend. It’s easy to get across how people behave in the evidence I gather. I have confidence that the evidence I gather will be sufficient to charge a suspect. .56 .65 .56 .61 3 Compliance I always follow the correct procedure when attending an incident. I make sure I stick to the protocols when dealing with an incident. I always do thing strictly by the book whilst I’m out on shift. .77 .86 .75 .78 3 Change in Accountability I think twice before I stop and search someone nowadays. I need stronger grounds to stop and search someone than I did last year. I need stronger justification to use force than I did last year. I feel I need to justify the actions I take at incidents more than I used to. .66 .88 .48 .50 .74 4 Certainty of Sanctions I would be held to account if I under-performed in my role. I would be challenged if I didn’t do my job properly. I would be disciplined if I broke the rules. .75 .71 .73 .77 3 Supervisor feedback I would gain recognition if I performed well in my job. It is acknowledged when I deal with an incident well. I would be praised if I showed notable improvement in my performance. If I delivered excellent service to a victim it would be recognised by my supervisors. My supervisor gives me constructive feedback that helps me to improve my performance. My supervisor helps me to identify areas where I could improve. My supervisor encourages and supports my development. .79 .80 .85 .80 .90 .90 .79 .94 7 38 Analysis of Scaled Variables (Officer Survey – Round 3) Scaled Variable Median Response Mann Whitney Treatment Group Control Group U P Effect Size Protection Against Complaints 4.50 3.75 30875.5 .000* 0.21 Procedurally Just Policing_v2 6.20 6.20 40028.5 .608 0.02 Public Co-operation_v2 4.40 4.40 38590.0 .216 0.05 Frequency of Force_v2 5.34 5.34 38527.5 .196 0.05 Quality of Evidence 5.50 5.00 32932.0 .000* 0.17 Compliance 6.00 6.00 39476.0 .424 0.03 Change in Accountability 5.13 5.25 36924.0 .037* 0.09 Certainty of Sanctions 6.33 6.33 40545.0 .800 0.01 Supervisor feedback 4.43 4.24 39955.5 .584 0.02 NB: T-tests were also performed on all scaled variables using the original or normalised data following transformation. Results for every variable was similar to that shown in the table. * Significant 39 Factor Analysis Results (Officer Survey Round 2) Scaled Variable Contributory Variables Factor Loading Reliability Statistics Cronbach’s Alpha N of Items Protection Against Complaints If a member of the public made a false complaint against me, it would get sorted out quickly. I’m confident people would believe my side of the story if a complaint was made against me. A complaint made against me with little evidence would be dropped quickly. I feel protected from malicious complaints when dealing with the public. .79 .65 .77 .76 .83 4 Procedurally Just Policing I explain the reason for a stop without fail. I make certain the person I am searching understands why they are being searched. I take the time to explain my decisions to members of the public. I allow members of the public to voice their opinion when decisions are made that affect them. .72 .70 .69 .50 .71 4 Public Co-operation The public I deal with willingly assist me when asked. The public I deal with willingly provide me with information about suspects. When dealing with an incident, the public will follow my instructions. People are usually polite to me when I’m dealing with an incident. .78 .60 .63 .66 .76 4 Frequency of Force On average, how often in the past 3 months have you drawn your baton? On average, how often in the past 3 months have you restrained a suspect on the floor (e.g. ground pin)? On average, how often in the past 3 months have you used a strike (e.g. palm heel strike/ knee strike)? .56 .56 .81 .63 3 Quality of Evidence I can gather good quality evidence at the incidents I attend. It’s easy to get across how people behave in the evidence I gather. I have confidence that the evidence I gather will be sufficient to charge a suspect. .60 .69 .56 .64 3 Change in Accountability I think twice before I stop and search someone nowadays. I need stronger grounds to stop and search someone than I did last year. I need stronger justification to use force than I did last year. .59 .90 .70 .76 3 Certainty of Sanctions I would be held to account if I under-performed in my role. I would be challenged if I didn’t do my job properly. I would be disciplined if I broke the rules. .67 .86 .57 .73 3 Supervisor Support My supervisor gives me constructive feedback that helps me to improve my performance. My supervisor helps me to identify areas where I could improve. My supervisor encourages and supports my development. .97 .89 .82 .92 3 40 Analysis of Scaled Variables (Officer Survey Round 2) Scaled Variable Median Response Mann Whitney Treatment Group Control Group U P Effect Size Protection Against Complaints 4.00 3.25 21678.5 .000* 0.26 Procedurally Just Policing_v1 6.00 6.25 28876.5 .204 -0.06 Public Co-operation_v1 4.00 4.25 29672.0 .441 -0.03 Frequency of Force_v1 6.33 6.33 30494.0 .793 0.01 Quality of Evidence 5.33 5.33 28828.5 .194 0.06 Change in Accountability_v1 4.33 4.33 29432.5 .358 -0.04 Certainty of Sanctions 6.33 6.33 30446,5 .770 0.01 Supervisor support 4.66 4.66 29293.0 .314 -0.04 NB: T-tests were also performed on all scaled variables using the original or normalised data following transformation. Results for every variable was similar to that shown in the table. * Significant 41 Annex C: Body Worn Video Usage BWV usage data was downloaded from the cloud storage system and used by MPS to save, share and use footage for the duration of the trial. The data included all clips uploaded between May 1st 2014 and April 30th 2015. Significant data cleansing was undertaken to exclude all duplicate clips and clips created from original footage, which would falsely inflate the volume of clips recorded. The analysis presented below is of footage uploaded by officers in treatment teams of the trial boroughs only and excludes any clips uploaded prior to the officer’s borough live date or any other officers in the MPS who had the equipment. This is to present usage patterns in the context of the outcome data analysed (e.g. complaints, violent incidents etc.), which also excludes any incident occurring prior to each borough’s trial ‘go live’ date. All clips recorded prior to borough ‘go live’ dates are considered to be for training and familiarisation purposes and are not included in the report. 1) Overall BWV usage 1.1. Total trial figures During the trial, BWV cameras were used to generate 48,086 clips, totalling 12,156 hours of video (an average of 14.9 minutes for each recording submitted). Of all the clips submitted, 13,616 (28%), totalling 4,650 hours, were tagged as “evidential” for potential use within the criminal justice system (CJS). Figure 10 presents the total number of clips recorded each month, with the red data points indicating months when boroughs became operational in the trial. Overall, the total number of clips recorded per month increased reaching 6,770 clips (1,726 hours) recorded in April 2015. From February 2015 to April 2015, when all boroughs were operational in the trial, the average number of clips per month was 6,127 (1,563 hours – just over 65 days of footage). Finally, the average length of clips remained steady throughout the trial ranging from 14.6 to 15.8 minutes. Figure 10: The total number of clips recorded per month by all boroughs: *Red data points indicate the month when different boroughs became operational in the trial. Table 2. shows overall usage during the trial at the borough level. The total number of clips recorded vary depending on the borough, as well as the total length and average length. However, at this level figures are not comparable between boroughs due to the different length of time each borough spent ‘live’ in the trial. 42 Table 2: Borough level clips recorded during the trial Boroughs Total Number of Clips Total Length of Clips (hours) Average length of clips (min) Month borough became operational Borough 1 3363 990 17.7 Sept 2014 Borough 2 3410 760 13.4 Jul 2014 Borough 3 321 71 13.3 Feb 2015 Borough 4 5404 1226 13.6 Aug 2014 Borough 5 7214 2004 16.7 May 2014 Borough 6 7368 1709 13.9 Jul 2014 Borough 7 6556 1674 15.3 Aug 2014 Borough 8 4648 1084 14.0 Jul 2014 Borough 9 4645 1488 19.2 Jul 2014 Borough 10 5157 1150 13.4 Aug 2014 1.2. Evidential Footage Table 3. shows the total number of clips flagged as evidential and the corresponding percentages per borough. The percentage of evidential clips tends greatly between boroughs, with Borough 8 being the lowest (14%) and Borough 3 being the highest (42%). Overall, from February 2015 to April 2015 the percentage of evidential footage remains quite steady ranging from 28% to 35%. Table 3: Evidential Footage recorded per borough Boroughs Total Number of Clips Recorded % of the overall footage collected Borough 1 1221 36% Borough 2 817 24% Borough 3 136 42% Borough 4 2216 41% Borough 5 1946 27% Borough 6 1714 23% Borough 7 2110 32% Borough 8 668 14% Borough 9 1262 27% Borough 10 1526 30% Average 28% 1.3. Storage Requirements Analysis of the amount of data in gigabytes (GB) generated during the trial has been undertaken based upon rolling totals of 31 day periods. This method is chosen over analysis of monthly totals to allow for busy periods that straddle two months to be identified. 31 day periods were also used as this is the length of time a clip not marked as evidential is retained for. The maximum number of GB generated within a 31 day period was between 28th of March 2015 to 27th of April 2015 – totalling 1377GB (7227 clips) from which 71% (5103 clips) were marked as non-evidential and deleted after 31 day period. The maximum GB generated for evidential footage was 566GB (2203 clips) in the period from 24th of January 2015 to 23rd of February 2015. The data used for this analysis does not show how long evidential footage was retained, as it is still required so not deleted during the trial. However, these figures are only applicable for the technology used in this trial, as different technology produces clips of different file sizes. 43 2) Temporal Analysis Figure 11. displays the total number of clips recorded and the corresponding length during each hour of each day of a week. These figures are based on the start date and time of each clip. The number of clips recorded fluctuate depending of the time of a day or day of a week. Between 23:00 – 00:00 on Friday has the largest number of clips recorded in any given hour - 591 (144 hours) and between 23:00 – 00:00 Saturday night has the second largest number of clips recorded – 517 (121 hours). An interesting pattern emerges from the analysis; weekdays tend to have fewer but longer clips, while the weekends have more but shorter clips. One explanation of this pattern may be officers have more time to deal with incidents during periods of the week that tend to have fewer incidents or are attending different sorts of incidents at different days/times. Additionally, morning periods between 4:00 to 7:00 tend to have the fewest and the shortest overall duration of clips. Peaks in recordings may be a result of shift patterns rather than a result of increased demand on officers. Figure 11: The total number and total length of clips recorded per hour of the week Table 4. displays the total number of clips and their corresponding total length depending of the day of a week. It should be noted that each total relates to all clips recorded between 00:00 – 23:59. Therefore a high proportion of clips recorded on Sundays will in fact relate to the early hours of the morning during officers’ Saturday night shifts. Table 4: Differences in clips recorded and total length of clips via the day of the week Day of Recording Total Number of Clips Recorded Total Length of Clips Recorded (hours) Sunday 7469 1875 Monday 6440 1654 Tuesday 6443 1666 Wednesday 6538 1657 Thursday 6945 1758 Friday 6762 1697 Saturday 7489 1849 44 3) Officer usage of BWV The profile of officer usage was covered in the implementation section. However, in addition. Table 5. displays the rate of clips recorded per officer per month depending for each borough. This was calculated by dividing the total number of clips in a given operational month with the total number of police officers with cameras per borough, to allow easier comparison despite the variance in the boroughs’ time in the trial. Figures 7 and 8 demonstrates the inconsistency in officers’ use of BWV throughout the trial. Table 5: Rate of clips recorded per officer per month Borough Rate: clips/per officer/per month Borough 1 12 Borough 2 10 Borough 3 5 Borough 4 15 Borough 5 11 Borough 6 10 Borough 7 10 Borough 8 11 Borough 9 9 Borough 10 9 Figure 12. illustrates the total number of clips recorded depending on their length in minutes, however due to infrequency of clips longer than 100minutes, these were not included in the Figure 9. The most frequent length of clip recorded in the trial was between 2 – 4 minutes long (15.7%, 7560 clips). The longest clip recorded during the trial was 482min (8 hours) and the median clip length was 5 minutes 48 seconds. 45 Figure 12: Number of clips recorded depending on the length of a clip in minutes . 46 Annex D: Criminal Justice Case Attrition All incidents reported to the Metropolitan Police between 1 st May 2014 and 30 th April 2015 147,986 Incidents where the 'reporting officer' was in treatment or control group, classified as 'Crimed' 64,355 Incidents removed due to classified as: CRI (3283), No Crime (1024), Transferred Out (390), Other Accepted Crime (22,184). 26,881 Incidents taken from after each borough 'Go live' date 92,035 Incidents removed where contamination identified (officers or boroughs) 784 Incidents where any suspect was identified 31,682 (49% of all cases for analysis) Incidents where no suspects were identified 32,673 Total number of suspect across all incidents 37,301 Cases with a suspect, but no arrest 10,078 Cases where at least one suspects was arrested 21,604 (68% of cases with a suspect) In half of cases with a suspect (17,781; 56%), they were present at the scene. The majority of these suspects (14,794; 83%) were arrested. Cases without an arrest with a judicial outcome (mainly cannabis warning/community resolution) 1,167 Arrested suspect, was Charged/Summons by the police 8,849 (41% of those arrested) Arrested suspect, was Cautioned, Warning or other outcome (e.g. community resolution, Cannabis warning, Penalty notice) from the police 3,868 (18% of those arrested) Currently No Outcome 8,887 (41% of those arrested) Cases where an arrested suspect received a Judicial Outcome 12,717 (59% of those arrested) 47 References iGuidance for the Police use of Body-Worn Video Devices (2007). Police and Crime Standards Directorate. Home Office and ACPO; College of Policing. (2014). Body Worn Video http://library.college.police.uk/docs/college-of- policing/Body-worn-video-guidance-2014.pdf iiDrover, P., & Ariel, B. (2015). Leading an Experiment in Police Body-Worn Video Cameras. International Criminal Justice Review, 1-18. iiiBromley, Camden, Havering and Wandsworth’s Transport OCU. iv Kingston, Camden, Enfield, Havering, Newham, Westminster, Sutton, Lambeth and at Heathrow. vhttp://content.met.police.uk/Site/about vi Ellis, T, Jenkins, C., and Smith, P. (2015) Evaluation of the Introduction of Personal Issue Body Worn Video Cameras (Operation Hyperion) on the Isle of Wight: Final Report To Hampshire Constabulary. http://www.bwvsg.com/wp-content/uploads/2015/03/Ellis-Evaluation-Worn-Cameras.pdf viiAriel, B., Farrar, W. A., & Sutherland, A. (2014). The effect of police body-worn cameras on use of force and citizens’ complaints against the police: A randomized controlled trial. Journal of Quantitative Criminology, 1-27. Advance online publication 19 November 2014. viiiRankin, L. (2013). End of Program Evaluation/Recommendations: On-Officer Body Camera System, Mesa Police Department Report (Mesa, AZ: Mesa Police Department). http://issuu.com/leerankin6/docs/final_axon_flex_evaluation_12-3-13- ixHome Office (2007) Plymouth Head Camera Project – Body Worn Video Recording System (Head Cameras): National Pilot, Final Report, April 2007, found as Annex A here: http://library.college.police.uk/docs/homeoffice/guidance-body-worn-devices.pdf xMiller, J., Bland, N., and Quinton, P. (2000) The Impact of Stops and Searches on Crime and the Community, Police Research Series Paper 127, London: Home Office. xiStanko, E., Jackson, J., Bradford, B., and Hohl, K. (2012) A golden thread, a presence amongst uniforms, and a good deal of data: studying public confidence in the London Metropolitan Police, Policing and Society: An International Journal of Research and Policy, 22:3, 317-331; Bradford, B. (2014) Policing and social identity: procedural justice, inclusion and cooperation between police and public, Policing and Society: An International Journal of Research and Policy, 24:1, 22-43. xiiPolice Executive Research Forum (PERF), with support from the U.S.Department of Justice’s Office of Community Oriented Policing Services (COPS Office). (2013). Implementing a Body-Worn Camera Program: Recommendations and Lessons Learned. http://www.policeforum.org/assets/docs/Free_Online_Documents/Technology/implementing%20a%20body- worn%20camera%20program.pdf; Bradford, B. (2014) Policing and social identity: procedural justice, inclusion and cooperation between police and public, Policing and Society: An International Journal of Research and Policy, 24:1, 22-43, DOI: 10.1080/10439463.2012.724068; Jackson, J., Bradford, B., Hough, M., Myhill, A., Quinton, P., and Tyler, T.R. (2012) Why do people comply with the law? Legitimacy and the influence of legal institutions. British Journal of Criminology 52 (6), pp. 1051-1071. ISSN 0007-0955. Downloaded from: http://eprints.bbk.ac.uk/5041/; Miller, L., & Toliver, J. (2014) Implementing a Body-Worn Camera Program: Recommendations and Lessons Learned. Police Executive Research Forum (PERF), Washington, DC: Office of Community Oriented Policing Services, with support from the U.S. Department of Justice’s Office of Community Oriented Policing Services (COPS Office). (2013). http://www.policeforum.org/assets/docs/Free_Online_Documents/Technology/implementing%20a%20body- worn%20camera%20program.pdf xiiiOwens, C., Mann, D., Mckenna, R. (2014). The Essex Body Worn Video Trial: The impact of Body Worn Video on criminal justice outcomes of domestic abuse incidents: http://college.pressofficeadmin.com/repository/files/BWV_Report.pdf xivHome Office (2007) Plymouth Head Camera Project – Body Worn Video Recording System (Head Cameras): National Pilot, Final Report, April 2007, found here: http://library.college.police.uk/docs/homeoffice/guidance-body-worn-devices.pdf xvMOPAC press release (2015) Mayor confirms plans to roll out 20,000 police body cameras across the Met. https://www.london.gov.uk/media/mayor-press-releases/2015/06/mayor-confirms-plans-to-roll-out-20000- police-body-cameras 48 xviStrang, H. (2012). Coalitions for common purpose: Managing relationships in experiments. Journal of Experimental Criminology, 8, 211-225; London Criminal Justice Partnership. (2011). An evaluation of the Diamond Initiative: year two findings http://www.londoncjp.gov.uk/publications/diamond_year2_FINAL_050511.pdf xviiDrover, P., & Ariel, B. (2015). Leading an Experiment in Police Body-Worn Video Cameras. International Criminal Justice Review, 1-18. xviiiOwens, C., Mann, D., Mckenna, R. (2014). The Essex Body Worn Video Trial: The impact of Body Worn Video on criminal justice outcomes of domestic abuse incidents: http://college.pressofficeadmin.com/repository/files/BWV_Report.pdf xix http://www.bwvsg.com/wp-content/uploads/2015/03/Ellis-Evaluation-Worn-Cameras.pdf xxOwens, C., Mann, D., Mckenna, R. (2014). The Essex Body Worn Video Trial: The impact of Body Worn Video on criminal justice outcomes of domestic abuse incidents: http://college.pressofficeadmin.com/repository/files/BWV_Report.pdf xxiMiller, L., & Toliver, J. (2014) Implementing a Body-Worn Camera Program: Recommendations and Lessons Learned. Police Executive Research Forum (PERF), Washington, DC: Office of Community Oriented Policing Services, with support from the U.S.Department of Justice’s Office of Community Oriented Policing Services (COPS Office). (2013). http://www.policeforum.org/assets/docs/Free_Online_Documents/Technology/implementing%20a%20body- worn%20camera%20program.pdf xxiiOpinion Research Services: Public Attitude Survey 2014-2015. Technical Report, Quarter 38. (2014). http://www.met.police.uk/about/performance/confidence.htm xxiiiLa Vigne, N. (2015). Five Myths about Body Cameras. Washington Post. http://crimeandjustice.org.uk/ xxivMPS policy is availbale here: http://www.mayorwatch.co.uk/labour-urge-met-to-listen-to-communities- over-body-camera-trial/ xxvAriel, B., Farrar, W. A., & Sutherland, A. (2014). The effect of police body-worn cameras on use of force and citizens’ complaints against the police: A randomized controlled trial. Journal of Quantitative Criminology, 1-27. Advance online publication 19 November 2014. xxviOwens, C., Mann, D., Mckenna, R. (2014). The Essex Body Worn Video Trial: The impact of Body Worn Video on criminal justice outcomes of domestic abuse incidents: http://college.pressofficeadmin.com/repository/files/BWV_Report.pdf; Ellis, T, Jenkins, C., and Smith, P. (2015) Evaluation of the Introduction of Personal Issue Body Worn Video Cameras (Operation Hyperion) on the Isle of Wight: Final Report To Hampshire Constabulary. http://www.bwvsg.com/wp-content/uploads/2015/03/Ellis- Evaluation-Worn-Cameras.pdf xxviiShadish, W.R., Cook, T.D., Campbell, D.T. (2002) Experimental and quasi-experimental designs for generalized causal inference, Houghton Mifflin, New York. xxviiiHome Office (2015).Victim Satisfaction Surveys – Survey Guidance 2015/16 xxixOpinion Research Services: Public Attitude Survey 2014-2015. Technical Report, Quarter 38. (2014). http://www.met.police.uk/about/performance/confidence.htm The Lighthouse: London’s Child House Initial Evaluation Report December 2018 Lynne Conroy, Dr. Zoe Hobson, Rachael Parker & Tim Read MOPAC Evidence & Insight 2 Contents Executive Summary 3 1. Background: Setting the Scene 5 The story so far… 5 Child House - a new approach 6 The Lighthouse - a Child House for London 7 Figure 1: Enhanced aspects of the Lighthouse 8 2. The Lighthouse Evaluation 9 What does success for the Lighthouse look like? 9 Table 1: Overview of Lighthouse Objectives 10 Methodology 11 Figure 2: E&I Evaluation Process 12 Figure 3: Evaluation progress 12 Ethics and Oversight 14 Research timeline 15 3. Learning from Baselining the Service 15 Base-lining Criminal Justice 15 Figure 4: The attrition of 20 test MPS cases 18 Base-lining Costs of the Lighthouse 18 4. Learning from Mobilisation & Early Implementation 20 Designing Child House: the vision 20 Designing Child House: the evidence & future success 20 Need for governance & Internal Communications 21 The benefits of multi-agency partnership 21 External Communications: the importance of engaging external stakeholders 22 Technical set-up 23 The future 23 5. Discussion 23 References 25 Appendices 28 Appendix 1: Lighthouse partner relationships 28 Appendix 2: Disclosures of Child Sexual Abuse requiring referral to the Child House 29 Appendix 3: Child House Logic Model 31 Appendix 4: Theory of change per outcome - why should there be an impact? 32 Appendix 5: E&I Research Product Timeline 39 Appendix 6: Criminal Justice Attrition Tree 40 3 Executive Summary The Lighthouse, London’s Child House, opened in October 2018 as part of a two-year pilot. Bringing together a range of organisations under one roof, the Lighthouse intends to be a child friendly, multidisciplinary service for victims1 of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden, it will replace the existing services to date and serve the five surrounding North Central London boroughs of Barnet, Camden, Enfield, Haringey and Islington. The Lighthouse intends to offer more enhanced features than previous services, with the foremost aim to be focused on the child. The Evidence and Insight (E&I) Unit are MOPAC’s in-house social research and analytical team and were commissioned to evaluate the Lighthouse. The E&I evaluation will focus on four distinct areas for analysis, these are; a performance review; a process evaluation; impact evaluation and an economic evaluation. This will be done using a variety of methods, balancing qualitative information from staff, stakeholder or clients alongside analysis of routine performance data as well as more complex quantitative analysis exploring the impact and costs. More details about obtaining a control group to assess impact will be included in the next evaluation report. This variety is important given the absence in impact and cost analysis in previous similar evaluations. The ability to successfully complete each element will depend on data quality, which will be reviewed throughout the life of the research. What does success look like for the lighthouse? The report documents the primary outcomes for the Lighthouse, these broadly cover: • Enhanced referral pathways into and out of the Lighthouse • Enhance CYP, family and carer experience of support received post disclosure • Enhance CYP experience of the criminal justice process post disclosure • Enhance mental health and well-being outcomes for CYP • Enhance professional awareness, competence and confidence • Increased likelihood of charge or conviction for those cases within the Lighthouse • Enhance partnership working • Providing CSA victims care and support to reduce the long-term impact of victimisation Early findings to date Base-lining Criminal Justice There is a lack of wider research pertaining to the attrition of sexual offences against children and no easy manner to obtain this data. This is problematic given the status as a key outcome. As a test of feasibility in obtaining such data, the research team hand coded 20 CSA/E cases from the Metropolitan Police Service (across 120 variables). Although labour intensive, the benefit for the evaluation is thought to warrant the resource required. Base-lining Cost RedQuadrant were commissioned to produce a cost-benefit model for the Lighthouse project – base-lining key capital costs (the one-off expenditures to set up the programme) and operating costs (the on-going costs associated with service delivery). In terms of the benefits RedQuadrant highlight these are wide-ranging. Outcomes were grouped from the literature 1 Referred to as victims throughout the remainder of the report 4 into three categories of social value: well-being to the client; useful savings from public sector spend; and additional public sector spend on essential activity. Estimates of savings for each theme are ‘best guesses’ based on many unknown assumptions and will therefore need to be substituted for actual costs and the Lighthouse (and counterfactual) throughput and prevalence data once available. More detailed figures will be provided in the next evaluation report. Process: Summary of Learning from Mobilisation & Early Implementation The learning presented herein is taken from a variety of sources, all from relatively early on in the Lighthouse service delivery. Process learning is valuable in reflecting the entire journey of a programme - from initial design and set up through implementation. Overall, in terms of results so far; all staff, partners and stakeholders are positive in terms of the vision, design, set-up, governance, and partnership approach of the Lighthouse. Certain aspects, such as partnership working and colocation have been especially praised. Indeed, a fundamental element of the Lighthouse is the multi-agency working, enabling different organisations to practically work towards a common goal. Given the level of underlying complexity and size of the model - this is something to be celebrated and is an important enabler in the ongoing effective implementation of the initiative. Stakeholders recognise the Lighthouse is evidence based, referring to it as an amalgamation of the U.S. Child Advocacy Centres and the Scandinavia Barnahus model. Stakeholders did identify some differences between Barnahus and the London Lighthouse design with respect to commissioning and criminal justice aspects. Given what is known about the importance of programme integrity, this will be important to monitor within the evaluation, as a move away from the original model has the potential to influence the outcomes. There were some challenges identified, such as the importance of linking with the judiciary; the use of Live link and section 28; Information Technology; sustainability of the service in the future - although it is hoped these issues will be addressed by the Lighthouse governance structure. This is the first in a series of MOPAC E&I Lighthouse evaluation reports to be released throughout the next two years, enabling learning both internally as a catalyst for improvement, and externally to advance the evidence base. 5 1. Background: Setting the Scene The Lighthouse, London’s Child House opened in October 2018 as part of a two-year pilot. Bringing together a range of services under one roof, the Lighthouse intends to be a child friendly, multidisciplinary service for victims and survivors 2 of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden it will replace the existing services to date and serve the five surrounding North Central London boroughs of Barnet; Camden; Enfield; Haringey; and Islington. The story so far… Across England and Wales, in recent years there has been an increase in reported cases of CSA/E. In 2016/17 there were 63,663 sexual offences against children3 recorded by the police in the UK4 – the highest recorded number in the past decade – with a 16% increase from the last year (54,898 sexual offences recorded in 2015/2016)5 and a 196% increase from 5 years ago (21,493 sexual offences recorded in 2011/12)6. Turning to London, it is difficult to obtain definitive figures to represent the number of children who report to health services because of CSA7]. However, in terms of comparable police-recorded CSA in London, there has been a 67% increase comparing financial Year 2011/12 to financial year 17/18 (from 2,208 to 3,685 cases). It is unknown how much of these increases are due to an escalation in prevalence; improved police recording; or a reflection of an increased willingness to report abuse following high profile cases in the media. Regardless, the critical aspect is the substantial increase in demand. Given the prevalence of CSA/E, it is increasingly being recognised as a public health problem8, impacting substantially on long-term outcomes including physical health 9 . The road to recovery following CSA/E is complex and requires specialist care and tailor-made support. However, this is not the routine provision - services are often delivered from a variety of agencies including, but not limited to: the NHS; police; and other criminal justice agencies; children’s social care; and agencies from the voluntary and community sector; making care disjointed. Concerns were raised in 2015 by the Office of the Children’s Commissioner, as only around one in eight victims of sexual abuse come to the notice of statutory authorities10. In response, MOPAC and NHS England (London region) commissioned the “Review of Child Sexual Assault Pathway for London”, to map the various pathways for children and young people following CSA. Findings from the Goddard Review highlighted variation in the available services across all London boroughs and gaps in medical provision, emotional support and the prosecution process. The report made recommendations advocating the need to establish better overall 2 Referred to as victims throughout the remainder of the report 3 Due to data availability, offences for England and Northern Ireland are against children under 18 and offences for Scotland and Wales are against children under 16 4 Bentley, et al., 2018 5 Bentley, H. et al (2017) 6 Harker, L. et al (2013) 7 Goddard, Harewood, & Brennan, 2015 8 Brown, O’Donnell & Erooga, 2011 9 Bellis, Lowey, Leckenby, Hughes, & Harrison, 2014 10 Children’s Commissioner for England, 2015 6 services for Children and Young People (CYP) who have experienced CSE/A11. A direct result of the review was the introduction of Child Sexual Abuse Hubs (CSA Hubs) across London. Designed and built on good practice, they created virtual teams of CSA/E experts in local areas. In 2016, two NHS sector Hubs were established, the first located in the North Central NHS sector and a second established in South West London. These provided medical and short-term emotional support for victims of CSA/E and an integrated response for the families, but the police and social care were not directly involved. However, the Goddard Review identified a better approach would be to introduce ‘Child Houses’ to London12]. Child House - a new approach Child House is a term that groups several similar models adopted internationally together. Initially developed in the US in the 1980s, the Child Advocacy Centre (CAC) model was proposed as a solution to many problems associated with standard responses to CSA, including: lack of therapeutic services; low conviction rates; traumatic investigation processes; and inter-agency conflicts13. Research into CACs in the US found positive results, particularly around reducing the trauma experienced by victims of CSA and improving levels of satisfaction with the overall service for both children and parents14. Barnahus (Children’s House), a model used in Iceland since 1998, was inspired by the CACs (although there are some differences in service delivery between CACs and a ‘Child House’ model). Barnahus is a child-friendly, interdisciplinary and multi-agency centre where different professionals work under one roof in investigating suspected CSA cases and providing appropriate support for victims15. The main components of the Barnahus model are: • A home-like setting with all services delivered under one roof; • Helping victims disclose abuse through exploratory interviewing, conducted by child psychologists; • Use the least possible number of interviews conducted by child-expert staff; • Improved evidence through the reduced need for children to testify in court; and • Guaranteed and rapid access to therapy for abused children. Compared to before and after its inception, the initial Icelandic Barnahus model was considered to yield positive results across many outcomes, such as: improved partnership working between police and social services; improved therapeutic outcomes for children and their families; improvements in children’s and families’ experiences of the criminal justice process; and improvements in the quality of investigations; trebling the number of perpetrators charged; and doubled the number of convictions 16. As a result, the Barnahus model has since been adopted in several other countries such as Sweden, Norway and Denmark. The EU Promise project brings together research across European pilot countries and provides standards, learning and best practice17. However, although findings across these later adaptions appear to be relatively positive, most evaluations are based in the US and on 11 Goddard, Harewood, & Brennan, 2015 12 Goddard, Harewood, & Brennan, 2015 13 Herbert and Bromfield, 2016 14 Elmquist et al., 2015 15 Children’s Commissioner, 2016 16 Children’s Commissioner 2016 17 http://www.childcentre.info/promise/pilot-country-updates/ 7 the CAC model, rather than on Barnahus (see appendix 4 for more research). In addition, previous evaluations primarily concentrate on the underlying processes, with fewer robustly assessing impact or economics. This is a considerable gap in research learning and one that the current evaluation seeks to address. The Lighthouse - a Child House for London Following the Goddard Review, NHS England (London region) approached MOPAC for support with a bid to Home Office Police Innovation Fund to pilot two Child Houses in London. The joint bid was successful and funding was awarded in April 2016. The Child Houses subsequently formed a commitment within the new Mayor’s Police and Crime Plan (2017 – 2021) 18, as part of keeping London safe for children and young people. The Mayor and Home Secretary announced the plans for two Child Houses in September 2016, originally anticipated to open in April 2017. When deciding on service location there were several criteria used, one of the key aspects being a suitable property to house multi-agency services. Other considerations were the state of readiness of the NHS Sector with regards to seeing Child Abuse as a high priority, and a wider prevalence of CSA in those areas. Based upon these, a decision was made to develop services in the North Central London NHS Sector and the South West. However, it became apparent there was insufficient funding to run across both sites. A decision was made to proceed with one location, making it possible to include the enhanced staffing levels; an extension of the service to 18-25-year olds with additional needs; extended opening hours; and consideration of accepting neighbouring sector/out of sector referrals. In deciding the one location - again, criteria were employed (i.e., need; strategic alignment of the wider health community; existing clinical leadership; demand projections; premises availability) and it was decided that North Central London was best place to proceed with the pilot. Although based in a Camden property, the service would take referrals from Barnet; Camden; Enfield; Haringey; and Islington. This geography would form a coherent area and these boroughs present a substantial demand – with 2016/17 MPS data indicating a total of 683 victims of sexual offences aged 17 or under. Borough level data shows each area has seen a steady increase in MPS CSA/E figures, with Barnet the largest increase of 61% in yearly totals. Enfield has the highest levels overall and Islington and Camden have consistently the lowest levels of CSA among the North Central boroughs19. In February 2018, the commissioned contract was awarded to the University College London Hospitals NHS Foundations Trust (UCLH) and their sub-providers, brought in to deliver specialist elements of the service, namely - the Tavistock and Portman NHS Foundation Trust; the National Society for the Prevention of Cruelty to Children (NSPCC). Delivery of the two- year pilot was to follow a six-month mobilisation period. One of the most unique aspects for the programme is the complex partnership arrangements required to fund and deliver the enhanced service across multiple London boroughs (see appendix 1). This includes embedding key organisations physically into the premises, such as the two dedicated Metropolitan Police Service (MPS) officers. Although recognised as logistically challenging, this approach is paramount to providing a smooth and efficient service to the CYP and one of the unique elements of the model. 18 https://www.london.gov.uk/mopac-publications/police-and-crime-plan-2017-2021 19 Data downloaded from MPS CRIS September 2018 by MOPAC E&I. 8 The Lighthouse will provide enhanced support to CYP aged between 0 – 17 years old (or those between 18-25 years of age with learning delay or disability for whom a child or young person- oriented service appears more suitable), as well as non-offending parents/carers/family for up to two years. It intends to offer a joined-up approach where, if required, you can get access to all medical; practical; social care; police; and therapeutic support ‘under one roof’. Services will be offered at the Child House during extended opening hours (Monday to Saturday 10:00 to 20:00 and by outreach Sunday 10:00 to 13:00) and CYP Havens will continue to provide the acute/Forensic Medical Exam (FME) service. The Lighthouse builds on the CSA Hubs, but offers enhanced service features which include the elements in Figure 1: There have been several discussions in terms of estimating demand to the Lighthouse. Based on the throughput of cases into the pre-existing CSA Hub services, the lead provider has been commissioned to deliver a total of 544 CYP accessing the Lighthouse per year. However, estimates from police data are higher in terms of those who may disclose/be eligible for Achieving Best Evidence (ABE) interviews (around 700 CYP). As outlined later, one aspect of the evaluation will be to monitor such numbers. Figure 1: Enhanced aspects of the Lighthouse In addition, the Lighthouse has extended the eligibility criteria for access to the service following a ‘disclosure’. It is recognised disclosing abuse is a difficult and often a negative experience, with a child rarely using a straightforward process of telling someone they have been abused – in fact, research has indicated on average it can take 7 years for CYP to disclose sexual abuse, with the earlier it started taking the longest to disclose20. As a result, many CYP may present to medical or support services without having verbally disclosed the offence – something which can be problematic when accessing pre-trial therapeutic support 21 . Nevertheless, it was considered vital to include children where there is a ‘significant suspicion of CSA/E’, where practitioners conclude it is highly likely abuse has occurred and there are either behaviours or physical symptoms to support this, or a history of risk or harm to the child or a sibling (see appendix 2 for eligibility criteria). 20 Allnock, D. and Miller, P., 2013 21 Warrington, C., 2015 9 2. The Lighthouse Evaluation The Evidence and Insight (E&I) Unit is MOPAC’s in-house social research and analytical team which has been commissioned to undertake an evaluation of the Lighthouse. The two-year evaluation will cover the processes of the Lighthouse (from design through implementation), monitor routine performance, as well as seek to explore robust impact and cost benefit. This report seeks to outline the broad approach to the evaluation, measuring success and the methodologies used, as well as present some early results from fieldwork. This has been split into base-lining the service (criminal justice and cost), and key themes from mobilisation and early implementation (including: design; need for clear governance & internal communication; external communications: the importance of engaging external stakeholders; making a multi-agency partnership work; technical set-up; and the future). What does success for the Lighthouse look like? A key first step, prior to commissioning the service to a provider, was to develop the core outcomes for the service. To achieve this, E&I and key staff from the Lighthouse commissioning team (MOPAC & NHS England (London region)), developed the underlying logic of the service (see appendix 3 completed Logic Model). The main emphasis was for the service to be focused on the child – reducing the risk of re-traumatisation; ensuring timely access to medical and therapeutic support; supporting CYP and the non-offending members of their families. To achieve this, overall aims cut across delivery partners. Outcomes were split into primary and longer term, each with a theory of change and evidence base for why there is a potential to measure impact (see appendix 4 for detailed research findings). See Table 1 for an overview of the outcomes and underpinning rationale. The evaluation will look to monitor as many of these key outcomes as the data allows. The evaluation has been designed across 2-years utilising a mixed methodology approach – balancing qualitative context from staff, stakeholder or client feedback, particularly in the shorter-term, with the ‘harder’ performance figures, indicating how the service is running on a day-today basis. 10 Table 1: Overview of Lighthouse Objectives Lighthouse Objective Is this informed by evidence? Short/ Medium/ Long Measurable by… Enhanced referral pathways into and out of the Lighthouse: Better identification of CSA/E by local professionals, better knowledge of referral pathways and speedier referrals into the Lighthouse. One single referral pathway for all five boroughs - stops multiple agencies acting as gate keepers. Greater standardisation of service for those engaging, and where needed, more appropriate referrals onwards. ✓ These outcomes are consistently found in the evidence base (see appendix 4). Short term Performance data around the number and timing of referrals. An increase in professionals self-reported confidence/knowledge of referral pathways. Appropriate referrals to other services matched to CYP needs. Enhance CYP, family and carer experience of support received post disclosure: Longer availability of support (2 years) and streamlined services should provide high levels of satisfaction with the service. ✓ Satisfaction often found to be high, but usually no difference to comparison groups. Medium term Self-reported satisfaction from CYP and families/carers assessed throughout service. As there is usually no difference to the comparison groups, measures should look to assess how unique service elements (e.g. extended support/opening hours) are viewed. CYP may set goal-based outcomes for themselves during therapy – their self- reported achievement can be assessed. Enhance CYP experience of the criminal justice process post disclosure: Advocate support; not being required to attend a police station to be interviewed; and potentially not being required to attend court due to Live Link, should reduce stress for the CYP and their caregivers. Satisfaction with the process (rather than court outcome) should be high. ✓ Satisfaction found to be higher than comparison groups. Medium term & Long term Self-reported satisfaction from CYP and families/carers can be assessed throughout service provision – can start with initial police investigation in the medium term, whilst court process/outcome maybe longer term. Assess unique service elements against comparison group (see methodology – impact). Enhance mental health and well-being outcomes for CYP: Easier and quicker access to longer term support should increase satisfaction and specific wellbeing/mental health outcomes. ✓ Evidence found (see appendix 4). Medium term & Long term Self-reported wellbeing can be assessed throughout service provision. Psychometric measures (e.g. Trauma Symptom Checklist for Children (TSCC)) can be compared pre/post and potentially to a comparison group. Enhance professional awareness, competence and confidence in working with CSA/CSE: Raising the local profile of the Lighthouse via communication and engagement to ensure knowledge of, and confidence in, referral pathways. ✓ Evidence found (see appendix 4). Short term & Medium term The number of referrals. An increase in professional’s self-reported confidence/knowledge of referral pathways. 11 Increased likelihood for CYP who received a Lighthouse service to have cases charged by CPS and; Increased likelihood for CYP who received a Lighthouse service to have their case end in conviction: More referrals into the service; advocate support leading to better engagement; better evidence collected by ABE interviews; and better cross examination via Live Link should increase the likelihood of reaching charging thresholds and receiving convictions. ✓ Evidence is mixed and maybe linked to which part of the process is assessed - outcomes earlier in the CJ process maybe more favourable than later in the process, like criminal convictions (see appendix 4). Medium term & Long term Tracking case attrition and reasons for drop outs can start with initial police investigation in the medium term and then move to court process/outcome in the longer term. Number of cases charged. Number of cases convicted. Sentences received. Enhance partnership working: Multi-agencies working in collaboration under one roof will make communication better and data sharing issues less problematic. ✓ Evidence found (see appendix 4). Short term & Medium term Self-reported staff satisfaction assessed throughout provision. Case studies can assess how organisations have come together to make decisions regarding care. Providing CSA victims care and support to reduce the long-term impact of victimisation: Impact on well-being after the service.  Little evidence found in terms of child and family outcomes Long term Limited predications can be made from case outcomes on future benefits based on literature findings. May need longer term tracking outside of evaluation scope. Organisations are committed to being victim focused in their support of CSA victims. Organisations not directly involved in service delivery supporting the Lighthouse e.g. courts open to using Live Link and S.28. Changes in current practice to be more victim focused. ✓ Evidence is mixed. Most research focuses on criminal justice – with significant differences found earlier in the process. Some studies found practices have become incorporated more broadly, but what elements have an effect are unclear. Medium term & Long term Self-reported staff accounts of wider partnership involvement can be assessed throughout service provision. Case studies can assess how organisations have come together to make decisions regarding care. May need longer term tracking outside of evaluation scope. Methodology The Lighthouse evaluation focuses on four distinct areas. These are: performance monitoring; process; impact and economic analysis (see below diagram). The ability to successfully complete each element will depend on the quality and quantity of data and will be reviewed throughout the life of the research, as it is subject to change. MOPAC E&I follow a pragmatic ‘action research’ approach, feeding key findings back to the service and commissioning team in a timely manner, so improvements may occur. The evaluation aims to holistically assess the programme and will therefore draw on a range of quantitative and qualitative methods. 12 Figure 2: E&I Evaluation Process The performance monitoring aspect will use data and management information captured during the everyday running of the Lighthouse to track actual service delivery. For example, how many clients are using the service and when; what needs they present with; what services they receive and for how long. For this purpose, a bespoke Case Management System (CMS) has been set up to store ‘Lighthouse data’, inputted from delivery partners – namely the lead NHS provider UCLH and sub-providers brought in to deliver specialist elements of the service. The process aspect will explore the underlying learning, good practice and challenges across the entire two-year Lighthouse pilot (i.e. from design through partnership working to implementation and end user experience). Drawing upon a range of sources, learning to date has been taken from those illustrated below in figure 3. Figure 3: Evaluation progress 13 Over the course of the two-year evaluation, the process aspect will be drawn from a range of methodologies (i.e., interviews, focus groups and surveys) to capture in-depth information from across stakeholders, practitioners and those within the service. The early findings reported here are based upon: • 18 face-to-face interviews in April/May 2018 conducted with key stakeholders and board members (undertaken by RedQuadrant alongside E&I). These interviews focused on the history of the project; procurement; commissioning; design specification; and initial implementation (see section 4. Learning from Mobilisation & Early Implementation). Learning was shared to inform the separate RedQuadrant and E&I products, but crossed referenced where applicable. • An online survey designed by E&I distributed in September 2018 to professional stakeholders (e.g. police, charity workers, mental health practitioners, etc) who work within the five Lighthouse boroughs and may come across CYP who have experienced CSA/E. A total of 54 people responded from a range of occupations (Police Officers (39%, n=21), Nurses (15%, n=8), Victim Charity workers (15%, n=8), Mental Health practitioner (11%, n=6) and Other (20%, n=11)) (see section 4. Learning from Mobilisation & Early Implementation). • To supplement delivery stakeholder interviews, all programme board members received a short online survey to capture opinions around the design and initial implementation of the Lighthouse. In total, 13 people responded22. • In October 2018, all Lighthouse staff attended an induction to the service and training. Afterwards, staff were asked to participate in a focus group to explore their views of the Lighthouse; how ready they felt to open the doors to the public; training needs; and potential challenges going forward. In total, 13 staff members attended from a potential 27 practitioners. • Finally, in November 2018 the Home Office produced a Health Check for the Child House programme, conducting semi-structured discussions with two key members of the MOPAC Programme Team, focusing on the sustainability requirements of the Lighthouse. Findings are highlighted where applicable. The impact analysis aims to examine if the Lighthouse has achieved a demonstrable impact across the key outcomes compared to a counterfactual (i.e., a matched group of individuals who do not receive the Lighthouse services). Estimating impact is always the most challenging aspect of evaluative research. The gold standard Randomised Control Trial was not possible for ethical reasons. The evaluation is therefore seeking to utilise the next most robust approach of a quasi-experimental design. This is where a comparison group is statistically generated and used to compare to the key outcomes. To gain access to relevant data, a control group will most likely be sourced from a ‘service that exist in London, had the Lighthouse not been established’, namely a CSA hub. Future reports will detail the development of the counterfactual and the associated validity checks. One aspect of the impact evaluation aims to track criminal justice cases. To explore the feasibility of this approach a small dip sample of 20 cases (10 identified as CSA and 10 as CSE) were randomly selected from the MPS Crime Reporting Information System (CRIS) to test the time-consuming methodology and explore case attrition. All cases were reported in April 22 Given the size of the research cohort (e.g. the number of respondents to the stakeholder survey/offender surveys), caution should be used when considering the results. Response base size is provided; however, this varies as not all respondents answered every question. 14 2016, so enough time had passed to collect meaningful information on case outcomes. Cases were coded using an extensive framework, to capture specific details about the victim; offender; aspects of the offence; police investigation; and case outcome. Key attrition points were coded e.g., suspects identified; arrests made; charges; and whether the cases reached court (see Learning from Base-lining the Service: Criminal Justice for early learning). Although time consuming, this approach proved to be viable and able to provide important learning on CJ progression. Future reports will use this method to track cases receiving a service from the Lighthouse; who report to the police; and who consent to their data being used for the evaluation. Simultaneously, many cases from the counterfactual will also be tracked. Full case coding will begin approximately nine months after the launch of the Lighthouse, to allow for sufficient time for cases to progress. Ideally around 300 cases will be identified for the control group, from the same timeframes as the Lighthouse cases. A qualitative assessment on the CJ process is also planned, in the form of interviews or focus groups with police officers who both work within the Lighthouse areas or those from across the counterfactual areas to explore and compare their experiences. It is anticipated (depending on data access, quantity and quality), any early effect of the Lighthouse on the initial stages of the CJ process can be reported in the E&I report currently scheduled for January 2020. The latter stages of the CJS (e.g. convictions and sentencing) will not be possible until the end of pilot. The economic analysis will seek to explore issues such as ‘Does the Lighthouse provide value for money?’ and ‘What are the public value benefits and what are the fiscal benefits?’. RedQuadrant have been commissioned to produce a cost-calculator grounded in learning from the wider evidence base. This report summaries the methodology used for base-lining service cost and testing against the benefits identified through the literature (see Learning from Base-lining the Service - Cost). E&I will use these principles to perform cost benefit once there is confirmed set up and running costs and actual service data is available on throughput and prevalence of relevant criteria (e.g., the number of CYP presenting with mental health needs and so on). In addition, to the overall E&I evaluation, the commissioning team procured RedQuadrant to conduct a Learning Strategy and toolkit to identify how best to sustain the Lighthouse at the end of the pilot and present key learning for setting up a programme elsewhere. The initial work conducted by RedQuadrant details the funding, commissioning and procurement of the service, to understand the learning from the initial two-year design phase of the project. As this will be included in their product, it will not be covered by this report. Ethics and oversight In line with Data Protection Act (DPA) 2018 and General Data Protection Regulation (GDPR) (2018), E&I have produced a Data Privacy Impact Assessment (DPIA) Ref: [MOPAC/2018/CHEVALDPIA/001] appended to the programme documentation, to outline the bespoke risks and mitigation for them, which will be reviewed throughout the lifetime of the project. 15 In addition, an independent research advisory board was developed and will meet quarterly (the first in January 2019). It consists of both subject matter and methodological experts, to act as critical guiding experts to the evaluation team. Whilst collating direct feedback from service users is a desired component of the research, E&I are aware of the additional risks of re-traumatisation and will work with all partners; the academic advisory group; and any necessary ethical bodies (e.g. Health Research Authority) to explore how to do this in a safe manner. This will include only undertaking direct contact with CYP through specialist professionals, who can ensure the correct support and safeguarding processes are in place. Research timeline Over the course of the 2-year pilot there are four (including this one), E&I evaluation reports planned (see appendix 5 for the E&I research products timeline). Currently there is set to be an interim report delivered in May 2019, after 6 months of data collection and delivered in time to inform the sustainability work around funding cycles. This report will include the first look at performance monitoring; additional interview findings regarding continuation of care; and cost-base-lining using actual rather than predicted data. The year 1 interim report is due early 2020 and, along with refreshed performance monitoring and qualitative findings from surveys; interviews; and focus groups, will include a detailed account of the development of the quasi-experimental approach (the counterfactual) and will take a first look at the effect on key outcomes. Criminal justice analysis will focus mainly on the initial police stages of an investigation, but along with health and well-being outcomes if possible compared to the counterfactual. The final report will include a summary of all findings to date and is due mid-2021. It will look to include additional qualitative/case study analysis; analysis on the latter end of the CJS (such as convictions and sentencing); and where possible comparisons of all key outcomes to a control group and detailed cost benefit analysis. Findings from the evaluation will also be routinely fed into the commissioning and programme teams and will update partners at the official Partnership Oversight Board and any other relevant meetings. Reporting timelines are provided as a guide (see appendix 5) and are subject to change. 3. Learning from Base-lining the Service The remainder of the report will discuss early learning to date resulting from initial fieldwork drawn from the above methodology. It should be noted such results are interim and should be seen as indicative, as they are subject to change as the service progresses. Base-lining Criminal Justice As outlined, this section presents early insights from analysis into criminal justice baselines. Assessing case attrition is important, as achieving increased CPS charges and court convictions are two key outcomes for the Lighthouse. Attrition refers to the process whereby cases drop out of the criminal justice system, at one of several possible exit points before an outcome at court (see appendix 6 attrition tree). The journey through the CJS is often lengthy and complex and attrition can occur for several reasons, including crime occurring outside of the jurisdiction of the police service in which it 16 was reported; the police or Crown Prosecution Service (CPS) deciding there is not enough evidence to proceed; or the victim withdrawing from the investigation. Before turning to early results from the feasibility work on base-lining criminal justice, it is useful to present some learning from the wider literature. Unfortunately, there is little UK research on the attrition of sexual offences against children. Although it is thought cases of CSA are less likely than adult cases to be assigned ‘no further action’ by the police23 and more likely than adult sexual offences to result in a conviction24, one paper acknowledges overall high attrition rates leads to a minority progressing past a police sanction detection to prosecution – with government statistics at the time reporting just a 30% (34% for sexual activity with minors) sanction detection rate25. Recent figures indicate only 16% of the 54,000 sexual offences recorded against children by 43 police forces in England and Wales (between October 2015 and September 2016), resulted in a charge; summons; community resolution; or caution against the perpetrator, although rates differ greatly between forces, ranging from 4% to over 35%. For those offences not resulting in action against a perpetrator, the most common reason was the victim did not support the police in the investigation, which happened in 27% of cases26. Australian case progression and attrition research found reasons for not progressing to trial included the offence not being reported to police; parents wanted to protect their children, perpetrator or other family members; evidence was not strong enough; the child was too young; the offender threatened the family; or the child was too distressed27. It is a Lighthouse aspiration for the service to increase CPS and court charges through supporting the CYP and caregivers to report to the police; maintaining engagement using an advocate; obtaining better evidence collected by psychologist led ABE interviews; and better cross examination via Live Link. It is unknown how the latter two procedural aspirations will affect CPS/court outcomes as research exploring if they account for a low prosecution rate are mixed. Some studies suggest poor quality interviews28 contribute, whereas others have found the quality of interviews within a Swedish Child Advocacy Centre did not have any predictive value on the prosecution rate29. However, it is acknowledged the importance of gathering robust evidence (including victim account and where possible forensics), to ensure cases pass the threshold to enter criminal proceedings. In the United States (US) acceptance for prosecution is one of the main barriers in the progression of sexual abuse cases. A study examining the relationship of CSA case characteristics and acceptance for prosecution found across the 431 cases referred the background characteristics of the perpetrator and victim; severity of abuse; and nature of available evidence were all significantly related to acceptance for prosecution30. When the impact of different types and amounts of evidence have been explored, it was found across 329 US CSA cases four types of evidence uniquely predicted whether charges were filed following an investigation: a victim disclosure; a corroborating witness; an offender 23 Hohl and Stanko (2015) 24 Allnock, 2015 25 MoJ, Home Office and ONS 2013 cited from Warrington, C. 2015 26 The Children’s Society (2018) 27 Parkinson et al., 2002 28 Diesen & Diesen 2009 29 Hagborg et al. 2012 30 Cross et al. 1994 17 confession; or an additional report against the offender. Cases with a corroborating witness – even when lacking stronger levels of evidence such as a confession or physical evidence – were nearly twice as likely to be charged31. It will be a key aim of the evaluation to understand the nature of the evidence presented to determine what aspects maybe effecting case attrition. As outlined, to better understand the nature of the CSA/E cases receiving a Lighthouse service, a small number of current MPS cases were identified, to test the time-consuming coding methodology. From the 10 CSA cases (of which 5 were rape cases), and 10 CSE cases included in the analysis it was found attrition for CSE happened almost immediately. All CSE reports were classified as ‘non-crime’ - that is, they were created because of safeguarding or concerns about a potential victim. In most cases there was not a specific CSE offence, instead an event or situation led authorities to have concerns, such as the victim running away from home; a notable behaviour change; or associations with risky older individuals. However, of the 10 non-crime CSE reports, 5 also had related reports whereby CSA against the same victim was being investigated, suggesting the investigation of those initial concerns led to the detection of a CSA offence. See figure 4 for the overview. In most cases (n=19 of 20), reporting to the police was done by a third party/witness, predominantly a parental phone call to police (n=11). A suspect was identified in 12 cases – the majority (n=9) were for CSA cases. In eight of the CSA cases the victim was also able to identify the suspect. Of the 10 CSA cases, 3 were No Further Actioned (NFA’d) by the police, but 6 were submitted to the CPS and 5 were subsequently referred to court. Whilst this sample is too small to suggest findings are indicative of all CSA/CSE cases, their coding has developed the methodology ready to track the Lighthouse cases and has been insightful as to how the police record CSE – in that the associated CSA cases should be tracked instead. Although the approach is labour intensive (around 120 variables are hand coded from the case files and a sub-set is double coded for validity checking), the value added to the evaluation is thought to warrant the resource required. Key health, wellbeing and criminal justice elements can be combined to potentially understand their influence on the progression of cases through the CJS. 31 Walsh et al. 2008 18 Figure 4: The attrition of 20 test MPS cases Base-lining costs of the Lighthouse. According to the Treasury’s guidance for evaluation32: ‘a reliable impact evaluation might be able to demonstrate and quantify the outcomes generated by a policy, but will not on its own be able to show whether those outcomes justified that policy. Economic evaluation can consider such issues, including whether the costs of the policy have been outweighed by the benefits’, adding ‘economic approaches value inputs and outcomes in quite particular ways, and it is crucial that the needs of any economic evaluation are considered at the design stage.’. Economic analysis has been noticeably absent from previous evaluations of child advocacy centres, something the current evaluation is seeking to address, with MOPAC commissioning RedQuadrant to produce a ‘difference in difference’ cost-benefit model for the Lighthouse project. The benefits identified (provided below) are those believed to have the most significant impact; to be measurable; and to be attributable to the interventions delivered via the Lighthouse. RedQuadrant also designed a cost-analysis process to be completed by E&I throughout the life of the project, once Lighthouse and counterfactual throughput data is available. Base-lining a project looks to identify key capital costs (the one-off expenditures to set up the program) and the operating costs (the on-going costs associated with service delivery), so if a model were to be replicated elsewhere there is an understanding of all associated expenditure. In terms of the benefits, RedQuadrant highlighted a wide-ranging selection – covering the child; family members; involved organisations; or wider society. Some will be evident in the short term, whilst others may not be apparent for many years. They decided to group outcomes identified from the literature into three categories of social value: well-being to the client; useful savings from public sector spend; and additional public sector spend on essential activity. Within these categories are thematic areas which have been considered, these are: 32 HM Treasury 2011 p20 19 • Health (sexual health, physical health and substance misuse); • Well-being (from the perspective of the NHS and local authorities, the individual child or young person, and the family); • Children’s services: (possible impact on child protection action and on the need for any additional school support); • Employment: (possible impact of the Lighthouse on loss of earnings and take-up of benefits otherwise resulting from CASE); • Criminal justice: (costs of crime and enforcement action against alleged perpetrators, and the impact on possible criminal activity committed by victims of CSAE during adulthood because of their abuse. The possible reduction in the number of ‘cracked trials’ is also estimated); and • System effects: (improved productivity because of better co-ordination of multi-agency services at the Lighthouse). Depending on the availability of prevalence and cost information in the literature, RedQuadrant used a methodology to estimate the social value for each of the thematic areas based on: • the estimated unit cost of treating an issue multiplied by an estimated proportion of people who present with an issue; or • the estimated reduction the Lighthouse is thought to have on an issue multiplied by societies average spend on the issue. For example, under the thematic areas of ‘system effects’, RedQuadrant explored how the introduction of the Lighthouse may improve productivity, stipulated in the literature to be because of better co-ordination between agencies. A system cost can be calculated from combining court; police; CPS; local authority; and NHS costs taken from relevant NSPCC and NAO publications. The literature then provides an assumption of improvement in productivity following better co-ordination (taken from 2012 Nat Cen study). A financial benefit as a result of the introduction of Lighthouse can then be calculated per client. Although this methodology provides estimates of savings for each thematic area, they are ‘best guesses’ based on many unknown assumptions and will therefore need to be substituted for confirmed costs and the Lighthouse (and counterfactual) throughput and prevalence data once available. More detailed figures will be provided in the next evaluation report. The importance of using a robust counterfactual to compare change against is again stressed, as without it impact cannot be tied to the services provided by the Lighthouse. For example, without a counterfactual it will be unknown if any changes in health and well-being are actually due to the quality of the therapeutic intervention provided or down to chance. In addition, the cost: benefit ratio will depend significantly on how many people use the service. The current calculations undertaken by RedQuadrant are based on the higher estimate of 700 C&YP per year using the service. If this is an overestimate, the cost per child will be significantly higher than anticipated. 20 4. Learning from mobilisation & early implementation This section explores learning from across a range of sources including internal MOPAC E&I primary research, MOPAC commissioned research from RedQuadrant and the Home Office Assurance Review. Although products focus on different aspects of the programme, including commissioning; procurement, early mobilisation; initial implantation; and learning for sustainability, there are common themes present. Key thematic areas of learning were identified, which are: design; need for clear governance & internal communication; external communications: the importance of engaging external stakeholders; making a multi-agency partnership work; technical set-up; and the future. Designing the Lighthouse: the vision It was a positive that from across the sources there was a consensus around the necessity for Lighthouse due to the amount of unmet need; limitations in current provision to victims and delays in the health, social care and the criminal justice system. Stakeholders felt the Lighthouse will not only “meet a service gap” but will provide better support for CYP, as it has been designed with them at the centre of vision, wrapping services around them. This was considered to result in better identification of abuse and the ability to provide an effective, consistent and trusted response, throughout the process and across wider partners/agencies. Some responses went as far as hoping the Lighthouse plays a role around community change and education. Stakeholders emphasised a key aspect of the vision which was bringing currently fragmented service provision under one roof, reducing the need for travel, integrating the response and enabling the CYP to access support and provide evidence of abuse hopefully without re-traumatisation. Stakeholders were also positive towards the strong leadership “providing good scrutiny and direction”. It was felt dedicated involvement and enthusiasm, especially from clinical consultants who played a leading role in defining the mobilisation stage, demonstrated their commitment and passion to the subject area. The Home Office assurance review identified the shared common goals of key stakeholders enabled the vision to be realised and was a key success for the programme. Designing the Lighthouse: the evidence & future success The Lighthouse design was evidence based, something all sources agreed with. Stakeholders referred to it as an amalgamation of the Child Advocacy Centres in the United States and the Barnahus model in Scandinavia. Whilst some differences between the Barnahus model and the subsequent London Lighthouse design were identified - this was expected given the London setting. For example, in Iceland, the Barnahus model has an inquisitorial justice system but delivery stakeholders were unable to see how the London model could completely align to the UK adversarial justice system or how to positively change or challenge court proceedings. As the evaluation progresses, a core aspect of the process aspect will be monitoring implementation and adherence to the model - especially important given what is known about the importance of programme integrity (e.g., the better a programme is designed and delivered, the more likely it is to achieve its outcomes). “Overall there was a shared vision, a central place where those different services could together respond to child sexual abuse and provide a coordinated response”. Stakeholder interviewee 21 Delivery stakeholders identified one area they felt was missing from the London Lighthouse design was the exploratory interview – that is speaking to CYP where there are concerns of CSA/E, but where disclosure is yet to occur and therefore statutory agencies are yet to be involved. Whilst the Lighthouse has been keen to extend the eligibility criteria for the service (see appendix 2), there still needs to be significant suspicion of abuse from partner agencies. Another challenge outlined by staff has been the delayed national implementation of Section 2833 and the use of the Lighthouse as a Live Link location, which are part of the enhanced features due to be tested. These would enable a CYP to give evidence without needing to attend court, in a physical place that is familiar and hopefully safe. Stakeholders reported the unavailability of the technology for Section 28, along with a perceived lack of acceptance from of the judiciary to use both methods has been problematic. Delivery stakeholders emphasised importance of getting support from across the judiciary, at a national, rather than purely local level. The need for governance & internal communications A cross-organisational Programme Board, led by MOPAC has been in place during mobilisation to oversee delivery of the programme. This has ensured stakeholder relationships and a collective understanding have been developed - a clear positive finding. Since the Lighthouse opened in October 2018, this board has transitioned to an oversight function, with a concentration on the pilot’s future sustainability and options for further development of the model – something the Home Office Assurance review saw as a positive development to ensure continued stakeholder buy-in. RedQuadrant highlighted that some stakeholders on the ground raised more practical aspects they still felt needed to be addressed - particularly around specific information sharing policies, organisational funding and clearer feedback around decision making. The Home Office Assurance Review highlighted the importance of clarifying the scope and improving control and planning activities for the next phase of the programme. This will be crucial to securing future financing and completing the review and evaluation of the project. The benefits of a multi-agency partnership Sources consistently found the programme employed a thorough approach to stakeholder engagement and as a result staff were positive the right people and organisations have been included throughout (i.e., initial design process and beyond) and are very confident34 all partners will work well together. Establishing a collective approach, particularly through the existing CSA Hubs and CSA Transformation Programme is thought to have been a key enabler. This “shared ownership” was deemed critical for the Lighthouse to embed into the complex local environment of the North Central sector. With five different local authorities, several NHS trusts and at a time when the MPS are re-structuring their teams (including those who focus on CSA/E), it is unsurprising interviewees highlighted early links made with local safeguarding Children’s Boards and the MASHs as integral for the Lighthouse to become operational. One 33 Section 28 is of the Youth Justice and Criminal Evidence Act 1999, is using pre-recorded cross examination evidence of vulnerable and or intimidated victims and witnesses 34 Survey question: Please rate how confident you are that the partners involved in the Child House will work well together (scale of 1 = extremely not confident – 7 = extremely confident). 9/13 scored 6 or 7. 22 organisational relationship frequently identified to be working well was with the voluntary sector, as much effort has been put into including them and ensuring the smaller, local organisations have not been marginalised by the larger voluntary agencies. Another positive across multiple sources was the inclusion of the voice of the CSA/E previous victims and survivors, adult survivors and young people from across London. Stakeholders felt including their views was vital and very influential in the development of services - in particularly survivor views regarding health and well-being services were included in the look and feel of the Lighthouse infrastructure. The benefit of the Professional Advisory Group on collaborative working was also highlighted. Set up to capture the opinions of professionals from across disciplines to inform the project, it created key champions willing to promote the Lighthouse and push decisions forward. This was considered vital throughout mobilisation. However, one interviewee mentioned the sheer volume of engagement across stakeholders makes practical decision-making and moving forward difficult, particularly when sub-groups are not interlinking effectively. The co-located service design is especially thought to have facilitated multi-agency relationships, as it was both convenient and aided communication. During focus groups, partnership working and the resulting combination of expert knowledge was emphasised as one of the main attractions to working for the Lighthouse. Many staff saw their role as an “opportunity” to learn about other services/disciplines, and together provide better support for CYP. Staff were also positive that the partnership aspect will enable greater staff support during emotionally demanding cases. This was starkly compared to reflections from their previous roles where they had felt isolated. Bringing staff cultures together will require a continuing focus. Even through positive, staff identified challenges in understanding everyone’s roles and organisational structures, and alignment - although practices have been put in place to mitigate this gap, for example ‘lunchtime talks’. This is an issue that can be further developed through staff training, using a clear operational model and clarity in job roles. External Communications: the importance of engaging external stakeholders There have been a range of external communication regarding the Lighthouse - this includes media, all the way to communication from the Major and Home Secretary. This was a positive in raising the profile of the service. However, some delivery stakeholders wanted to see more done to promote the work, locally sharing internal workings with key borough professionals (i.e., referral processes), all the way to internationally promoting the work (i.e., through the EU promise). Positively, there is already a commitment to increase external publicity and promotion around referrals. Most of the external professionals who answered the survey were aware the Lighthouse would be opening (80%, n=35) and were positive towards the service. They were also (91%, “the other day we needed a report for a child and because I had a connection with a colleague I was able to get that really quickly instead of emailing and phoning. To get a medical appointment I just walked down the corridor to speak to the paediatrician”. Lighthouse Staff 23 n=49) confident35 in their own knowledge of CSA/E and identifying CSA (96%, n=52). External professionals also stressed the importance of agencies practically working together and adequate training in place for them to understand the service. The majority (84%, n=37) indicated they would like more training on CSA/E themselves – something the Lighthouse could offer to relevant external colleagues. Technical Set-up Most large-scale programmes encounter implementation challenges - perhaps the largest the Lighthouse encountered related to the procurement, design and implementation of the bespoke electronic Case Management System (CMS). The bulk of development was unable to start until after the lead provider was appointed. This left an ambitious time-scale and meant development had to run in parallel to mobilisation of the service itself. Delivery stakeholders identified this resulted in delays to the technology being ready for the opening of the service and impacted on when the service could ‘go live’, with staff preferring to have receive their training on the final system. The actual delivery of CMS within this time, being such a complex and ambitious IT solution has been a huge achievement and staff positively described how the face-to-face IT support from NELCSU (the IT provider) was very beneficial. Aside from the CMS, other technology issues were discussed, such as not having fully functioning phone lines with little instructions on their use; no printer ink; lack of laptops available; and no photocopier. Although commonplace in a new service, practical issues will be monitored over the course of the evaluation, to ensure the Lighthouse is fully operational and delivering an optimal service. The future One of the key themes to emerge from all sources of fieldwork was staff and stakeholders are already thinking of the sustainability and the future of the Lighthouse post the pilot period. In a sense this illustrates the level of support towards the service. Time-limited pilots do suffer from this uncertainty. The staff themselves will feel this, especially as the programme moves forward and into the second year. The Home Office Assurance Review highlighted the importance of continued stakeholder buy-in and a concern programme delivery is out of sync with potential funding timelines. However, it is positive this has been identified as a pivotal future risk by the Programme Board and has commissioned RedQuadrant to identify potential sources and methods of sustainable funding. 5. Discussion The current document is the first within the MOPAC Evidence and Insight two-year evaluation of the London Lighthouse. The report aims to establish the broad parameters of the evaluation as well as presenting very early insights from cost analysis and fieldwork As outlined, the evaluation will cover the performance, process, impact and economics of the Lighthouse. This breadth is positive given the lack of robust evidence in terms of impact and costings around such Child House models. In this way, it is hoped the evaluation will become a benchmark, robustly exploring aspects out of the scope of previous evaluations. 35 scored between 5-7 on the scale “It’s really good having NELCSU around, having face-to-face tech support when all these issues come up” and it “takes the pressure off the admin team”. Lighthouse Staff 24 There is consensus the delivery of the Lighthouse vision into a service which opened its doors in October 2018 is huge achievement for all partners and stakeholders. The complicated practical set up of the procurement, commissioning, estates, IT and mobilisation have incurred minimal delays, given the complex multi-agency nature of the service. In terms of results so far, all staff, partners and stakeholders were very positive in terms of the vision, design, governance, set-up and partnership approach of the Lighthouse. Given the level of underlying complexity and size of the model - this is something to be celebrated and itself is an important enabler in the ongoing effective implementation of the initiative. Some challenges have been raised by staff, such as the importance of linking with the judiciary; the use of Live link and section 28 and so on - although it is hoped these issues will be subsequently picked up and addressed by the Lighthouse governance structure. Likewise, staff are already aware of the time limited nature of the Lighthouse pilot, and it is positive sustainability is already being planned. As the service continues, it is advised to keep staff informed of progress to ensure ongoing awareness and maintain enthusiasm. The Home Office Assurance Review highlights this as a necessity, as a reduction in commitment from key stakeholders could affect the required funding for the future and the ability to reach benefit realisation. From an evaluation perspective, there is recognition the design of the Lighthouse has remained consistent from the initial vision and model – again this is positive given what is known around the importance of programme integrity. However, there are some changes and although the Lighthouse is the first opportunity to test the model in a UK setting, it will be important to monitor these. Evaluation reports released across the next two years will enable learning both internally, as a catalyst for improvement, and externally to advance the evidence base. 25 References: Allnock, D. and Miller, P. (2013) No one noticed, no one heard: a study of disclosures of childhood abuse. London: NSPCC. Bellis, M., Lowey, H., Leckenby, N., Hughes, K. & Harrison, D. (2014). Adverse childhood experiences: Retrospective study to determine their impact on adult health behaviours and health outcomes in a UK population. Journal of Public Health, 36, 1, 88 – 91. Bentley, H. et al (2017) How safe are our children? The most comprehensive overview of child protection in the UK 2017. London: NSPCC. Bentley, H. et al (2018) How safe are our children? The most comprehensive overview of child protection in the UK 2018. Brown, J. O’Donnell, T. & Erooga, M. (2011). Sexual Abuse: A Public Health Challenge. NSPCC. Children’s Commissioner for England, (2015). Protecting children from harm: A critical assessment of child sexual abuse in the family network in England and priorities for action. London. https://www.childrenscommissioner.gov.uk/wp-content/uploads/2017/06/Protecting- children-from-harm-executive-summary_0.pdf Children’s Commissioner for England (2016). Barnahus: Improving the response to child sexual abuse in England. London Cross, T.P., Jones, L.M., Walsh, W.A., Simone, M., & Kolko, D. (2007). Child forensic interviewing in Children’s Advocacy Centers: Empirical data on a practice model. Child Abuse & Neglect, 31(10), 1031-52. CSA Centre (2017) Measuring the scale and changing nature of child sexual abuse and child sexual exploitation CSA Centre (2018) Improving understanding of the scale and nature of child sexual abuse: Briefing. Diesen, C. & Diesen, E.F. (2009) Assault on women and children: The legal practise. Stockholm. Edinburgh, L., Saewyc, E., & Levitt, C. (2008). Caring for young adolescent sexual abuse victims in a hospital-based children’s advocacy center. Child Abuse & Neglect, 32, 1119–1126. Elmquist, J., Shorey, R.C., Febres, J., Zapor, H., Klostermann, K., Schratter, A., Stuart, G.L. (2015). A review of Children’s Advocacy Centers’ (CACs) response to cases of child maltreatment in the United States. Aggression and Violent Behavior, 25, 26-34. Goddard, A., Harewood, E., & Brennan, L. (2015). Review of Pathway following sexual assault for children and young people in London. Kings College Hospital NHS Foundation Trust, NHS England (London). 26 Hagborg, J.M., Stromwall, L.A. & Tidefors, I. (2012) Prosecution rate and quality of the investigative interview in child sexual abuse cases. Journal of Investigative Psychology and Offender Profiling, 9, 161-173. Harker, L. et al (2013) How safe are our children? London: NSPCC. HM Treasury. (2011) The Magenta Book: Guidance for evaluation. HM Treasury: London Herbert, J L and Bromfield L (2016) Evidence for the Efficacy of the Child Advocacy Center Model: A Systematic Review Trauma Violence and Abuse 2016 Vol 17 (3) 341-357. Herbert, J L and Bromfield L (2017) Better Together? A Review of Evidence for Multi- Disciplinary Teams Responding to Physical and Sexual Child Abuse Trauma Violence and Abuse 2017 1-15. Herbert, Walsh and Bromfield (2018) ‘A national survey of characteristics of child advocacy centers in the United States: Do the flagship models match those in broader practice?’ Child Abuse & Neglect 76 p583–595). Joa, D., & Edelson, M. G. (2004). Legal outcomes for children who have been sexually abused: The impact of child abuse assessment center evaluations. Child Maltreatment, 9, 263–276. Jones, L. M., Cross, T. P., Walsh, W. A., & Simone, M. (2007). Do children’s advocacy centers improve families’ experience of child sexual abuse investigations? Child Abuse & Neglect, 31, 1069–1085. Kaldal, A (2015). A Child evidence – a comparative study on handling, protection and testing evidence in legal proceedings within states in the Baltic Sea Region. Landberg, A and Svedin, C G (2013). Inuti ett Barnahus: a quality review of 23 Swedish barnahus. Radda Barnen Lippert, T., Cross, T.P., Jones. L.M., & Walsh, W.A. (2009) Telling interviewers about sexual abuse: Predictors of child disclosure at forensic interviews. Child Maltreatment, 14, 100-113 Miller, A., & Rubin, D. (2009). The Contribution of children's advocacy centers to felony prosecutions of child sexual abuse. Child Abuse & Neglect, 33(1), 12-18 ONS (2016) Abuse during childhood: Findings from the Crime Survey for England and Wales, year ending March 2016 ONS (2017) Sexual offences in England and Wales: year ending March 2017: Analyses on sexual offences from the year ending March 2017 Crime Survey for England and Wales and crimes recorded by police. https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/sexualoffencesi nenglandandwales/yearendingmarch2017#which-groups-of-people-are-most-likely-to-be-victims-of- sexual-assault) 27 Parkinson, P.N., Shrimpton, S., Swanston, H.Y., O’Toole, B.I., Oates, R.K. (2002) The process of attrition in child sexual assault cases: a case flow analysis of criminal investigations and prosecutions. The Australian and New Zealand Journal of Criminology, 35, 347-362. Rasmusson, B. (2011) Children’s advocacy centers (Barnahus) in Sweden: Experiences of children and parents. Child Indicators Research 4 301-321. Smith, D.W., Witte, T.H., & Fricker-Elhai, A.E. (2006). Service outcomes in physical and sexual abuse cases: A comparison of child advocacy center-based and standard services. Child Maltreatment, 11, 354-360 Walsh, W., Cross, T., Jones, L., Simone, M., & Kolko, D. (2007). Which sexual abuse victims receive a forensic medical examination? Child Abuse & Neglect, Volume 31, Issue 10, Pages 1053-1068 Walsh, W., Lippert, T. Cross, T., Maurice, D., & Davidson, K. (2008). How Long to Prosecute Child Sexual Abuse for a Community Using a Children’s Advocacy Center and Two Comparison Communities? Child Maltreatment, 13, 1, 3-13 Warrington, C with Beckett, H Ackerley E, Walker M and Allnock D (2017). Making Noise: Children’s voices for positive change after sexual abuse: Children’s experiences of help- seeking and support after sexual abuse in the family environment. Children’s Commissioner Wolfteich, P., & Loggins, B. (2007, August). Evaluation of the children’s advocacy center model: Efficiency, legal and revictimization outcomes. Child & Adolescent Social Work Journal, 24(4), 333-352 28 Appendices Appendix 1: Lighthouse partner relationships • NHS England (London Region) is the commissioner of the service & have no data access rights. • MOPAC is joint programme lead & a joint controller of CH data. MOPAC are also a joint controller of MPS data. • RedQuadrant is MOPAC’s contracted processor for the cost-benefit analysis. • Opinion Research Services (ORS) is MOPAC’s contracted processor for online surveys. • NELCSU MOPAC’s IT contractor have no data access rights • Excelicare MOPAC’s IT contractor have data access but no processing rights • The Home Office is MOPAC’s co-funder, they have no data access rights • The Department for Education (DfE) is MOPAC’s co-funder, they have no data access rights • The University College London Hospitals NHS Foundations Trust (UCLH) is the contracted Lighthouse lead provider & joint controller of all data inputted onto the CMS. They determine the purposes for which the data is to be processed for service users of Lighthouse • Camden Council is a sub provider for the Lighthouse service, is UCLH’s processor and provides elements of the service. • The Metropolitan Police Service (MPS) is a sub provider for the Lighthouse service & a controller of MPS data (joint with MOPAC) • The Tavistock and Portman NHS Foundation Trust is UCLH’s sub provider for the Lighthouse service & joint controller of CMS data • The National Society for the Prevention of Cruelty to Children (NSPCC) - is UCLH’s sub provider for the Lighthouse service & joint controller of CMS data • Morgan & Stanley are the NCPCC’s co-funder but have no data access rights • Solace are the NSPCC’s processor and sub provider for the Lighthouse service 29 Appendix 2: Disclosures of Child Sexual Abuse requiring referral to the Child House 1. Allegation of child sexual abuse made to police or social care Non-acute Child Sexual Abuse reported to police or social care directly or via school or other practitioner; and outside forensic window/not needing DNA swabs 2. Referral to Child House from the CYP Haven following forensic examination at the Haven Children and Young people examined at the CYP Havens for a forensic medical examination (FME), will be transferred to the Child House once the acute FME has taken place. 3. Significant suspicion of Child Sexual Abuse: Practitioners conclude, during a Section 47 discussion, that it is highly likely that sexual abuse has occurred and there are signs from Category A and/or B below: Category A: History of risk and some evidence of harm to the child or a sibling • They have been in contact with a known individual or alleged person who poses a risk of sexual harm. • They have a history of disclosure and retraction • There is a history of sexual abuse in their extended family • They are the sibling of a child who has disclosed or retracted sexual abuse Category B: Behaviours and physical symptoms: symptoms that lead practitioners to suspect child sexual abuse (as defined in the NICE guideline -NG76) Suspect current or past child sexual abuse: • If a pre-pubertal child displays or is reported to display repeated or coercive sexualised behaviours or preoccupation (for example, sexual talk associated with knowledge, emulating sexual activity with another child). • If a pre-pubertal child displays or is reported to display unusual sexualised behaviours, including: • oral–genital contact with another child or a doll • requesting to be touched in the genital area • inserting or attempting to insert an object, finger or penis into another child's vagina or anus • If there are persistent or recurrent genital or anal symptom (for example, bleeding or discharge) that is associated with behavioural or emotional change and that has no medical explanation • If a child younger than 13 years has gonorrhoea, chlamydia, syphilis, genital herpes, hepatitis C, HIV or trichomonas infection unless there is clear evidence of mother-to-child transmission during birth or blood contamination 4. Self-referral following child sexual abuse made to the Child House Self-disclosure by a young person or child and family/carer following non-acute Child Sexual Abuse, reported to the Child House directly 30 5. CONSULTATION with Child House team for advice on referral Practitioners can seek advice from the Child House if they CONSIDER child sexual abuse if the signs and symptoms below are associated with other concerns such as Domestic Violence, not attending school etc. and seek advice from the Child House team. • Ano-genital signs and symptoms e.g. gaping anus, dysuria (discomfort on passing urine), evidence of one or more foreign bodies in the vagina or anus. • Sexually transmitted infections e.g. hepatitis B, gonorrhoea or ano-genital warts, unless there is clear evidence of mother-to-child transmission during birth, non-sexual transmission from a member of the household, blood contamination or that the infection was acquired from consensual sexual activity with a peer • Pregnancy in a young woman aged 13 to 15 years • Pregnancy in a young woman over 16 years where there is a clear difference in power or mental capacity between the young person and their sexual partner, in particular when the relationship is incestuous or is with a person in a position of trust (for example, teacher, sports coach, minister of religion) • Concern that the young person is being exploited Exclusion criteria: • Children and young people requiring acute forensic medical examination • Victims who are also perpetrators or at high risk of offending (based on the professional judgement of the Child House team). The ‘status’ of a child or young person attending the Lighthouse may not become clear until sessions have commenced. Decisions are made locally with exception reporting used to inform the commissioner quarterly to facilitate a shared understanding as operational experience develops. • Those where an ‘exploratory interview’ is required to determine whether or not sexual abuse has occurred • Those living outside the geographical boundaries of the 5 London Boroughs served by the Lighthouse, based on the address at which the child or young person is living • Young people over the age of 18 years (although those between 18-25 years of age with learning delay or disability for whom a child or young person-oriented service appears more suitable will be accommodated. Exception reporting will be used to identify the volume of ‘clients’ falling into the 18-25-year-old age range so that this can be monitored, and the approach regularly reviewed. 31 Appendix 3: Child House Logic Model What is a logic model? A logic model describes the theory, assumptions and evidence underlying the rationale behind a project. It is a key tool to embed the evaluation within policy. The logic model is beneficial because it: ▪ adds clarity to the conversation; ▪ ensures the project’s motivations stay true; ▪ steers implementation and set out clear outcomes for measuring impact; and ▪ helps in setting a robust and cognitive data capture to allow evaluation 32 Appendix 4: Theory of change per outcome - why should there be an impact? Outcome Theory for change Evidence Base Enhance referral pathways into and out of the Lighthouse One single referral pathway directly into the Lighthouse for all five boroughs - stops multiple agencies across boroughs acting as gate keepers. There should be clearer guidelines on how to directly refer. Once in the Lighthouse, there will be greater standardisation for engaging with CYP and, where needed, referring them on to other services. A 2008 evaluation of the first Swedish Barnahus found cooperation between various authorities intensified and become more efficient. Contact increased; understanding for and knowledge about each other’s areas of expertise increased; and case conferences acquired more structure.36 Similar findings confirmed by a 2010 study and a 2012 evaluation of Norway’s Barnahus, which stressed benefits of all co-located professionals in minimising the child’s need to travel and in strengthening professionals’ competence and the coordination of their interventions37. Improved CYP, family and carer experience of support received post disclosure The Lighthouse service model offers 2 years’ worth of support for CYP and family not found elsewhere. All services under one roof should mean the CYP will only be required to relay events once and there should be more streamlined support to navigate services on offer. A 2016 systematic review of Child Advocacy Centres (CACs) concluded ‘it was striking how few studies were directed toward assessing child and family outcomes’38. Only 3 studies39 examined how CAC led to recovery from trauma, with varying results, and none measured benefits against standard service delivery. Whilst pre- and post-treatment comparisons40 found no significant improvement in measures of family empowerment, there were significant improvement in the use of community resources (i.e., any type of support services). Hubel et al (2014) also found significant improvements in family functioning following group treatment within a CAC41. In terms of caregiver satisfaction, whilst research indicates it is often high, there is usually no difference to the comparison groups. This is true for overall satisfaction with the services offered by CACs42, where users also rated the performance of the centres highly (although there was no standard service delivery comparison) and specifically with medical exams, where there were high levels of satisfaction across both CAC and comparison community cases43. 36 Landberg and Svedin 2013 37 Landberg and Svedin 2013 38 Herbert and Bromfield 2016 39 Hubel et al 2014; Brown 2007 and Jensen et al 1996 cited in Herbert and Bromfield 2016 40 Carman (2004) cited in Herbert and Bromfield 2016 41 Herbert and Bromfield 2016 42 Carman (2004); Bonach, Mabry, and Potts-Henry (2010); Rasmusson (2011) and Klenig (2012) cited in Herbert and Bromfield 2016 43 Walsh et al. (2007) cited in Herbert and Bromfield 2016 33 This may be expected as caregivers are not to know how the features of service are different and any intervention/support at a difficult time would be beneficial. A 2017 report identified many benefits for families following therapeutic support, including: a safe space in which to process what happened; knowing others have comparable experiences; being believed; countering stigma, isolation and self-blame; the development of coping strategies; and wider confidence and resilience building. Research concluded while benefits could be achieved in non-specialist settings, specialist CSA services were particularly helpful in countering stigma, isolation and understanding the complexities of CSA in the family environment44. Improved CYP experience of the criminal justice process post disclosure CYP will be supported throughout the process by an advocate. They will not be required to go to a police station as all interviews are conducted in the Lighthouse (e.g. psychology led ABE). The use of live link should negate the need to go to court and the associated stress with the process. A 2012 evaluation of the Barnahus model in Norway suggested children interviewed by the police in the facility received better care than those interviewed at a police station or in a district court.45 An evaluation of Barnahus in Sweden found it led to improvements in children and families’ experiences of the criminal justice process.46 When comparing satisfaction with the investigation of CSA, research47 found caregivers in CAC cases were significantly more satisfied than comparisons, although children reported no significant difference in satisfaction with the investigation between the two conditions. Improved mental health and well-being outcomes for CYP It will be easier and quicker access to longer term support for mental health, rather than having to meet the high criteria to access CAMHS who have long waiting lists and potentially shorter-term interventions. A 2017 systematic review of multi-disciplinary teams (MDTs) found a significant difference between MDTs and comparisons in increasing the uptake of mental health and support services. Three studies48 compared the extent of referrals to the use of services, and all found outcomes related to service use were significantly greater. In terms of medical outcomes, an earlier systematic review (2016) found CAC cases were more likely than comparisons to have a medical exam49, although this was not associated with offenders being charged. Saewyc et al also found CAC were more likely to receive referrals to counselling, for STI tests and treatment when needed50. 44 Warrington et al 2017 45 Landberg and Svedin 2013 46 Rasmusson 2011 47 Jones, Cross, Walsh, and Simone (2007) cited in Herbert and Bromfield 2016 48 Edinburgh et al.,2008; Smith et al., 2006; Turner, 1997 cited in Herbert and Bromfield, 2017 49 Smith, Witte and Fricker-Elhai (2006); Walsh, Cross, Jones, Simone and Kolko (2007), and Edinburgh, Saewyc and Levitt (2008) cited in Herbert and Bromfield, 2016 50 Saewyc and Levitt (2008) cited in Herbert and Bromfield, 2016 34 Compared to before its inception, the Icelandic Barnahus yielded positive results, improving therapeutic outcomes for children and their families.51 In a pre-post study, Hubel et al. (2014) found cognitive behavioural group therapy delivered to sexually abused children and non-offending caregivers within the CAC, led to decreases in depression; loneliness; anxiety; fears about victimisation; post-traumatic stress; social functioning; and attributions about the abuse52 Jenson et al. (1996) found significant reductions in parents’ sense their child was ‘‘demanding of their time,’’ ‘‘was not their regular self,’’ ‘‘had trouble falling asleep,’’ and ‘‘was more afraid than usual.’’ However, when examining the reduction in trauma symptoms Brown (2007) found there were no significant changes in measures of depression and post-traumatic stress over the course of the therapy provided at the CAC53 Improved professionals’ awareness, competence and confidence in working with CSA/CSE The Lighthouse should raise its profile in local community, ensuring everyone knows the new referral pathways. Professionals should become more aware and competent in what to do with eligible cases. A 2012 evaluation of the Barnahus in Norway found its introduction led to greater coordination among professionals and an increase in awareness of CSA in the general population54 A 2013 Swedish study reported the Barnahus developed into local or regional knowledge centres for questions about violence against children55 Increased likelihood for CYP who received a Lighthouse service to have CYP are more likely to report in the first instance and, as they are supported by advocate, stay engaged with the process. Better evidence collected through ABE Compared to the period before its inception, the Icelandic Barnahus yielded positive results, trebling the number of perpetrators charged56. A 2017 systematic review suggested evidence was mixed as to whether MDTs resulted in more arrests and prosecutions than comparisons. Whilst many earlier studies57 found significant differences, this has not been the case more recently58 leading to the suggestion many practices of MDTs and CACs had diffused into ‘practice as usual’, resulting in a higher baseline for MDTs in later studies. 51 Children’s Commissioner 2016 52 Hubel et al. (2014) cited in Herbert and Bromfield 2016 53 Jenson et al. (1996) and Brown (2007) cited in Herbert and Bromfield 2016 54 Landberg and Svedin 2013 55 Landberg and Svedin 2013 56 Children’s Commissioner, 2016 57 Jaudes & Martone, 1992; Turner, 1997 cited in Herbert and Bromfield 2017 58 Edinburgh, Sawyc and Levitt 2008, Wolfteich and Loggins 2007 cited in Herbert and Bromfield 2017 35 cases charged by CPS. increasing the chance of meeting thresholds for charging. It appears differences in CJ outcomes may be linked to which part of the process is assessed. Outcomes earlier in the CJ process (e.g. the level of police substantiations) were more likely to be significantly different between MDTs and their comparisons59 than not60. However, results were less consistent for outcomes later in the process, like criminal charges filed/prosecutions for abuse. While some studies found significant increases61, others found no difference between MDTs and comparisons62. An earlier 2016 systematic review mainly supported the idea that CACs resulted in favourable criminal justice outcomes 63 , although unsurprisingly, this finding was more equivocal when compared to police/child protection partnership models64. Bradford (2005) found rates of charges and convictions increased significantly after the introduction of CACs and Joa and Edelson (2004) found CACs were more likely to have charges filed, cases charged, and guilty pleas than traditional practice, but no significant difference between guilty and non-guilty verdicts.65 In terms of the speed taken to progress through the CJS, research suggests when an arrest was made, communities with CACs had a significantly faster case resolution time compared with demographically equivalent communities66. Wolfteich and Loggins (2007) found significant differences in the time between initial reports and cases authentication across child protection teams (100 days av), CACs (225) and traditional service delivery (311), although the authors noted figures might reflect the referral of complex cases to CACs.67 Increased likelihood for CYP who Better evidence collected through ABE and better cross examination Compared to the period before its inception, the Icelandic Barnahus doubled the number of convictions.68 However, a 2017 systematic review found results for convictions following a CAC were mixed - three studies suggesting a significant difference69 compared to two that did not.70 59 Jaudes & Martone, 1992; Ruggieri, 2011; Smith et al., 2006; Wolfteich & Loggins, 2007 cited in Herbert and Bromfield 2017 60 Wolfteich & Loggins, 2007 cited in Herbert and Bromfield 2017 61 Bradford, 2005; Joa &Edelson, 2004; Miller & Rubin, 2009; Turner, 1997 cited in Herbert and Bromfield 2017 62 Campbell, Greeson, Bybee, & Fehler-cabral, 2012; Edinburgh et al., 2008; Goldbeck et al., 2007; Wolfteich & Loggins, 2007 cited in Herbert and Bromfield 2017 63 Joa and Eddelson, 2004; Miller & Rubin 2009 cited in Herbert and Bromfield 2016 64 Wolfteich & Loggins, 2007 cited in Herbert and Bromfield 2016 65Bradford (2005) and Joa and Edelson 2004 cited in Herbert and Bromfield 2016 66 Walsh, Lippert, Cross, Maurice and Davison 2008 cited in Herbert and Bromfield 2016 67 Wolfteich and Loggins 2007 cited in Herbert and Bromfield 2016 68 Children’s Commissioner 2016 69 Bradford,2005; Joa & Edelson, 2004; Turner, 1997 cited in Herbert and Bromfield 2017 70 Edinburgh et al., 2008; Joa & Edelson, 2004 cited in Herbert and Bromfield 2017 36 received a Lighthouse service to have their case end in conviction through live link therefore CYP more likely to be believed in court. A 2004 study71 found CACs were more likely to have charges filed, cases charged, and guilty pleas (the higher proportion of guilty pleas were attributed to the fact CACs produced more compelling evidence). However, there was no significant difference between guilty and non-guilty verdicts when comparing to traditional practice. Similarly, a 2005 study72 found rates of charges and convictions increased significantly after the introduction of CACs. Improved partnership working All multi-agencies under one roof will make communication better and data sharing issues less problematic. Cases are less likely to fall through the net and partners can work in collaboration with each other. The 2017 systematic review concluded the five studies which examined a multi-disciplinary response in a CAC ‘found mostly significant results for the effect of increased collaboration or ties between service agencies’ in terms of service provision, mental health service use or mental health improvement.73 A 2012 study of the Norwegian Barnahus model found its introduction led to greater coordination among professionals, the justice system, and in those who provide crisis support and treatment.74 A 2010 Swedish evaluation concluded the introduction of Barnahus led to increased collaboration between the various professional groups and disciplines. This resulted in a broader knowledge base and a deeper foundation for the investigative and the cross-disciplinary collaboration enabled the ability to tackle CYP’s problems from different perspectives.75 A 2007 study76 found CAC cases were more likely than comparisons to feature multi-disciplinary team interviews; case reviews; joint police and child protective investigations; and video or audiotaping of interviews. Whilst this could suggest more effective partnership working in CACs could lead to less time interviewing victims, no significant differences were found in the number of interviews conducted by the CACs and comparator sites. Providing CSA victims care and support to reduce the long- term impact of victimisation: Impact on well-being after the service. Research is lacking. It was identified more research is needed in terms of child and family outcomes, both in terms of the effect of more child-friendly practices and of supported referrals to therapeutic services77 Organisations are committed Organisations not directly involved in service delivery There is some evidence of mainstreaming changes implemented as part of the new practice, however much research is based on the criminal justice outcomes, with studies generally finding significant 71 Joa and Edelson (2004) cited in Herbert and Bromfield 2016 72 Bradford (2005) cited in Herbert and Bromfield 2016 73 Herbert and Bromfield 2017 p7 74 Landberg and Svedin 2013 75 Landberg and Svedin 2013 76 Cross, Jones, Walsh, Simone and Kolko cited in Herbert and Bromfield 2016 77 Herbert and Bromfield 2017 37 to being victim focused in their support of CSA victims. supporting the Lighthouse e.g. courts open to using Live Link and S.28 Changes in current practice to be more victim focused differences earlier in the criminal justice process. As significant differences were more likely to be seen in the initial studies, there is a suggestion some CACs practices have been incorporated more broadly, but this is not systematically found. It is suggested although it is clear well-implemented teams are likely to lead to improved outcomes compared to responses built around individual agencies, little is known about the ‘optimal configurations of teams, and what needs to be done to be in place to foster effective teams’ 78. There is some evidence of other organisations supporting the new approach – e.g. the development of joint performance measurement and evaluations79 and improvements in information exchange between CACs and other agencies80 78 Herbert and Bromfield 2017 79 Lalayants 2015, Bertram, 2008, Ells 2000, Ferguson, Baines, Schneider and Galloway 1994 cited in Herbert and Bromfield 2017 80 Ruggieri 2011, cited in Herbert and Bromfield 2017 38 Appendix 5: E&I Research Product Timeline 39 Appendix 6: Criminal Justice Attrition Tree 1 The Lighthouse: Final evaluation report Rachael Parker, Tim Read, Alexander Scragg & Paul Dawson MOPAC Evidence & Insight June 2021 2 Contents Executive Summary .................................................................................................................. 4 Glossary of Acronyms .................................................................................................................. 8 Acknowledgements ..................................................................................................................... 9 1. The Lighthouse evaluation ............................................................................................... 10 Background ............................................................................................................................... 10 Methodology ............................................................................................................................ 14 Performance management data ................................................................................................ 15 Qualitative fieldwork undertaken during the evaluation ............................................................ 15 Attempting to understand impact of the Lighthouse ................................................................ 18 Economic analysis ..................................................................................................................... 20 2. Performance insights ........................................................................................................ 22 Referrals to the Lighthouse ....................................................................................................... 22 A deeper look - background of clients who consented to evaluation ........................................ 25 Offence details .......................................................................................................................... 27 Delivering the Lighthouse service ............................................................................................. 27 3. Implementing the Lighthouse .......................................................................................... 29 Initial set up and maturation of the Lighthouse ........................................................................ 29 Implementing during Covid-19 ................................................................................................. 33 The value and challenges of partnership working ...................................................................... 35 4. Impact of the Lighthouse ................................................................................................. 38 Comparison of Performance data: Lighthouse and NEL CSA hub .............................................. 38 Service users’ perceptions ......................................................................................................... 40 Parents’ perceptions of the Lighthouse..................................................................................... 42 Emotional wellbeing outcomes .................................................................................................. 43 Comparison of criminal justice outcomes and attrition .............................................................. 43 5. Economic analysis and the future ................................................................................... 46 The future ................................................................................................................................. 48 6. Discussion ........................................................................................................................... 50 References .............................................................................................................................. 54 Appendices ............................................................................................................................. 57 Appendix A – Full list of criminal Justice coding framework variables and their frequencies, comparing Lighthouse with NEL ............................................................................................... 57 3 Appendix B – Comparison of NEL criminal justice cases sample with final comparison group ... 60 Appendix C - Supplementary Lighthouse performance tables ................................................... 62 Appendix D - Description of service provision in NEL ............................................................... 65 Appendix E – Supplementary tables of comparisons between Lighthouse and NEL services ..... 69 Appendix F – Calculations underpinning economic evaluation .................................................. 72 Appendix G - Summary of Early Emotional Support (EES) Services commissioned in London ... 82 4 Executive Summary The Lighthouse, London’s Child House, opened at the end of October 2018 initially as part of a two-year pilot, although now funded until March 2022. Bringing together a range of organisations under one roof, the Lighthouse is a child friendly, multidisciplinary service for victims1 of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden, it serves the five surrounding North Central London boroughs of Barnet, Camden, Enfield, Haringey and Islington. The Evidence and Insight (E&I) Unit are the Mayor’s Office for Policing and Crime’s (MOPAC) in- house social research and analytical team and were commissioned to evaluate the Lighthouse. The E&I evaluation focuses on four distinct areas for analysis; a performance review; a process evaluation; impact evaluation and an economic evaluation. This report brings the evaluation to a close. Results Summary of performance insights • Between the end of October 2018, when the Lighthouse launched, and the end of March 2021 there were a total of 889 referrals to the service, which works out as an average of around 30 referrals per month, slightly lower than anticipated. Out of the total referrals – the majority were female (82%, n=730), with over half in the older age group between 13-17 years (57%, n=505). • Between the end of October 2018 and the end of March 2021 the service carried out 510 Initial Assessment (IA). This calculates at an overall conversion rate of referral to IA at 57%. • Of these cases receiving an IA (and whom consented, n=360) the demographics were similar to the overall cohort of referrals. The majority were female (85%, n=306); an average age of 13; nearly half were between 13-17 (n=171, 48%). Of these, ethnicity data was recorded for 301 service users and there was a fairly even split between BAME and non-BAME clients (n=161, 53% and n=140, 47% respectively). • In terms of vulnerabilities – 84% (297/354) of those receiving an IA were reported to present at least one type of vulnerability (at a total of 793 vulnerabilities with an average of 2.7 each). 201 (68%) of these service users have at least 2 types of vulnerability. Among the most frequent categories of vulnerability were anxiety and/or depression (60%, n=121), followed by history of domestic violence (55% DV, n=110) and education problems (41%, n=82). • Across the Lighthouse lifespan there was a considerable amount of delivery – at a glance the data indicates: - A total of 4780 telephone, video or face to face sessions. - 936 professional meetings (excluding strategy and consultation meetings). 1 Referred to as victims throughout the remainder of the report. 5 - There were 29 psychologist-led ABE (Achieving Best Evidence) interviews that took place at the Lighthouse (averaging at 1.6 per month).2 Over the same period there were 43 police led ABE interviews at the Lighthouse.3 - The Lighthouse made 91 onward referrals to local services, most commonly within the voluntary sector (34%, n=31), followed by Social services (26%, n=24). - Between April 20 and March 21, 137 strategy discussions that took place and 118 consultations delivered by the Social Care Liaison Officers (SCLOs). Summary of learning from implementation • Overall, the pilot was well implemented. Staff, partners, children & young people (CYP) and parents were positive about the general service. As expected, given the size and scale of the Lighthouse, the pilot experienced several implementation and maturation challenges – some of these were required and indeed instigated by staff, others were due to wider factors that staff had to work around, and these resulted in changes to the model and often staff not working in a way initially envisioned. However, staff responded well, and although some aspects were not totally addressed, many were resolved over the duration of the pilot. • Implementation of the Lighthouse was heavily disrupted by the CV-19 pandemic. In response, considerable changes were made in order to deliver the pilot – and to the credit of staff the service continued. Indeed, many of these virtual changes became embedded into the routine way of working, but overall staff felt that virtual working whilst necessary, was lacking therapeutically compared to face to face. • Overall, partnership working was a resounding positive to emerge from the evaluation, appearing to bring a wide range of benefits to the service as well as the clients – something supported by specific roles (Police Liaison Officer [PLO], Social Care Liaison Officer [SCLO]) that enabled such a way of working. However, it is clear some tensions as a result of different organisational cultures and working practices were evident which were not able to be fully reconciled over the pilot duration. This serves to illustrate how such working cultures are very difficult to avoid and should be borne in mind when devising and running any similar programme. Exploring impact of the service • In order to explore the question of ‘impact’ the evaluation seeks to triangulate across a range of data sets. This is the most suitable approach given the limitations across timeframes, sample sizes and methodology. The analysis sees the Lighthouse compared to a Business as Usual (BAU) service from a nearby geographic area – North East London (NEL) – across issues such as performance, referrals and Criminal Justice Outcomes. Key results indicate: - As expected, Lighthouse reached a far larger number of clients and delivered far more outputs compared to the BAU site. - Emotional Wellbeing outcome data is variable across sites, however comparing available data between Lighthouse and NEL indicates 89% of Lighthouse goals were achieved or partially achieved – whereas for NEL there were improvements in 47% of 2There were more psychology led ABEs planned between April and June 2020, but a number were cancelled due to family anxiety about travelling in for an interview during COVID lockdown. 3 The Lighthouse is contracted to offer 3 psychologist-led ABEs a month. 6 outcomes (see Appendix for full breakdown of outcome results). This is a significant difference although caveats should be borne in mind. - Upon comparing two key outcomes of cases progressing through the Criminal Justice System (charge and conviction) there was almost no difference between Lighthouse and NEL two groups. The Lighthouse cases had 7% (n=6) cases charged by the Crown Prosecution Service (CPS) and a 5% (n=4)4 conviction rate; compared to NEL who had 6% (n=4) charged by the CPS and a 4% (n=3) conviction rate. The analysis was limited by the proportion of cases progressing through the CJS and the sample sizes. - Comparing across a range of investigative actions between the two groups, there were some positive and encouraging findings. For example, the Lighthouse had significantly higher instances of positive investigative actions such as increased suspect arrests (44% vs 27%, p<0.05), and proportion of cases submitted to the CPS (24% vs 10%, p<0.05). Additionally, more early investigative advice was sought from CPS as compared to NEL; something that was actively worked on between Lighthouse and CPS during the pilot through case progression meetings. - The qualitative data obtained indicated that staff, parents and clients themselves all spoke very positively about the impact of the service. Exploring the economics • Economic analysis was commissioned out and completed by RQ. They provided an indicative cost-benefit analysis (CBA) which involved exploring the benefits of the Lighthouse against the costs associated with setting up and delivering the Lighthouse. This was then compared against a comparison group and a difference-in-difference analytical approach which estimates the differential effect of the Lighthouse on its service users. • The overall costs of the Lighthouse pilot comprise the annual operating costs of £2.387m per year 5 plus the one-off capital costs of refurbishment, installation of IT and infrastructure (£3.9m allocated to cover a period of eight years at £0.49m per year). This equates to £2.88m per year in total. • With this overall cost, and 420 clients per year, the unit costs for the Lighthouse are of the order of £6,860. Costs of a Havens service were estimated at £4,925 per case. • Unfortunately, despite development of a comprehensive dataset for monitoring performance and utilisation rate, accurate data on many outcomes is not available. • Outcome measurement is particularly challenging for CSA services due to the complexity of the recovery with every child’s journey likely to be different. • This total cost can then be set against the wider social value identified across three broad areas: wellbeing to the client, useful savings from public sector spend and additional public sector spend on essential activity. • These results show that there is a financial return to the public sector from the operation of the Lighthouse, with a net gain in public expenditure per client of £14,570. The ratio 4 One of the guilty outcomes includes two suspects within the same case (both convicted). 5 Comprising the costs of directly employed staff, diagnostic equipment, drug costs, clinical supplies and IT, and overheads together with rent, service charge and rates (£0.38m per year), and police and social care costs (£0.28m per year). 7 between future savings and cost of the intervention is 3.12 : 1.0 (calculated as £21,430 ÷ £6,860). However, the comparator scheme also shows a good financial return, with a net gain in public expenditure per client of £9,460. The ratio between future savings and cost of the intervention is 2.92 : 1.0 (calculated as £14,385 ÷ £4,925). These were calculated over a period of 30 years. • The calculations suggest that the use of the Lighthouse compared to the Haven costs perhaps £1,935 per case more, but saves an additional £7,000 on future public expenditure, and improves wellbeing by an additional £10,3006. Conclusion This report brings the 3-year evaluation of the Lighthouse service to a conclusion. This was a substantial project, spanning 4 reports, and comprised of a multi-method action research approach incorporating the views of over 200 individuals (staff, clients etc…), performance analysis, wider analysis on outcomes as well as economic methods. Lighthouse received 889 referrals and completed 510 IAs over the evaluation. Clients were highly vulnerable. There was considerable delivery of services. Overall, perceptions towards the Lighthouse were positive across all groups. Implementation was good, and whilst not entirely as expected this was often due to staff agency in reacting to unexpected challenges as well as wider issues out of their control. There was a huge effort in response to CV-19 to continue the service and many aspects of virtual working became embedded into routine practice. Partnership working emerging as a critical benefit, although some organisational tensions remained throughout. Exploring impact triangulated across a range of data – compared to a BAU; Lighthouse delivered more, had more positive wellbeing outcomes, had more positive investigative actions but no difference in terms of charges or convictions. Cost data indicated that Lighthouse was more expensive than the BAU but may be able to generate a more positive financial return over a long time period. The evaluation is the largest of its type in the country and has generated a wealth of useful information for practitioners, staff and wider academics within the CSA/E field. There is a need for continuing research into the impact of the Lighthouse, particularly around criminal justice outcomes because of the small number of cases available. Safeguarding will remain a key issue for London and beyond and it is hoped the findings herein can contribute to keeping young Londoners safe and protected. 6 Details of the calculations that underpin these figures are provided in Appendix F of this report 8 Glossary of Acronyms ABE Achieving Best Evidence ACE Adverse Childhood Experience BAU Business As Usual CAMHS Child and Adolescent Mental Health Services CBA Cost-Benefit Analysis CMS Case Management System CPS Crown Prosecution Service CRIS Crime Record Information System CSA Child Sexual Abuse CSE Child Sexual Exploitation CYP Children & Young People DfE Department for Education DWP Department for Work and Pensions E&I Evidence & Insight HMCTS Her Majesty’s Courts and Tribunals Service IA Initial Assessment LTFI Letting the Future In MASH Multi-Agency Safeguarding Hub MOPAC Mayor's Office for Policing and Crime MPS Metropolitan Police Service NEL North East London NFA No Further Action NHSE National Health Service England NSPCC National Society for the Prevention of Cruelty to Children P&R Protect & Respect PLI Psychologist-Led Interview PLO Police Liaison Officer RCADS Revised Children's Anxiety and Depression Scale SCLO Social Care Liaison Officer TSCC Trauma Symptom Checklist for Children UoB University of Bedfordshire VRI Video Recorded Interview 9 Acknowledgements MOPAC E&I would like to thank all the practitioners, stakeholders and service users who gave up their time to share their views and perspectives with the evaluation team. In particular we would like to thank; Staff at the Lighthouse, particularly Emma Harewood, Georgia Johnson, Meta Madden, Vicky Mattison, Prince Nwanguma, Martin Slack and Eimear Timmons. Chloe Collins at NHS London, Dr Anna Riddell at the Royal London Hospital and Justine Wilson-Darke at Barnardos. Claire Bethel and Neil Reeder at RedQuadrant for their work on the cost benefit analysis. Dr Helen Beckett, Claire Soares and Dr Camille Warrington at the University of Bedfordshire for undertaking the service user evaluation. Our policy colleagues in MOPAC (particularly) Gareth Linington and Dean Nevill. The following colleagues at Evidence and Insight for their assistance with the research; Lynne Conroy and Dr Zoe Hobson for their earlier work on the evaluation, Barry Charleton and Emanus Prospere for their assistance coding CJS cases, Laura Duckworth for help with transcription and analysis, Valerie Forrester for providing CSA data. The members of the Lighthouse Academic Advisory Board for their advice and guidance; Professor Simon Hackett, Dr Julie Harris, Dr Katrin Hohl, Professor Liz Kelly, Professor Michael Lamb, Lisa McCrindle. 10 1. The Lighthouse evaluation Background This report presents the findings of a three-and-a-half-year evaluation of the Lighthouse, London’s Child House – an innovative service to support victims of Child Sexual Abuse and Child Sexual Exploitation (CSA/E). This first section provides an overview of the rationale of the model and the previous evidence relating to it. The Lighthouse, London’s Child House, opened in October 2018 initially as part of a two-year pilot, although funding was subsequently extended until March 2022. Bringing together a range of services (medical, social care, police, advocacy and therapeutic support) under one roof, the Lighthouse is a child friendly, multidisciplinary service for victims and survivors of CSA/E. Based in Camden it replaces the existing services7 and serves the five surrounding North Central London boroughs of Barnet; Camden; Enfield; Haringey; and Islington. The Lighthouse builds on the CSA Hubs but offers enhanced support to children and young people (CYP) aged between 0 – 18 years old (or those aged over 18 to 25 years of age with learning delay or disability for whom a child or young person-oriented service appears more suitable), as well as non-offending parents/carers/family for up to two years. The service is provided by University College London Hospitals National Health Service Foundations Trust (UCLH) in partnership with the Tavistock and Portman National Health Service (NHS) Foundation Trust and the National Society for the Prevention of Cruelty to Children (NSPCC). It offers a joined-up approach where, if required, service users can get access to medical, social care, police, advocacy and therapeutic support ‘under one roof’ (for example, the service has two dedicated Metropolitan Police Service (MPS) liaison officers (PLOs) and two Social Care Liaison officers (SCLOs) working from the building). Levels of CSA and CSE in England and Wales Measuring the scale and nature of child sexual abuse is challenging. While administrative data sources do not represent the full scale of the issue, these sources of information can provide some indication of the resources needed to support the child protection system. In 2018/19, there were 73,260 sexual offences against children recorded by the police in England and Wales, a 3% increase compared to the previous year (70,869 sexual offences recorded in 2017/18). In 2018/19, there were 30,720 assessments following a referral to local authority children’s services in England where sexual abuse was identified as a factor and 18,720 assessments where sexual exploitation was identified as a factor. At the end of March 2019, 2,230 children (2 per 10,000) in England were subsequently the subject to a child protection plan with a primary concern of sexual abuse, a 2% increase from the previous year.8 Turning to London, in 2020/21 the MPS recorded 1,891 offences flagged as child sexual abuse, an increase of 20% compared to the previous year, and 4,357 sexual offences involving a victim aged under 16, a decrease of 18% compared to the previous year. The same year the MPS recorded 1,004 CSE offences, an increase of 55% compared to the previous year. It is unknown how much of these increases are due to an escalation in prevalence; improved police recording; or a reflection of an increased willingness to report abuse following high profile 7 NB CYP Havens continue to provide the acute/Forensic Medical Exam (FME) service. 8 Source: Child sexual abuse in England and Wales - Office for National Statistics (ons.gov.uk) 11 cases in the media. Regardless, the critical aspect is the substantial increase in demand. Given the prevalence of CSA/E, it is increasingly being recognised as a public health problem (Brown et al, 2011), impacting substantially on long-term outcomes including physical health (Bellis et al, 2014). The road to recovery following CSA/E is complex and requires specialist care and tailor- made support. However, this is not the routine provision - services are often delivered from a variety of agencies including, but not limited to: the NHS; police; and other criminal justice agencies; children’s social care; and agencies from the voluntary and community sector; making care disjointed. Concerns were raised in 2015 by the Office of the Children’s Commissioner, as only around one in eight victims of sexual abuse come to the notice of statutory authorities (Children’s Commissioner for England, 2015). In response, NHS England (London region) commissioned the “Review of Child Sexual Assault Pathway for London”, to map the various pathways for CYP following CSA. Findings from the London CSA Review highlighted variation in the available services across all London boroughs and gaps in medical provision, emotional support and the prosecution process. The report made recommendations advocating the need to establish better overall services for CYP who have experienced CSA/E (Goddard et al. 2015). A direct result of the review was the introduction of Child Sexual Abuse Hubs (CSA Hubs) across London. Designed and built on good practice, they created virtual teams of CSA/E experts in local areas. In 2016, two NHS sector Hubs were established, the first located in the North Central NHS sector and a second established in Southwest London. These provided medical and short-term emotional support for victims of CSA/E and an integrated response for the families, but the police and social care were not directly involved. However, the London CSA Review identified a better approach would be to introduce ‘Child Houses’ to London (Goddard et al. 2015). Child House Child House is the term that refers to the British model for survivors of CSA based on international best-practice. Initially developed in the US in the 1980s, the Child Advocacy Centre (CAC) model was proposed as a solution to many problems associated with standard responses to CSA, including: lack of therapeutic services; low conviction rates; traumatic investigation processes; and inter-agency conflicts (Herbert and Bromfield, 2016). Research into CACs in the US found positive results, particularly around reducing the trauma experienced by victims of CSA and improving levels of satisfaction with the overall service for both children and parents (Elmquist et al, 2015). Barnahus (Children’s House), a model used in Iceland since 1998, was inspired by the CACs (although there are some differences in service delivery between CACs and a ‘Child House’ model). Barnahus is a child-friendly, interdisciplinary and multi-agency centre where different professionals work under one roof in investigating suspected CSA cases and providing appropriate support for victims (Children’s Commissioner for England, 2016). The main components of the Barnahus model are: • A home-like setting with all services delivered under one roof; • Helping victims disclose abuse through exploratory interviewing, conducted by child psychologists; • Use the least possible number of interviews conducted by child-expert staff; • Improved evidence through the reduced need for children to testify in court; and • Guaranteed and rapid access to therapy for abused children. Compared to before and after its inception, the initial Icelandic Barnahus model was considered to yield positive results across many outcomes, such as: improved partnership working between 12 police and social services; improved therapeutic outcomes for children and their families; improvements in children’s and families’ experiences of the criminal justice process; and improvements in the quality of investigations; trebling the number of perpetrators charged; and doubled the number of convictions (Children’s Commissioner for England, 2016). As a result, the Barnahus model has since been adopted in several other countries such as Sweden, Norway and Denmark. The EU Promise project brings together research across European pilot countries and provides standards, learning and best practice.9 However, although findings across these later adaptions appear to be relatively positive, most evaluations are based in the US and the CAC model, rather than on Barnahus. In addition, previous evaluations primarily concentrate on the underlying processes, with fewer robustly assessing impact or economics. This is a considerable gap in research learning and one that the current evaluation has sought to address. The Lighthouse - a Child House for London Following the London CSA Review, NHS England (London region) approached MOPAC for support with a bid to the Home Office Police Innovation Fund, initially to pilot two Child Houses in London. The joint bid was successful, and funding was awarded in April 2016. The Child Houses subsequently formed a commitment within the new Mayor’s Police and Crime Plan (2017 – 2021), as part of keeping London safe for CYP. The Mayor and Home Secretary announced the plans for two Child Houses in September 2016, originally anticipated to open in April 2017. When deciding on service location there were several criteria used, one of the key aspects being a suitable property to house multi-agency services. Other considerations were the state of readiness of the NHS Sector with regards to seeing Child Abuse as a high priority, and a wider prevalence of CSA in those areas. Based upon these, a decision was made to develop services in the North Central London NHS Sector and the Southwest. However, it became apparent there was insufficient funding to run across both sites. A decision was made to proceed with one location, making it possible to include the enhanced staffing levels; an extension of the service to 18-25-year-olds with additional needs; extended opening hours; and consideration of accepting neighbouring sector/out of sector referrals. In deciding the one location - again, criteria were employed (i.e., need; strategic alignment of the wider health community; existing clinical leadership; demand projections; premises availability) and it was decided that North Central London was best place to proceed with the pilot. Although based in a Camden property, the service would take referrals from Barnet; Camden; Enfield; Haringey; and Islington. This geography would form a coherent area and the boroughs presented a substantial demand – with 2016/17 MPS data indicating a total of 683 victims of sexual offences aged 17 or under. Borough level data showed each area had seen a steady increase in MPS CSA/E figures, with Barnet the largest increase of 61% in yearly totals. Enfield the highest levels overall and Islington and Camden had consistently the lowest levels of CSA among the North Central boroughs. In February 2018, the commissioned contract was awarded to the University College London Hospitals NHS Foundations Trust (UCLH) and their sub-providers, brought in to deliver specialist elements of the service, namely - the Tavistock and Portman NHS Foundation Trust; the NSPCC. Delivery of what was initially to be a two-year pilot (subsequently extended by 18 months) was to follow a six-month mobilisation period. The Lighthouse provides enhanced support to CYP aged between 0 – 17 years old (or those between 18-25 years of age with learning delay or disability for whom a child or young person- oriented service appears more suitable), as well as non-offending parents/carers/family for up 9 https://www.promise-project.eu/ 13 to two years. It offers a joined-up approach where, if required, the CYP can get access to all medical; practical; social care; police; and therapeutic support ‘under one roof’, although the CYP Havens continue to provide the acute/Forensic Medical Exam (FME) service. Figure 1 below provides a summary of the service provided by the Lighthouse. Figure 1: the Lighthouse model Included within the Lighthouse model above are a number of roles which are self-explanatory (such as Paediatricians, Play Specialists and Mental health professionals), however there are other unique roles at the Lighthouse: - The CYP’s Advocates support the service users throughout their time with The Lighthouse, ensuring that the voice of the child is heard by professionals within and outside of the Lighthouse. They can be involved at any stage of the pathway including: a show around prior to first appointment, work with the young person outside of the Lighthouse before attending, support on the day of their evidential interview following Achieving Best Evidence (ABE) guidance, at the initial assessment (IA), advocating with school, social care and the police; and finally, a key role in support throughout the police investigation, court preparation and support during the trial. The advocates are trained Independent Sexual Violence Advisors (ISVAs). - Police Liaison Officers (PLOs) bring criminal justice expertise in child abuse investigations to the Lighthouse and assist with the flow of information between organisations at policy and individual case level. They also manage the Psychologist- led (PLI) interview process and provide safeguarding advice to outside agencies. - Social Care Liaison Officers (SCLOs) triage the referrals that come into the Lighthouse, ensuring all appropriate information has been received from the referrer. They also act as a go-between for Lighthouse and borough social care teams and provide expert advice and support in safeguarding to the local teams. The Lighthouse pilot set out to meet the following outcomes: • Enhanced referral pathways into and out of the Lighthouse, • Enhanced CYP, family and carer experience of support received post disclosure, Enhanced CYP experience of the criminal justice process post disclosure, • Enhanced mental health and well-being outcomes for CYP, 14 • Enhanced professional awareness, competence and confidence, • Increased likelihood of charge or conviction for those cases within the Lighthouse, • Enhanced partnership working, and • To provide CSA victims with care and support to reduce the long-term impact of victimisation. The Evidence and Insight (E&I) Unit is MOPAC’s in-house social research and analytical team which has been commissioned to undertake a multi-year mixed methods evaluation of the Lighthouse. This is the fourth and final E&I evaluation report 10 and in presenting closing results seeks to provide valuable learning for many interested parties, related to not only the Lighthouse, commissioners, but wider areas such as CSE, as well as playing a significant contributor to the national and international evidence base. Methodology This section sets out the methodology used within the evaluation. Overall, a broad action research perspective was utilised (Avison et al. 1999). That is, findings are continuously fed back to the programme teams, the academic advisory group, to the official Partnership Oversight Board and other relevant meetings to ensure learning is continually shared as part of an active feedback loop. To note, this has led to many practical improvements being made to the service during the evaluation timeframe and could well be seen as best practice in how an evaluation can actively support a programme over the duration. The evaluation used a mixed methodology approach – balancing qualitative context from staff, stakeholder or service user feedback, with ‘harder’ performance figures indicating how the service is running on a day-today basis. It focuses on four distinct areas; performance insights; process; impact and economic analysis. See Figure 2 below for the overall timeline for the evaluation. 10 The previous evaluation reports can be found on MOPAC’s website at: https://www.london.gov.uk/sites/default/files/childhouse_jan19_report.pdf https://www.london.gov.uk/sites/default/files/2019_117_childhouse_2nd_evaluation_report_for_publication.pdf https://www.london.gov.uk/sites/default/files/childhouse_nov_2020_interim_evaluation_report_for_publication.pdf 15 Figure 2: Evaluation timeline The methodology behind each of these will now be expanded upon. Performance management data Performance data is a key ingredient in understanding the basic delivery of the Lighthouse – that is - what has been delivered, when, and to whom. The evaluation team had several data sources to contribute to the performance insights covering the entire length of the evaluation (end of October 2018 to March 2021). Firstly, aggregate data was provided by the Lighthouse Data Officer which included total referrals by month, referring borough and referral source, as well as the age categories and gender of the CYP. The second, and more comprehensive, data source was individual-level data produced from Excelicare (Lighthouse’s case management system [CMS] devised specifically for the service). This data was for all clients who consented to have their data used in the MOPAC evaluation and included detailed information on their demographics, vulnerabilities and risk. Additional data on service activities (such as open cases to services, onward referrals, numbers of consultations and meetings with professionals and service users) was also provided by the Lighthouse through their quarterly monitoring returns which date back to October 2019. Overall, this data was able to provide the evaluation with an excellent description of the core programme details and cohort over time. Qualitative fieldwork undertaken during the evaluation Over the duration of the evaluation - a key aspect has been the capture of the voices of those individuals that can provide insights into the implementation and working of the Lighthouse. This has ranged from practitioners, key stakeholders, to family members and the young people themselves. As the evaluation timeline above presents, this has spanned multiple years so to explore not just the general perceptions, but whether these perceptions have changed during the life of the programme. This capture over time was especially valuable in understanding the entire life of the programme – such as the initial set up, delivery, the impact of COVID-19, as well as exploring views towards sustainability and future working. March 2018 March 2021 Dec 2018 Sept 2019 Oct 2018 April & May 2019 May 2021 April & May 2018 Sept 2018 Feb & Mar 2020 July 2020 Nov 2020 Feb & Mar 2021 CMS data Evaluation report – 1st interim Evaluation report – 2nd interim Evaluation report 3rd interim Final evaluation report Key stakeholders & Lighthouse board member interviews Telephone interviews with Police Officers April 2021 Barnardo’s Tiger Light Interview Lighthouse staff attended an induction to the service, training and focus group Lighthouse staff Focus groups and interviews Lighthouse Focus groups and interviews Lighthouse virtual focus groups Professional stakeholder online survey Professional stakeholder online survey Virtual focus group – social care CPS Staff interviews Interviews with North East London CSA hub Lighthouse virtual focus groups Parent interviews Lighthouse service goes live 16 In total, the evaluation over the 3-year period has captured the views of over 200 individuals through a variety of methods. This has consisted of: Staff focus groups and interviews The details of the staff focus groups and interviews are as follows: • In October 2018, all Lighthouse staff attended an induction to the service and training. Afterwards, staff were asked to participate in a focus group to explore their views of the Lighthouse; how ready they felt to open the doors to the public; training needs; and potential challenges going forward. In total, 13 staff members attended from a potential 27 practitioners. • Focus groups and interviews were held in April and May 2019 with Lighthouse staff to explore their perceptions of the early implementation of the Lighthouse. Data were collected from 23 individuals (all but one staff members); 15 people attended one of two focus groups, and in-depth face to face interviews were undertaken with 7 practitioners, and written feedback to the interview questions was obtained from a practitioner with whom it proved impossible to arrange an interview. The staff who participated were from a range of occupations and included paediatricians, advocates, psychologists, admin staff, SCLOs and PLOs. • Focus groups and interviews were held between July and August 2020 to explore perceptions of the implementation of the Lighthouse: undertaken face to face and virtual interviews and focus groups with 14 members of Lighthouse staff from several teams including senior management, Letting the Future In (LTFI), PLOs, SCLOs and health staff. • Seven virtual focus groups were undertaken between March and May 2021 with approximately 25 staff, including the following: the medical team, LTFI, Children and Adolescent Mental Health Services (CAMHS), a combination of SCLOs, PLOs and administrative staff, and advocates. Two further focus groups were conducted with CAMHS staff, one to specifically discuss the work that had been undertaken in the development of the parents’ group, and another with the CAMHS staff who had developed their model of working. Stakeholders and external agencies • In April and May 2018, during the design phase of the pilot 18 face-to-face interviews were undertaken by E&I and RedQuadrant with key stakeholders and Lighthouse board members. These interviews focused on the history of the project; procurement; commissioning; design specification; and initial implementation. To supplement the interviews, all programme board members received a short online survey to capture opinions around the design and initial implementation of the Lighthouse. In total, 13 people responded. • In September 2018, immediately prior to go-live, an online survey was distributed to professional stakeholders (police, charity workers, mental health practitioners, etc) who worked in the five Lighthouse boroughs with CYP who might have experienced CSA/E. The survey covered themes such as the respondent’s confidence in identifying and addressing CSA and CSE, and knowledge and awareness of the Lighthouse service. A total of 54 people responded from a range of occupations (Police Officers (39%, n=21), 17 Nurses (15%, n=8), Victim Charity workers (15%, n=8), Mental Health practitioner (11%, n=6) and Other (20%, n=11)). The survey was repeated in April and May 2019 asking similar questions about CSA/E, awareness of the Lighthouse, and experience of the service provided by the Lighthouse if applicable. This time there were 75 respondents, predominantly from schools and the police but also some medical professionals. • In July 2020 a virtual focus group was undertaken with representatives from social care from 4 of the boroughs covered by the Lighthouse and in February and March 2020 telephone interviews were undertaken with 13 police officers and written email feedback obtained from a further 2 officers. The focus of this work was to specifically explore perceptions of the SCLO and PLOs. • Between February and March 2021 Crown Prosecution Service (CPS) staff working with the Lighthouse or on Lighthouse cases were interviewed to ascertain their views on the service, gauge the level of CPS-Lighthouse interaction and reveal any concerns about the service when it comes to its dealings with the CPS. Overall, 6 interviews were conducted; two with CPS staff from the Strategy and Policy Directorate, 2 with CPS lawyers who managed colleagues working on Lighthouse cases, and 2 with CPS lawyers that had worked on Lighthouse cases. Interviews were conducted using Microsoft Teams because of COVID-19 restrictions. • Two semi-structured interviews were conducted with staff from the North East London (NEL) CSA hub in February 2021 and the Barnardo’s Tiger Light programme between April and June 2021. These enable the evaluation to offer qualitative reflections from a Business As Usual (BAU) CSA service. Service users and parents • A vital aspect of the evaluation was the capture of the voices of the young people themselves that experienced the Lighthouse service first-hand. This work was undertaken on E&I’s behalf by staff from the Safer Young Lives Research Centre at the University of Bedfordshire11. In total, 11 young people were interviewed about topics such as their general perceptions of the Lighthouse (aspects of service delivery and physical space), the benefits of the model and its impact, and suggested areas for improvement. The interviews took part between May and July 2021. 28 children and young people were put forward by Lighthouse staff as potential participants. Risk and needs assessments were carried out for all. 17 of the 28 young people risk and needs assessed were not subsequently interviewed, for various reasons. All the young people who have taken part in the interviews are females aged 15+. • Virtual semi-structured interviews were undertaken in April 2021 using Microsoft Teams with 4 parents of children who had attended the Lighthouse, three of whom had been engaged with the Lighthouse’s Parents’ Group. The parents were asked for their perceptions of the service they and their child/children had received, what they thought had worked well, and areas for improvement. 11 The University of Bedfordshire have produced a separate report which provides further detail about the methodology they developed for the service user interviews, and the findings from the young people themselves. The report (‘’There’s something there for everyone’- Learning about the Lighthouse: Young people’s perspectives on London’s Child House’, Beckett et al 2022) is available on the university’s website. 18 Attempting to understand impact of the Lighthouse A unique aspect of the current evaluation - when compared with comparable evaluations of such programmes - is the examination of a range of outcome measures that seek to explore the impact of the Lighthouse across a variety of different measures (i.e., criminal justice, health and well- being) comparing these to a BAU service provided to victims of CSA in another area of London (the CSA hub in NEL). This section outlines the broad approach to exploring impact. It should be stated up front that exploring ‘impact’ herein seeks to triangulate across data sets - both in terms of the service delivery, as well as some core relevant outcomes. This was necessary because many of the traditional means of developing a counterfactual were not feasible (i.e., Randomisation was not ethical or an option). Evaluating such a locally enhanced service offer also presents challenges – as no other service is directly comparable to the same degree. To counter this, efforts were made to select the most appropriate BAU service for CSE as a broad counterfactual.12 As will be outlined, whilst a unique feature, exploring impact herein was not without limitations and these are acknowledged (i.e., sample sizes, timescales, particularly in relation to the criminal justice outcomes, impact of the COVID 19 pandemic) – and these should be borne in mind - but never-the-less - the results can be seen as indicative of what the Lighthouse was able to deliver, at what cost and how this compares to a BAU service and is able to provide valuable insights that could be built upon in future research. The comparison group selected was NEL and the evaluation team engaged with NEL Clinical Commissioning Group (CCG) to gain an understanding of the provision there for victims of CSA/E. The services engaged with included Barnardo’s Tiger Light, 13 and a service provided by a Paediatrician. NEL was selected as a comparison area for the Lighthouse due to its geographical proximity, and that the CSA Hub in that area was more established in comparison to other areas of London. The Lighthouse and NEL are compared across performance and outcome data. These cover: - Throughput, number of referrals, referral sources – all drawn from performance data; - Service provision and staffing details drawn from qualitative interviews with managers; - Service user information: age, gender, ethnicity, disability – all drawn from performance data; - Mental health and well-being measures; and - Criminal Justice Outcomes (i.e., charge, conviction, wider attrition through the system). Most of the above measures are relatively straightforward, taken from routine performance data. However, it is worth outlining more detail on the Criminal Justice Outcomes. As outlined by previous E&I evaluation reports,14 one of the methods of impact was to undertake in-depth and resource intensive case study coding on these core elements related to Lighthouse service. Given the resources required to conduct the coding, sample sizes are modest. However, this ought to be adequate for the development of insights that when paired with wider performance and outcome data can be triangulated to generate insights. 12 C&YP who experience an acute sexual assault in London generally attend the Havens for immediate treatment and support, including a forensic examination. The Lighthouse is being set up to improve the services delivered to C&YP who are victims or survivors of non-recent CSA/E. The aim is to bring together a range of services delivered by a specialised multi-disciplinary team with outreach to a wide range of services. 13 https://www.Barnardo’s.org.uk/what-we-do/services/tiger-light-nel 14 Conroy et al. (2018) 19 For Criminal Justice outcomes - Lighthouse cases were selected during 2019 – the rationale being that this time period enabled the programme to have had a bedding in period whilst also allowing time for cases to progress. All Lighthouse cases in 2019 were eligible if consent was given. This resulted in a total of 139 cases identified. From these 53 were deemed inappropriate for a variety of reasons,15 leaving a final cohort of 86 cases to track relating to CJ outcomes. The evaluation sought to identify equivalent CSA/E within NEL boroughs at a similar timeframe (i.e., cases reported to police in 2018 and 2019). To generate a comparison group, three separate queries were run on MPS systems (the Crime Record Information System – CRIS). The first was to search for victims aged under 18 and where the classification was one of rape, sexual assault, penetration female, penetration male, and CSE. The second search was to repeat the search with the same classification types, but with the victim aged 18-25 and with a recorded learning disability. The third search was to search again for victims aged 17 and under, using CSA/E to search instead of classification types (and then any duplicates from the first search were removed). These three separate searches were conducted to capture as wide a sample as possible to draw a representative sample of the referral criteria of the Lighthouse (i.e., a victim of CSA/E aged up to age 17, or aged 18-25 with a learning difficulty). Together, these searches generated a total sample of 2,955 cases. As per the Lighthouse sample, those inappropriate were removed resulting a sample of 1,869. Given the in-depth nature of the coding, a modest sample of 80 were randomly selected for the final coding. Comparisons were made between the sample of 80 and the wider sample. The evaluation team were satisfied with the similarities between them in terms of demographic and case type (see Appendix A for breakdown). Of these, a further 13 were excluded as there was no criminal investigation. Therefore, the final comparison sample for the analysis is 67 cases. In terms of these two groups, the Lighthouse and counterfactual were broadly comparable across a number of demographic and offence criteria. Overall, the Lighthouse group were slightly younger (a mean of 11 vs 14 at time of reporting); gender proportions were similar (Lighthouse 80% female vs NEL 88% female); no significant differences across ethnicity (the Lighthouse split of White and BAME victims was equal at 50% each, and for NEL there was a split of 56% white and 44% BAME). In terms of the case characteristics, there were some differences such as Lighthouse cases being more familial abuse (41% of cases compared to 19% for NEL) and NEL had a higher instance of peer on peer and stranger abuse. Full details can be found in Appendix B. However, the evaluation was not looking for an exact statistical match, rather, the aim was to identify who would have been likely to be eligible for a Lighthouse service within the new area. This is a subtle but important difference given how local demographics and offending profiles can vary. The coding itself involved in depth reading of the police cases (both Lighthouse and comparison) across 70 variables covering: victim characteristics; suspect characteristics; offence characteristics (the circumstances of what happened); procedural characteristics (the police response and investigation); and outcomes.16 A coding frame was devised, piloted and four coders completed 15 37 were deemed inappropriate due to the type of classification where there was no investigation (e.g., Crime Related Incident, Child Care Issue, No Crime, Non-Crime CSE, Rape not Confirmed, or Rape Transferred). A further 16 cases were excluded because the dates of the criminal justice investigation did not overlap with referral and support from the Lighthouse, therefore there was no opportunity for the Lighthouse to have an impact on the outcomes. 16 Appendix A presents the variables coded and the descriptive characteristics for the full samples. 20 the in-depth work, with supporting quality and reliability checks overseen by regular meetings. Analysis focused upon comparing group differences on the key outcomes of interest. Economic analysis Economic analysis was commissioned out and completed by RedQuadrant. They provided an indicative cost-benefit analysis (CBA) which involved exploring the benefits of the Lighthouse against the costs associated with setting up and delivering the Lighthouse. This was then compared within a difference-in-difference analytical approach estimating the differential effect of the Lighthouse on its service users. The focus of the economic analysis is on the comparison of Lighthouse against standard forms of provision in London, in particular those considered in Harewood and Baine (editors) (2018) “London Child Sexual Abuse Learning Report”.17 This report examined three models funded by the London CSA Transformation Programme, all aiming to support CYP after experiencing CSA. These models were (1) the Children and Young People’s Haven Service (CYP Havens) based within London’s Sexual Assault Referral Centre, (2) the Child House and (3) CSA hubs.18 The economic analysis involved RedQuadrant: - Collating evaluation data on key outcomes at baseline and after treatment from the Lighthouse; - Collating data on expected levels of improvements for these outcomes from the existing literature on the after-effects of CSA/E for victim-survivors; and - Reviewing the improvement for Lighthouse versus that identified by the literature. The intended benefits are wide-ranging and can be categorised in a number of different ways. Some of the expected benefits are likely to be evident in the short term, others will not manifest themselves for many years – until the victim or survivor has reached adulthood, in some instances. There were three broad areas of social value set out: wellbeing to the client, useful savings from public sector spend and additional public sector spend on essential activity. The model sought to identify the benefits that accrue. The following domains of social value were considered: - Health: This comprises sexual health, physical health and substance misuse. - Wellbeing: This looks at wellbeing from the perspective of the NHS and Local Authority, the individual child or young person, and the family. - Children’s services: The report looks at both the possible impact on child protection action and on the need for any additional school support. - Employment: CSA/E often has significant impact during adult life; the report considers the possible impact of the Lighthouse on loss of earnings and take-up of benefits. - Criminal justice: The benefits examined are the costs of enforcement action against alleged perpetrators, and the impact on possible criminal activity committed by victims of CSA/E during adulthood as a result of their abuse. - System effects: Better co-ordination of multi-agency services at the Lighthouse has a tendency to lead to improved productivity. The focus of the analysis is on those categories of benefits whose effects are identifiable, measurable, both in the short term as well as long term and these have been used in turn to 17 https://www.england.nhs.uk/london/wp-content/uploads/sites/8/2019/05/London-CSA-Services-Learning-Report-2018- v1.2-002.pdf 18 The analysis undertaken uses data taken from both the Havens and a CSA hub (North East London). This was dictated by the quality of the cost data available. 21 provide an extrapolation to those benefits. The CBA model identifies the benefits that accrue compared against two hypotheses: (a) treatment under non-specialist services; and (b) under a partly integrated approach. 22 2. Performance insights A core aspect of the evaluation was to monitor the performance data underpinning the Lighthouse. Performance insights have been generated at frequent intervals throughout the pilot and have enabled the evaluation to monitor overall throughput to the service, the internal workings and processes, activities, client demographics and a closer look at clients’ vulnerabilities. This section will present the findings of this analysis over the lifetime of the pilot – from October 2018 to the end of March 2021 - with insight into how these aspects may have changed or developed over time. Such data is crucial in understanding what was actually delivered, whilst also laying the key foundations that will enable considerations around impact later in the report. The data in this section will cover the following aspects: - Numbers of referrals - Referrals by month - Referring borough - Referral source - IAs at the Lighthouse - Client demographics – for those consented - Client vulnerabilities and disabilities – for those consented - Offence and criminal justice overview – for those consented - Service delivery and activities – for all clients Referrals to the Lighthouse Between the end of October 2018, when the Lighthouse launched, and the end of March 2021 there were a total of 889 referrals to the service, which works out as an average of around 30 referrals per month, slightly lower than originally predicted. Out of the total referrals – the majority were female (82%, n=730), with over half in the older age group between 13-17 years (57%, n=505) – something that was a consistent finding throughout the evaluation. Further detail on demographics of the cohort is found later when exploring those who consented to the evaluation. Between Year 1 (end of Oct 2018 and Oct 2019) and Year 2 (Nov 2019 – Oct 2020) there was a 20% decrease in total referrals from 399 to 320, however 31 of the first year referrals had originally been made to the CSA Hub, and which were subsequently transferred to the Lighthouse. The remaining decrease in referrals in Year 2 could be attributed to the Covid-19 pandemic where there was a notable decrease in referrals in April and May 2020, to 14 and 19 respectively, where up until that point the service had seen an average of 31 referrals per month. Since the end of Year 2 (November 2020) the average number of referrals has increased to 34, and March 2021 saw the highest number of referrals ever seen at the Lighthouse in one month (n=45; see Figure 3 below). 23 Figure 3. Total number of referrals, IAs, number of services users who consented to the evaluation. Looking at the entire 29-month period of the Lighthouse, all 5 boroughs referred a similar proportion of CYP to the service, with only 3 percentage points between Haringey who referred the most (21%, n=184), and Camden who referred the fewest (18%, n=158) CYP. There were very few referrals outside of the core 5 boroughs or from outside of London. Looking at the pattern of referrals over time, with the exception of Barnet, all other borough referrals reduced from Year one to Year two. This was likely influenced by CV-19. See Table 1 below. Table 1. Referrals to the Lighthouse by borough, by year (Note - Year 3 is year to date) Borough Year 1 (end of Oct 2018 - Oct 2019) Year 2 (Nov 2019 - Oct 2020) Year 3 (Nov 2020 - March 2021) Total Percentage Haringey 82 67 35 184 21% Barnet 76 78 25 179 20% Enfield 71 65 40 176 20% Islington 73 60 33 166 19% Camden 84 42 32 158 18% Other London Borough 13 3 4 20 2% Outside of London 0 5 1 6 1% Total 399 320 170 889 Children’s Social Care has consistently been the largest referral source over the length of the pilot, referring over half of all CYP to the service (52%, n=460), whereas medical sources (GPs, hospitals and sexual health clinics) made up only 6% (n=59). These patterns have remained consistent over the entire duration of the Lighthouse (see Table 2). 0 5 10 15 20 25 30 35 40 45 50 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 2018 2019 2020 2021 Referrals Initial assessments Consented 24 Table 2. Referral’s source to the Lighthouse Referral source Referrals Percentage Children's Social Care 460 52% Police 107 12% Other 72 8% Self-Referral 57 6% CAMHS 37 4% CYP Havens 33 4% CSA Hub 31 3% GP 29 3% Hospital 26 3% School 18 2% Third Sector Provider 15 2% Sexual Health Clinic 4 0.4% One of the aims of the Lighthouse was also to provide timely support to CYP, and the Lighthouse collected information on time taken from referral to support being provided or allocated. Overall there was an average of 2 days between referral to the Lighthouse and an intake meeting19 (as expected, due to the latter being twice-weekly), and an average of 30 days from intake meeting to an IA. Initial assessments at the Lighthouse Lighthouse referrals are discussed at twice-weekly intake meetings – with the SCLOs playing a key role within this process. For those referrals accepted, an action plan will be made which includes assigning the case to the primary case holder or lead practitioner and the CYP are then invited in for their IA at the Lighthouse (either face to face or virtually) where they meet with a number of Lighthouse practitioners to assess needs and decide on care plans.20 There are many reasons why a referral did not reach an IA – for example, lacking information or CYP/parent not wanting the service or not feeling ready. Between the end of October 2018 and the end of March 2021 the service carried out 510 IAs. Therefore, the overall conversion rate of referral to IA overall for the evaluation period was 57%.21 After October 2018 (where 1 IA was conducted)22, the level of IAs remained at an average of 18 per month and varied between 7 and 25 a month. There was a large drop in IAs in April and May 2020 (9 and 7 IAs respectively) due to the building being temporarily closed following the Covid- 19 lockdown. During this time, practitioners at the Lighthouse conducted IAs virtually using an NHS system called Attend Anywhere. After those two months, the number of IAs recovered well, although between Year 1 and Year 2 there was still a 23% decrease (n=57). See figure 1 above. 19 Previously called allocation meetings 20 On average there are three professionals present at each IA (this data relates only to those service users who consented to providing their data for the evaluation), most frequently a Paediatrician (for 91%, n=327) for a medical assessment, followed by a CAMHS worker (present at 48%, n=172 of IAs), and an Advocate (47%, n=170). 21 It is difficult to interpret whether this referral rate is appropriate or not. For the 379 referrals that did not reach an IA (yet), this could be for many valid reasons - pending further information; the CYP and/or parents/carers did not want the service; they did not meet the criteria; they received a consultation to the professional network; they attended for a VRI; they may be pending CYP availability, or the CYP are not feeling ready yet. There is somewhat of a delay between referrals and IAs and the conversion to IA may occur in subsequent months. 22 The service opened to referrals on 23rd October, therefore it was unlikely to start seeing children until after at least a week in order to give them enough notice to attend. 25 A deeper look - background of clients who consented to evaluation So far, the report has presented wholesale numbers on total throughput, but additional detail was available for those individuals whom consented for the evaluation. The performance analysis will now turn to this group. Over the entire timescale of the evaluation, there was a consent rate of 71%23, which equates to 360 clients out of the 510 clients who reached IA. This consent rate for the Lighthouse is good in comparison to a systematic literature review which found an average 65% consent rate across 43 studies involving schoolchildren24. This section will provide a summary of the demographics and vulnerability for the service users who consented for their data to be used in the evaluation. Consented cohort demographics The demographics for the 360 service users who consented were very similar to the overall cohort of the Lighthouse, something that has remained consistent over time: the majority were female (85%, n=306); an average age of 13; nearly half were between 13-17 (n=171, 48%). Ethnicity data was recorded for 301 service users and there was a fairly even split between BAME and non-BAME clients (n=161, 53% and n=140, 47% respectively; see table 3 below); and broadly mirrors the same splits seen across the same age groups in wider London25 (48% White and 52% BAME). Most service users are in full-time education (81%, n=285/354), half of which are in secondary school (n=139). There are 76 CYP recorded as having issues with non- attendance (63 of which are aged 14 and over). Table 3. Ethnic breakdown of service users Ethnic groups Service users Percentage White 140 47% Black 70 23% Mixed White and Black 31 10% Asian 23 8% Other Mixed 20 7% Other Ethnic group 10 3% Mixed White and Asian 7 2% Client vulnerabilities A variety of assessments are completed by Lighthouse staff, often at the IA stage – for example, the Adverse Childhood Experiences Questionnaire (ACE-Q). Staff will also assess and record current vulnerabilities of the children (such as depression, anxiety, eating disorders, drugs and alcohol) whilst also taking into account any history of self-harm or previous attempted suicide. Practitioners will also assess any future risk such as risk of suicide, self-harm, or risk to others. These are not all asked of every client and are recorded across different areas of Excelicare - so sample numbers on vulnerability data varies. However, one aspect is clear - looking across these datasets in summation the evaluation sees, as expected - a highly vulnerable client group. 23 The consent rate improved considerably since the start of the project, where the consent rate was just 24% for the first four months. The evaluation team worked closely with the Lighthouse staff on this issue (i.e. conducted a training session around consent and the evaluation and provided visual aids for staff to use with clients to explain consent). 24 Blom-Hoffman et. al (2009). 25 Wider population data derived from “2016-based Round of Demographic Projections Local authority population projections - Housing-led ethnic group projections November 2017”. © GLA 2016-based Demographic Projections, 2017. 26 The Adverse Childhood Experience Questionnaire (ACE-Q)26 was completed for 222 clients27 whose ages ranged from 1 to 25. For these service users, the ACE – with scores ranges from 0-9 – had an average of 3. The majority (n=193, 87%) of service users had a score of at least 1, and two in five (n=88, 40%) had a score of 4 or more. The most common ACE for the Lighthouse service users was sexual abuse (n=83, 82%),28followed by parents divorced/separated (n=102, 46%; see Appendix C for the prevalence of all ACEs). Putting these figures into context: Lighthouse ACE scores are above non-clinical national populations29, but lower than other at-risk youth populations.30 Moving into the wider data collected at the IA stage – again, to note this data is not routinely captured on all clients. In terms of vulnerabilities – 84% (297/354) were reported to present at least one type of vulnerability (between these 297 cases they presented a total of 793 vulnerabilities with an average of 2.7 each). 201 (68%) of these service users have at least 2 types of vulnerability. Among the most frequent categories of vulnerability were anxiety and/or depression (60%, n=121), followed by history of domestic violence (55% DV, n=110) and education problems (41%, n=82). A fifth of service users (39%, n=78) had a history of self-harm. Additionally, 22% of clients were recorded as having a disability (n=78/354)31, with 23 (6%) having more than one. Mild (n=20, 26%) or moderate (n=19, 24%) learning difficulties were the most common forms of disability. A full breakdown of service user vulnerabilities and disabilities is presented in Appendix C. Moving to a different dataset - risk assessments conducted by staff – again, there is a variety of valuable information captured to further the knowledge on vulnerabilities (again, the sample size varies as it is based on confirmed responses to each aspect). Risk assessment data illustrates information such as whether a service user is currently known to children’s social services (n=65% 219/337), or whether they have previously been known to children’s social care or early intervention support services (n=61%, 192/314).32 The assessments also show that 39 of 322 service users are subject to care order or child arrangement orders. Finally, there is data calculating future risk which further highlights client vulnerability, whereby the Lighthouse worker assessed if there is further risk to the CYP or others (see table 4 below). This was captured for 239 of those service users that consented. It is apparent that many of the service users were assessed as presenting a risk; 33% (n=82) of suicide, 46% (n= 110) self-harm, 60% (n=144) of further abuse and 10% (n=23) a risk to others. 26 The ACE-Q is an internationally validated self-report tool encompassing 10-items across 10 areas which cover household dysfunction (parental separation/divorce, parental domestic violence, parental substance misuse and mental illness, and parent incarceration), child abuse (sexual and physical), and child neglect (emotional and physical). The ACE questionnaire has been used internationally and the original ACE study used a 10-question tool (Felitti et al. 1998), however the ACE-Q has sometimes been adapted by other organisations or researchers and has either been shortened or lengthened in terms of the number of items (Bethell et al. 2017)The more events that a person experienced before the age of 18, the higher their ACE ‘score’ will be, and literature demonstrates that the higher the score (a maximum of 10) the greater the risk of health issues (i.e., mental or physical), substance misuse, victimisation and offending in adulthood. This emphasises the importance of providing holistic, integrated support to these young people to mitigate the risk of health and lifestyle problems in adulthood. 27 ACE-Qs were not completed for all clients due to an issue at the start of the pilot where there were issues as the service adjusted to using a new patient record system and ways of working. 28 NB previous caveat refers. 29 As a baseline, a nationally representative survey of adults in the UK found that 46% of respondents reported at least 1 ACE, and 8% reported at least 4. This was undertaken by Bellis et al. (2014) with 3885 18-69-year olds in the UK. There have also been many other studies, in various populations and nationalities, which have also shown that most adults (between 52%-75%) have experienced at least one ACE (Zarse et al, 2019). 30 In a study that looked at vulnerable young people with mental health problems in Scotland (who present serious harm to others), there was a much higher prevalence of ACEs; 93% (out of 130) had experienced at least 1 ACE, and 59% had experienced at least 4 (Vaswani, 2018). In another study which looked at a sample of 64,329 juvenile offenders in Florida: 87% reported at least two, 71% reported at least three, 51% reported at least four, and 31% reported five or more (Baglivio et. al, 2014). 31 Disability information was not recorded for the whole sample. 32 Please note that the base size differs due to gaps in data collection. 27 Table 4. Risk assessment details Risk Ongoing and potential immediate risk Some risks/concerns None Total Further Abuse 19 (8%) 125 (52%) 95 (40%) 239 Self-harm 20 (8%) 90 (38%) 128 (54%) 238 Suicide 8 (3%) 74 (31%) 156 (66%) 238 Risk to others 2 (1%) 21 (9%) 208 (90%) 231 Offence details Looking at the Excelicare data for the 346 clients who consented to the evaluation, the majority (n=317, 92%) made a disclosure of the offence, most commonly to someone in their family (n=169). Other individuals disclosed to were someone in their school (n=33), a health service (n=21), Police (n=18), social services (n=18). For 29 out of the 346 clients, a suspicion of CSA was reported. The most frequent offence was intra-familial sexual abuse (n=166), followed by peer-on-peer sexual abuse (n=90).33 112 CYP were recorded as being repeat victims. Delivering the Lighthouse service This section of performance insights now turns to the total activities and services delivered in aggregate (i.e., not just those that consented) at the Lighthouse. It should be noted that much of this data (provided by the Lighthouse’s quarterly monitoring returns) only became available from October 2019 (not Oct 2018 when the Lighthouse went live). Over the period examined, data indicates a considerable amount of service delivery (see Appendix C for a quarterly breakdown of the data) – to illustrate: • Between Oct 2019 and March 202134 there were 4780 telephone, video or face to face sessions. This breaks down as 3442 telephone sessions Following CV-19 (i.e., April 2020 to March 2021) there were 829 sessions delivered via video call, compared to 509 face to face sessions within the same timeframe (and only 15 face to face appointments between April and June 2020). • 936 professional meetings (excluding strategy and consultation meetings). • There were 29 psychologist-led ABE interviews that took place at the Lighthouse (averaging at 1.6 per month).35 Over the same period there were 43 police led ABE interviews at the Lighthouse.36 • The Lighthouse made 91 onward referrals to local services, most commonly within the voluntary sector (34%, n=31), followed by Social services (26%, n=24). • Between April 20 and March 21, 137 strategy discussions that took place and 118 consultations delivered by the SCLOs.37 33 No distinction was made between rape cases or other child sexual abuse cases. 34 This data comes from the Lighthouse’s quarterly monitoring returns which only dates back to Oct 2019, not Oct 2018 when the Lighthouse went live. 35There were more psychology led ABEs planned between April and June 2020, but a number were cancelled due to family anxiety about travelling in for an interview during COVID lockdown. 36 The Lighthouse is contracted to offer 3 psychologist-led ABEs a month. 37 A key part of the SCLO’s role is also to provide consultations, attend strategy meetings and offer signposting. The SCLOs conduct many consultations for several reasons and mainly with local Social Care teams, but other organisations in the community as well (such as healthcare and police). Amongst the specific reasons given were: advice to CAMHS (around referral to local authority, need for an immediate Multi-Agency Safeguarding Hub referral); consultation with Lighthouse staff about liaison with social worker around the young person’s entitlements; consultation and written support in drafting a letter to the local authority 28 Between April 2020 and March 2021, within any given quarter there were between 489 and 695 open cases to Lighthouse workers (note that there will be duplicates of CYP open to multiple practitioners at once). The Paediatrician and the CAMHS team had the largest proportion of overall cases (each 26%, n=608 and 604 respectively). Table 5. Breakdown of cases open to Lighthouse services, quarterly from April 2020 to March 2021 Lighthouse service Apr-Jun '20 Jul-Sep '20 Oct - Dec '20 Jan - Mar '21 Total Overall Percentage Paediatrician 134 114 162 198 608 26% CAMHS 146 126 163 169 604 26% Sexual Health Nurse 65 99 148 119 431 18% Advocacy 95 104 104 125 428 18% LTFI & P&R 34 40 51 46 171 7% Play Specialist 14 20 29 37 100 4% Dietitian 1 1 1 1 4 0.2% Total 489 504 658 695 2346 One of the key outcomes for the Lighthouse was to improve the CYP experience of the criminal justice system post disclosure, through having support from an Advocate. Data indicates the advocates recorded 258 different goals for a total of 98 clients, with an average of 2.6 goals each and a range of 1-1038. Against these goals, the Lighthouse staff recorded an outcome against 195 of them: indicated 63% (n=122) goals were achieved; 27% (n=53) were partially achieved and 10% (n=20) were not achieved (it is not recorded why the goal was not achieved).39 setting out safe-guarding concerns around insufficient safety for therapeutic work; advice to Lighthouse staff around local authorities’ duties around accommodation and support for a child under the Children’s Act 1989 (due to concern about the closure of police investigation, and the suspect being at home with the victim); advice to Lighthouse staff and written support in drafting a letter to mother concerning limits of advocacy, and signposting in relation to complex housing situation. 38 The most common type of goal was relating to criminal justice support (n=65) which could mean that the service user wanted to be kept updated on the proceedings or wanted to be supported by the Lighthouse throughout the investigation, which may include help in understanding their options or to demystify the criminal justice processes. The second most common type of goal was around education and support in school (n=38), which may include getting help getting back into school, or for the Lighthouse to liaise with the school around a support plan, or help communicating the service users’ needs to the school. 39 For the remainder there was no indication of the progression of the goal although it may be because they are still in progress or data error. 29 3. Implementing the Lighthouse This section seeks to present the challenges and learning relating to the design, delivery and implementation of the Lighthouse across the entire evaluation timescale reflecting upon key learning. Key issues focussed upon are the initial set up of the Lighthouse and the maturation of the model; A focus upon the impact of Covid-19; and the value and challenges of partnership working. Overall, these themes have remained consistent across the entire evaluation having emerged across our previous reports. Initial set up and maturation of the Lighthouse A key aspect within the delivery of any programme is initial design and subsequent implementation. There is a strong evidence base upon relating to the importance of programme integrity and resultant benefits (see, for example Duwe and Clark, 2015). Whilst this is the case for all programmes, it could be viewed as especially important for large scale, expensive and innovative pilot programmes – such as the Lighthouse. Overall, implementation of the Lighthouse can be seen as a huge achievement for all partners and stakeholders involved. The set up was massively complicated across procurement, commissioning, estates, IT and mobilisation and to the credit of staff only incurred minimal delays. Given the level of underlying complexity and size of the model this is something to be celebrated. On a wider note - all staff, partners, stakeholders, service users, parents and external agencies have been very positive in terms of the vision, design, governance, set-up and partnership approach of the Lighthouse across the entire evaluation period. ‘Service itself is excellent, and the concept is brilliant, the way it was explained to us, all of the different specialists under one roof, and the close work and interaction with the police, it works very well, conceptually it’s great’ (Parent 2021) ‘The response of the Lighthouse is tailormade to what is presented and requires sorting’ (Medical team member 2021) ‘[We] have time in IA to really try to understand the family and come together to come up with a bespoke support plan which is child focused but takes the whole family into consideration as well; the first time you do it, it feels quite amazing’ (Advocate 2021) One of key elements of the Lighthouse - critical from the very beginning of set up was the need for the ‘child-friendly’ nature of the premises, and positively, CYP had been involved in early discussions about the design of the facilities and continue to be involved via a number of consultative groups established by the Lighthouse in discussions about the design of the premises and aspects of service delivery. From the outset and throughout the evaluation, feedback about the service and facilities (i.e., the appearance of the building, the quality and nature of its decoration, its atmosphere) provided by the Lighthouse has been positive from service users, staff and external agencies alike. The young people interviewed were all very complimentary about the Lighthouse, describing it as welcoming, homely, safe, less clinical and more ‘youth-led’ 30 than other services they had been to before. 40 Parents described their children’s keenness to attend the premises; police officers favourably contrasting the ‘child oriented’ video recorded interview (VRI) facilities available at the Lighthouse with those available in police stations, and CPS staff commented upon how welcoming the facilities were. Again, this was a consistent positive throughout the evaluation. ‘The building was very nice. It was all new. There was a hammock in the reception, and books. It was colourful and relaxing. Like you think ‘wow, this place is really nice’. It was very calm. There is a little kitchen with a snack cupboard. And toys for young people, and books, things for people from primary age up to my age. It was warm and welcoming and made you think you’d want to go there again’ (Young person 2021) ‘I’ve been to the Lighthouse. It's a lovely space. I physically went into the room where she would be interviewed. I know what an interview suite looks like in police stations, and the physical environment was infinitely superior to the environment in police stations’. (Parent 2021) There were a number of aspects of the original Lighthouse model that were amended or evolved in the light of operational experience and feedback from service users. Within such a large and complex initiative, this is not unexpected and exploring these facets in more detail can provide valuable insights into implementation. A number of these changes seemed to arise out of staff agency and responsiveness – that is, seeking to respond to unforeseen or necessary circumstances. For example, in April 2019 the decision was taken not to open the Lighthouse on Saturdays because of low levels of usage and feedback received from service users (as opposed to the original planned extended hours and weekend operation).41 Changes were made to the IA in response to feedback from children and parents feeling ‘overwhelmed’ that one of each team at the Lighthouse was present at the IA. Subsequently the IA met the child and family initially and then split up into separate rooms so that the child was able to meet them individually and focus on the area that mattered to them. This resulted in a more focussed – timelier – IA to which staff were far more positive about. There were many other examples where the initial model was refined – whether this was changing the frequency of intake meetings (daily to twice weekly), and retaining a consistent chair function (Consultant Psychiatrist) at these meetings to provide more time to collect the information required, ensure greater continuity overall, and to lessen the time/resource implications of the meeting, or the work conducted in Summer 2019 to develop vision and values for the Lighthouse resulting in a change to the management structure. In addition, in mid-2020 the senior leadership team was expanded to include Heads of Service: office manager, an SCLO, PLO, health team, CAMHS, LTFI/Protect and Respect (P&R) and an advocacy lead. All such changes can be seen as a highly responsive action throughout delivery of the Lighthouse. Perhaps the best illustration of the responsive nature of service development at the Lighthouse is to be found in the groupwork developed with parents and CYPs. The Parents’ Psychoeducation Course was developed 6 months into the Lighthouse’s implementation partly due to the amount of sexual abuse disclosed by parents and lack of wider support for them. 40 The CYP interviewed expressed mixed feelings about the location of the building – they liked the fact that it is relatively private and tucked away but also made it feel inaccessible/unsafe for some. 41 It was not possible to open the Lighthouse on Saturdays for the first 3 months of operation because security staff were not available at weekends. After the reduced hours the weekends were still used for activities such as groupwork, if requested by parents, and members of staff continued to attend case conferences off-site at the weekend. 31 ‘Over half of families that we see, the parent themselves has also been exposed to sexual violence, victim of sexual violence or abusive relationship, often have two clients in the room’ (CAMHS member of staff 2021) ‘There’s no spaces for parents, just a huge amount of shame and stigma. Our hope that if there was a criminal justice process going on that it wouldn’t prevent or preclude people from accessing the opportunity to meet with other people, [the] key thing that people draw out is the opportunity of meeting other people who understand and how it reduces shame’ (CAMHS member of staff 2021). As there was no appropriate model in the UK, members of the CAMHS team looked at the international literature and built a course around the work that had previously been undertaken by the Washington Coalition of sexual assault in their Circle of Hope programme.42 It was stressed that this was not a therapeutic model, but focused on the impact of disclosure, and moving from disclosure to healthy relationships.43 Much of the work in the parents’ group was about trauma caused to family/impact on family arising from the child’s disclosure, and the destructive impact of that, often inter-generational, trauma. While it was not the intention of the group to provide a space for parents for their own therapeutic recovery, one of the consequences of having the group was that it had led parents to seek support for themselves, which in turn had led to the formation of the Parents’ Discussion Group. Parents stressed the benefits of being able to talk to individuals who had been through the same experience as they had (the importance of the role played by the external consultant who had lived experience, and who had been brought in to assist with the running of the Parents Group and had subsequently joined the Lighthouse staff was specifically mentioned in this respect). In terms of the developed group work with CYP, in 2020 the Lighthouse implemented the Young Persons Feedback Group to provide CYPs with an opportunity to provide insight, guidance and suggestions related to service development. This group in turn had led to the creation of the Youth Forum and the Lantern Project. The latter, still ongoing (as of May 2021) provided an opportunity for young people to work alongside an artist to create a new lantern/lamp to be installed on the street outside the Lighthouse. The Youth Forum (also ongoing) was launched in early 2021 providing a service user participation opportunity for 14–18-year-olds who meet with Lighthouse facilitators every 6 weeks to discuss issues related to service development. The young people are paid for their time and are consulted as experts (a recent session had involved consultation with the group about the language being used on forms being developed by the Lighthouse). Finally, a Young Women's Wellbeing Group had also been launched on-line in April 2021 for young women (15-18) accessing the Lighthouse to come together and support one another. Two additional service user consultation groups (one for younger children, and one for children with disabilities) and a repeat of the art project were also in development. Feedback from the CYP interviewed indicated that those who had participated in the Youth Forum welcomed the opportunity to influence developments at the Lighthouse, and the opportunity to meet others in a similar situation to themselves. Outside of these responsive changes to the model, there were many changes that were brought about by wider issues – such as IT-related, external factors beyond the Lighthouse’s control, 42 Details of the Circle of Hope course can be found at https://www.wcsap.org/resources/publications/tips-guides/support- group-guides/parent-support-group-guide. As well as this material staff also drew on ‘a mixture of Webster Stratton, [the] Cygnet parenting course and 2 other [courses] for parents of children with disabilities that were added because there wasn’t a fit for purpose model’. 43 Parents attending the group consented to not share any details of their child’s case with each other, so as to protect any criminal justice processes underway. 32 or a combination of both. These presented staff a different type of challenge to overcome. The best example of this type of implementation hurdle can be seen within topics such as Section 28 and Live link. One of the aims for the Lighthouse was that its facilities would be used for both s.28 and Live link.44 Interviews with CPS staff indicated that they felt the Lighthouse “was so well placed for [trialling Section 28]. It fully understood the purpose of how it would fit in the process and the pathway.” While this remains the long-term ambition, and permission has been obtained from the judiciary in principle for both to take place, in practice neither has yet occurred. The national implementation of s.28 was delayed, and the technology to allow s28 was not installed at the Lighthouse until October 2020. At present the Local Implementation Team (LIT) is developing local procedures for the use of s.28 and it is unlikely that the Lighthouse suite will be used for a s.28 hearing before September 2021. Similarly, although permission from the senior presiding judge for the use of the Lighthouse as a Live-link site was received in the summer of 2020, and the technology was installed by October 2020, no applications to use the Lighthouse’s facilities for a case have yet taken place. The use of psychologist-led rather than police-led interviews could be seen as another delay out of the control of staff – and staff interviewed suggested that this process had been more difficult and taken longer, than anticipated, as in practice it wasn’t a simple transfer of skills. In the early months of the Lighthouse service, one of the PLOs had spent much of their time training the psychologists. A contributory factor to the delay was the psychologists’ clinical work and “full caseloads” reducing their availability to conduct or practice VRIs.45 What the above serves to illustrate are the challenges in implementing such a complex and innovative programme, and the extent to which practice at the Lighthouse has evolved over the two-and a half year since its inception. To an extent this is unsurprising, bearing in mind the complexity of the model. Indeed, staff commented that the Lighthouse’s pilot status meant that it, and they, had the latitude to work in different ways, and to develop practice in the light of what they observed, or in the light of feedback provided by service users. Notwithstanding the above where there were changes to the proposed implementation, there were other challenges during maturation. Throughout the evaluation staff spoke frankly across a range of issues related to building design or IT, such as soundproofing, poor wi-fi connectivity, layout of medical rooms, and the limitations of CMS; Excelicare (i.e., the absence of an appointment system from the latter has been particularly criticised). These issues caused staff some frustration – but it should be noted – they sought out many proactive solutions around many of these issues (i.e., walking without shoes on, or not booking two adjacent therapeutic rooms at the same time). 44 The Youth Justice and Criminal Evidence Act 1999 (YJCEA) introduced ‘special measures’ which aim to help vulnerable and intimidated witnesses give their best evidence in court and relieve some of the stress associated with giving evidence. Special measures apply to prosecution and defence witnesses, but not to the defendant and are subject to the discretion of the court. Amongst the special measures Section 28 enables the recording of evidence and cross-examination prior to trial away from the court room and includes all child witnesses. This evidence is then played during the live trial, which, in most cases, means the vulnerable person does not need to attend in person. Another special measure is Live link (again, available for vulnerable and intimidated witnesses) which enables the witness to give evidence during the trial from outside the court through a visual link to the courtroom. The witness may be accommodated either within the court building or in a suitable location outside the court https://www.gov.uk/government/news/section-28-for-vulnerable-victims-and-witnesses-in-crown-courts https://www.cps.gov.uk/legal-guidance/special-measures. 45Feedback obtained from the CPS suggested that their views on PLIs were positive, with respondents suggesting they were of higher quality. However, it was also suggested that they were longer compared to police-led equivalents, leading to longer charging decisions. The increased length was attributed to the work being in its early stages, and the fact that techniques by the psychologists to build rapport with the CYP were recorded in the VRI, but not in the police equivalent. It was also suggested that the fact that psychologist-led VRIs at the Lighthouse were undertaken with police present might also contribute to their length. Feedback from police officers about PLIs, where expressed, was also generally positive. 33 Over the duration of the pilot, remedial work to try and address many of these concerns was undertaken (something highlighted in previous evaluation reports). Improved soundproofing in the form of carpeting, and better doors and grills to minimise sound carrying between rooms/floors were installed. In addition, the decision to add more desks to the mezzanine level at the Lighthouse, sacrificing the 1st floor staff rest-area meant noise carrying from the staff kitchen down to the Lighthouse’s reception on the ground floor ceased to be a problem, and increased desk capacity at the same time. The recruitment of a dedicated data officer to assist with the operation of the CMS and provide training for staff was also highly beneficial, and there were changes to the functionality of the CMS during the pilot’s lifetime, and more are about to take place, but even with these efforts, concerns about soundproofing, telephony (poor wi-fi reception) and the CMS remained to a degree. Overall, the pilot was well implemented. Staff, partners, children & young people (CYP) and parents were positive about the general service. As expected, given the size and scale of the Lighthouse, the pilot experienced several implementation and maturation challenges – some of these were required and indeed instigated by staff, others were due to wider factors that staff had to work around, and these resulted in changes to the model and often staff not working in a way initially envisioned. However, staff responded well, and although some aspects were not totally addressed, many were resolved over the duration of the pilot. Implementing during Covid-19 It is not possible to consider issues around implementation and maturation without discussing the impact of the CV-19 pandemic and wider societal response on delivery of the Lighthouse. Following lockdown in March 2020, the Lighthouse closed, apart from allowing occasional urgent ABE interviews and medicals (which were also available at UCLH), although staff stressed that ‘most people chose to wait’ for a session or support. The Lighthouse would not reopen until June 8th 2020. However, before this reopening a host of work was undertaken by staff to enable and embrace virtual working. Understandably, the focus at the beginning of lockdown was upon staying connected with the child or young person rather than delivering therapeutic work (i.e., at this time is was not possible to offer medical examination inspections). For example, intake meetings continued but moved to video conference with triage and consultation for new referrals. IAs were offered as ‘Virtual IAs’ with the CYP and family meeting the full Lighthouse IA team on an NHS video consultation facility called ‘Attend Anywhere’46. Medical history details were taken but examinations were delayed until reopening. Unsurprisingly, as shown in the performance analysis, the number of referrals to the Lighthouse and IAs declined in the months following Lockdown but recovered relatively quickly. During this time there was an increase in telephone contact and ad hoc virtual appointments to keep in contact with families and staff were positive around the attendance at such virtual appointments. Meetings between professionals took place on Microsoft Teams – and attendance was felt to be good, with GPs and CAMHS attending. Generally, under lockdown there was an increased frequency in the number of virtual team meetings. CAMHS set up a daily meeting with the SCLOs to provide safe-guarding advice because of concerns about the impact of Covid-19 restrictions on the CYP they were managing in terms of the mental health, isolation and increased anxiety. Heads of service met initially 3 times a week, although the frequency of the meetings subsequently decreased to twice a week and then once a week. 46 Although concerns were expressed by staff that the quality of the virtual IA was poorer than that offered in person. 34 The Lighthouse reopened for face-to-face IAs from 8th June 2020 onwards and VRIs and IAs took place at the Lighthouse regularly, together with limited therapeutic sessions (the latter tends to be a mix of virtual and face to face work). The introduction of social distancing at the Lighthouse here also meant that working practice had to change (consultations with masks on, restricting the number of people in rooms). Some CYP remain on remote appointments only. In addition, the parent group course that began in July 2020 was undertaken virtually. At the time of writing, May 2021, the Lighthouse operates a ‘hybrid’ model with some activities being undertaken at the Lighthouse as before, and others being offered virtually – such as Teams online meetings, or in combination. Similarly, while some members of staff are back working at the Lighthouse, others remain working at home, working alternates between weeks at home and in the Lighthouse, providing services face to face and virtually as circumstances demand. When considering these changes brought upon as a response to CV-19, this was a mixed picture. There were many benefits identified to the virtual methods – less time was taken up with travel, especially with external agencies (for example, the SCLOs mentioned they were attending more consultations and strategy meetings when working virtually), staff were able to attend more meetings. Staff also mentioned that attendance by professionals at virtual meetings had improved. In addition, the introduction of virtual consultations meant youths who lived further away or were not able to travel to the Lighthouse, could receive a service. The virtual consultation also allowed the CYP to ‘test the water’ in terms of what the Lighthouse offered, prior to attending in person. It was also suggested that virtual work suited some children, particularly teenagers, and one parent suggested that the advent of the virtual parent group meeting allowed people to ‘open- up’ in a way that had not been the case for face-to-face meetings. However, on balance, the positives provided by virtual working appeared to be outweighed by a range of negatives. These covered both therapeutic challenges and some wider reflection on impact upon the youths or themselves. In terms of therapeutic challenges, feedback from across those interviewed indicated that some CYP did not want to engage virtually or were uncomfortable accessing services by video call or phone. In addition, staff reflected that working virtually made it more difficult therapeutically to gauge how the young person was reacting versus working face to face. The virtual working resulted in the reduction in the use of measures of stress and anxiety (for example the Trauma Symptom Checklist for Children and the Revised Children's Anxiety and Depression Scale) due to difficulties completing detailed evaluation questions over video link due to reduced ability to engage with the child, issues around confidentiality in their own home for CYP when answering, and shorter appointments generally. Staff had tried to arrange more frequent but shorter contacts with CYP. This had also had an impact on waiting lists and waiting times for services as staff had not been able to close cases, and pick-up new ones, as previously. ‘I haven’t been able to do all the work I could have done as it hasn’t felt safe to do virtually and go into the depth required. In order to contain and appropriately assess how children are doing and their mental health and emotional need it needs to be face-to-face, also working from home has been really challenging for lots of reasons which impacts your capacity to work with young people virtually’. (LTFI worker 2021) 35 ‘If I’m assessing risk of harm and disclosures it’s hard when I won’t visibly see that young person. I felt like I was retraining myself to do this work virtually’ (LTFI worker 2021) Staff also reflected upon the wider impact of CV-19 on CYP’s mental health, for example; outlining that young people may not being able to access alternative support (youth clubs, for example), or that youths were more isolated, and more reliant on Lighthouse support subsequent to CV-19. Furthermore, the impact that remote working had on their own morale and mental health was also raised by staff, especially dealing with traumatic and difficult subjects (often) within their own homes, especially where there was ‘no commute to distance yourself from work’. For other members of staff (i.e., the administrative team) the ability to work from home had not been an option as they were expected to maintain a presence at the Lighthouse, and this had been a source of frustration. Implementation of the Lighthouse was heavily disrupted by the CV-19 pandemic. In response, considerable changes were made in order to deliver the pilot – and to the credit of staff the service continued. Indeed, many of these virtual changes became embedded into the routine way of working, but overall staff felt that virtual working whilst necessary, was lacking therapeutically compared to face to face. The value and challenges of partnership working A common theme across many largescale programmes, indeed incorporating child advocacy centres and Barnahus are the manifest benefits of partnership working (Herbert & Bromfield, 2016, 2017; Landberg & Svedin, 2013). Throughout the evaluation - results were very clear – feedback was highly positive around the many benefits of the partnership model employed at the Lighthouse. One of the advocates said that she had been drawn to the Lighthouse from her old job because of its multi-agency approach. As she remarked ‘I see what I thought would happen actually does happen in practice’. Specific aspects focussed upon were the benefits of co- location, joint working (within the Lighthouse and across external agencies, for example the establishment of case-progression meetings with the CPS) and information sharing, with benefits for both the service itself (i.e., speedier conversations, access to varied agencies, wider skills and disciplines) and the service user (i.e., quicker access, joined up support). Feedback from the service users at the Lighthouse reflected on the benefits of having access to different professionals in the one place – the sense of a ‘wrap around’ service and a team of people helping them. ‘I was told there was a nurse there. And therapy was always there. It was a safe space. [Was that good to have things like a nurse there?] I think it’s great. If I had any problems – there would be a solution for most things. Like if you’d headaches, the nurse could help. Everything being under one roof is very reassuring, which you need when you are anxious’ (Young person 2021) The benefit of not having to repeat themselves was consistently raised by service users, as was being able to get all the help they needed in one place. Service users made specific reference to getting support with criminal justice processes (including doing the VRI at the Lighthouse) and the benefits of advocates being able to liaise with others as needed; and being able to get (specialist) health support rather than having to go to a community GP. ‘Just yesterday I was doing an appointment with a young person and could link in with the health team......not being in the same building I don’t think I would have been able to get her to 36 a health team, that’s enough effort on her part to go to the health team, and where is there a set up where they can access consultant paediatricians? And being able to build up that relationship with doctors, especially in Protect and Respect where prior to the Lighthouse I would spend so much of my time sitting in sexual health clinics with them, going to meet social work and police with them travelling around it’s amazing they can have that all under one roof’ (LTFI worker 2021) Similarly, this was also acknowledged by parents in terms of the partnership benefits. “through my discussions with (the advocate) because at times she hasn't known how, how to help. There's a sort of in-house consultant, child and adolescent psychiatrist too, so you know, sometimes when (the advocate) was like, ‘Oh my goodness, I don't quite know what to do and I don't know how to help, and I'm not a mental health professional’, she has liaised with those services and even though they don't know my child, they come back with far more helpful responses than the services who do know my child.” (Parent 2021) This was often conceptualised as a ‘Lighthouse Way’ and covered both practical staff but also Lighthouse Senior Management, which itself was expanded in 2020 across many varied Heads of Service – another positive example of partnership working. However, there are specific aspects of good practice around partnerships – and a good example can be seen within the PLO and SCLO roles. These were new, innovative, roles established within the Lighthouse and a considerable amount of work was undertaken to communicate and embed these roles within the service and wider partners. To illustrate, staff commented on how the PLO performed an important liaison role with police officers outside the Lighthouse, who commented very positively upon the position in arranging VRIs at the Lighthouse and for ‘chasing up’ and facilitating the CSA cases with which the officers were involved. ‘I have no words, they are fantastic. They are just an absolutely fantastic resource to have’ (Police officer 2020). Likewise, SCLOs were described as ‘the glue between social care and other professionals, paediatricians, psychologists within the Lighthouse’. Respondents also stressed the SCLOs’ usefulness in offering suggestions and guidance about how to navigate referral pathways, and their role in terms of quality assurance. Both of these roles offer insights and demonstrate the inherent value in such roles that specifically set out to work across partners and to a degree acted as a wider enabler of positive partnership working. When the referral comes in the fact it’s screened and triaged by police and social care so any safeguarding issues can be addressed straight away. That intake process and making decisions as a group, there’s already been a layer of additional safeguarding which wouldn’t have occurred if the Lighthouse wasn’t available. Traditionally, the police etc would have put safeguarding in, but here we can address as a group and get additional safeguarding issues addressed. In the intake process there is critical analysis of that family. (LTFI worker 2021) Whilst the benefits of partnership working were clear to all staff; as outlined in multiple previous interim evaluation reports (and wider literature) the issue of partnerships can also bring frustrations. Tensions as a result of different organisational cultures and working practices were apparent during the pilot. To illustrate, staff highlighted differences across terminology, pay, annual leave, training (determined by the individual’s employing organisation rather than being 37 standard across the Lighthouse); disagreements on relative contribution of certain aspects (i.e., was it too medically orientated), and can be illustrated by differences across many staff as to whether they saw themselves as employees of the Lighthouse or of their parent organisation. Overall, partnership working was a resounding positive to emerge from the evaluation, appearing to bring a wide range of benefits to the service as well as the clients – something supported by specific roles (PLO, SCLO) that enabled such a way of working. However, it is clear some cultural tensions were evident which were not able to be fully reconciled over the pilot duration. This serves to illustrate how such working cultures are very difficult to avoid and should be borne in mind when devising and running any similar programme. 38 4. Impact of the Lighthouse One of the most important questions behind any programme is the question of ‘impact’. That is, did the programme achieve what it intended do – and can we compare such outcomes to other groups that did not receive the service. This section seeks to explore the impact of the Lighthouse – by drawing upon a variety of data (referrals and throughput, health and wellbeing outcomes, criminal justice outcomes, and views from participants such as practitioners, and service users) that when triangulated can generate insights. The methodology section outlined the rationale behind the selection of the Northeast London CSA Hub as the comparison group for the Lighthouse. Table 6 below provides a high-level overview comparison of both services – the Lighthouse and NEL CSA Hub. For more detail on service provision in NEL please see Appendix D. As can be seen from this table, the Lighthouse covers fewer boroughs but provides a larger number of services and as such – a larger number of staff members – compared to NEL. The Lighthouse is also eleven times more expensive than NEL CSA Hub. Table 6. Comparison of Lighthouse with NEL service provision Service Lighthouse NEL CSA hub Boroughs covered Barnet, Camden, Enfield, Haringey and Islington Barking and Dagenham, Hackney, Havering, Newham, Redbridge, Tower Hamlets and Waltham Forest Services provided CSA clinic, play therapy, sexual health, Advocacy, LTFI/P&R, VRI, CAMHS, PLO, SCLO, groupwork (parents/CYP), dietician. CSA clinic (Paediatrician) at Royal London Hospital – 1 day a week (with occasional support from hospital play therapist) Barnardo’s Tiger Light EES (8 – 11 weeks of support) Number of referrals 889 CSA clinic - 90 Tiger Light programme - 141 Throughput 510 IAs Tiger Light – 120 CYP PA Staffing details Approx. 30 staff overall Delivery manager, strategic lead; clinical lead, consultant psychiatrist; paediatricians, clinical nurse specialist, play specialist, advocates, CAMHS, PLOs, SCLO, office manager, admin support, data officer. RLH Paediatrician x .25 WTE Barnardo’s 2.6 WTE Cost of service (£ PA) £2.3 million See Economic Evaluation section for further detail. CSA clinic £60,390 (staffing costs of £59,640 plus £750 annually for maintenance of colposcope) Barnardo’s Tiger Light costs £148,000 (staffing costs of £131,590 and other expenditure of £16,410) Total = £200,390. Comparison of Performance data: Lighthouse and NEL CSA hub This first section exploring seeks to compare referrals data between the Lighthouse and NEL (both Barnardo’s Tiger Light and the Paediatric services), and a comparison of service user demographics (largely as contextual setting). 39 Referrals and throughput As outlined, Lighthouse overall received 889 referrals and conducted 510 IAs. Over a comparable period, Barnardo’s Tiger Light in NEL had a total of 141 referrals to the service, most of which were accepted, working out at an average of 5 referrals per month.47 In the same time period, the Paediatrician within NEL CSA Hub received 90 referrals relating to CSA (although did receive a total of 131 to their service in that time),48 which works out at an average of 3 referrals per month (see Appendix E for breakdown of referrals by borough in NEL). In this case, the Lighthouse saw on average 22 additional CYP per month. In terms of referral sources, the Lighthouse had a much wider spread of sources (Social care, police, Self-referral, CYP Havens) whereas NEL predominant source was from Social Care (87%, n=116). See table 7 below for full breakdown. Table 7. Comparison of Lighthouse and Barnardo’s Tiger Light referral sources Lighthouse NEL Tiger Light Referral source CYP Percentage CYP Percentage Children's Social Care 196 54% 116 87% NHS 24 7% 10 7% Early Help Services 0 0% 3 2% Education 7 2% 2 1% MASH 0 0% 1 1% Police 34 9% 1 1% Youth Offending 0 0% 1 1% CYP Havens 16 4% 0 0% Other 32 9% 0 0% CSA Hub 3 1% 0 0% Self-referral 29 8% 0 0% Third sector provider 7 2% 0 0% CAMHS 12 3% 0 0% Total 360 134 Comparing the Lighthouse and NEL (Barnardo’s Tiger Light and Paediatrician) clients on demographics there are some differences. Overall NEL referrals were younger (the average age for CYP being referred to NEL services is 12 years for Tiger Light, and 8 for the Paediatrician) compared to the Lighthouse (average age of 14). NEL (combined Tiger Light and Paediatrician) had significantly more referrals in the younger age group of 12 years and under (n=159, 60%) compared to the Lighthouse (n=367, 41%). Gender proportions were comparable between the two areas, and ethnicity was overall comparable with the exception of NEL which had a significantly higher proportion of Asian CYP (20% compared to Lighthouse’s 8%; p<0.05). It is not possible to fully compare the vulnerability of the clients given data limitations in NEL – however, this is possible in one area. A total of 22% of Lighthouse clients were recorded as having a disability (n=78/354), whereas 7% (n=13/177) of Tiger Light CYP had a disability (most commonly a learning disability, n=6); this is a significant difference (p<0.05). 47 There were likely just over 200 total referrals to NEL within that time (provided anecdotally) although due to Barnardo’s reporting system it is not possible to accurately calculate those which were ‘rejected’ referrals. In January 2020 NEL switched to a new referral system where only ‘accepted’ referrals were added to the system (prior to this, all referrals included those rejected were added to the system). Reasons for referrals being ‘rejected’ were due to them not being suitable; referral declined by the CYP/family; referred elsewhere; or withdrawn by the Social Worker. 48 Other reasons for referrals to the Paediatrician were relating to abnormal genital anatomy and unrelated to CSA. 40 Service users’ perceptions A critical aspect of the evaluation was the voice of the clients themselves – and the final evaluation sought to present their views. Overall, a total of 11 CYP were interviewed to obtain their perceptions of the service they had received at the Lighthouse. While there were some areas young people felt could be improved – explanations re the Lighthouse offer; consistency of communication and clarity re information sharing, overall the young people were very positive about the Lighthouse, commenting that it had exceeded their expectations, which were typically informed by experiences of other services.. When asked to score the service out of 10, scores were all 8 plus (3 gave it 10 out of 10); there were references to ‘family’/’homeliness’; and a general feeling that it was different to other services they had been to before. They spoke positively about their engagement with different workers, including comments on how everyone (from reception on) treated them well and with respect. They described staff as genuinely caring and invested. ‘Looking back on the Lighthouse, even though obviously I wish I hadn’t had to go there, but I think they just made the experience of having to go there a lot less harder than it had to be. And yeah, I did feel like almost loved there, yeah. I guess that’s my, looking back I don’t realise, I didn’t realise at the time how easier things were made for me with the Lighthouse being there. So, yeah, and it’s just helped, you know, like yeah, no, they were able to just give me things that I’ve been able to still carry on, even though I might not need to apply them for like, just I don’t know, yeah’. (Young person 2021) Many talked about having choice and control, being able to go at their own pace. ‘Firstly, you have more control over your experience. As I have anxiety, being in control makes me relaxed. Control over your experiences, and they respect your choices. CAMHS had a very long wait list – I’ve personally experienced it – and when you finish they give your place to someone else. In the Lighthouse, its my choice. And they are very respectful of your choices. They are extremely kind. Anything they can do, they are there to help – even the very smallest things – like being stressed about your exams’ (Young person 2021) Some small details stood out for them – such as recalling being asked for their opinions about which artwork should go on the walls – making them feel they were part of the service. The importance of practicalities was also mentioned – the availability of the kitchen area and snacks, clean toilets and nice rooms. ‘Little things’ like being made a cup of tea/coffee helped to make them feel cared for. The flexibility of provision was consistently described by young people – feeling it was ok to cancel an appointment if they wanted to; not always needing an appointment to see the health team, for example, and experiencing staff as always willing to try and see them when needed. They compared this to negative experiences of missing appointments elsewhere and losing access to services. They also described the Lighthouse as being able to offer them help to ‘do life’ - with sleeping, exam stress, substance misuse etc., things that were not all directly about the abuse; ‘you get to do other stuff’. 41 In terms of specific aspects of the service at the Lighthouse, the young people were generally very positive about the counselling/therapeutic work they had received. Several reflected that it can take time to ‘open up’ and felt they were given the time and space to do this and build relationships and ways of working at their own pace. They described therapists as being ‘like a formal friend’; and having helped them in developing coping strategies; helping them make sense of their feelings; and supporting them to grow and develop as a person. Two also specifically mentioned valuing the therapy offered to their family, which helped relationships at home. [Young person was asked to select an image that described their Lighthouse experience, picked an image of seeds] ‘Because for me it represents growth, and for me like through growth in the Lighthouse I grew into a different person, like for example, and my confidence as well. And I picked this [image of sunshine] because at first you might feel like upset and down, and you might feel, oh it’s not helpful, then over time it became very helpful, and for me like going there first it was just like oh, I can’t be bothered, and then after a while I actually like was kind of excited to go there’ (Young person 2021) The advocates played a really important role for young people, sometimes in relation to support with the criminal justice process and sometimes a wider role. Advocates were described as helpful/valuable for explaining things throughout the investigation; checking in on them throughout; answering any questions they had; liaising and following up with police and others on their behalf and updating the CYP on the status of things – ‘That extra little support makes you feel very reassured’. While not all young people knew/used the term ‘advocate’ there was a sense of them as ‘their person’ and they were positive about having a single point of contact. ‘So, she was kind of like a bridge between you and the police or you and the court. And there were times where I would go and, before the trial started I went to the courthouse and I would see the rooms and stuff and she would come with me, just so it wasn’t as scary... So, it was just nice having someone, I didn’t feel like I was in the dark…You don’t really know what can happen with the court and stuff, and obviously she knew the most so she would be like, and she’d be realistic and say like, ‘The outcomes could be this, the outcomes could be like that’, and for me I found that reassuring, knowing the way the outcomes that could happen, but also still being reassured’. (Young person 2021) Longer term engagement with, and support from, the advocate when things had ended with the service was also described as important, as was the option to get back in touch with the service should circumstances change. One young person reflected positively on doing a VRI at the Lighthouse compared to doing it in a police station. The key things she noted were – the waiting space for family, the set-up of the interview space itself, the fact that the police came to you in plain clothes; her mum was in waiting room, the therapist and paediatrician were also about; and nice refreshments. Asked if she would recommend it, having done both LH and police, her response was – “yeah, yeah, 100%. It just works” Young people were also very positive about the health team, describing staff as very flexible and approachable; they emphasised how good it was to be able to have all aspects of their health addressed through them (because they understood their circumstances) rather than having to go to a GP separately – ‘Everything being under one roof is very reassuring, which you need when you are anxious.’ An example was given of ending up in hospital (outside of Lighthouse engagement) and the paediatrician and advocate coming to make sure the CYP was treated well. 42 Overall, the young people were very positive about the impact the Lighthouse service had had on their lives. A number of specific impacts were described such as increased confidence, feeling back to their normal self. When asked how she’d describe the lighthouse to an alien who had no idea what it was one service user said she would describe it as ‘somewhere that helps young people become a better version of themselves really’. Some also commented that the Lighthouse had helped them to develop short-term coping mechanisms, but also longer-term ones they could use in the future. Parents’ perceptions of the Lighthouse The evaluation sought the views of the parents who had individuals experiencing the Lighthouse. Whilst not robust ‘impact’ it is essential their voices, as well as staff and youths – are included when exploring impact. Overall parents were highly complementary about the service provided by the Lighthouse, about the information provided, and about the high levels of training of Lighthouse staff. ‘I’ve only got good things to say about it. If we didn’t have that I’ve no idea what we would have done, because there doesn’t seem to be anything similar to it, or not readily known or available even from the internet’ (Parent 2021) In terms of impact, parents stressed the benefit of the long-term holistic support provided to the child by the Lighthouse, support that they felt was not available elsewhere. One parent described what her family had received as ‘holistic family support’. Not only had the victim received support, but the Lighthouse had also provided support to a sibling, and to the parent, provided communication with the school, and had provided onward referrals for both the mother and the sibling. The parent, describing the advocate remarked they were; ‘so proactive, so when we liaise with the school, the school are actually quite good, but (the advocate) provides information, teaching information, trauma based teaching, information to the school. I'm lucky in that school, you know, picked it up and sent it to every single member of teaching staff’. (Parent 2021) Another parent said his young children; ‘[they] looked forward to it, they liked the person they were working with, you could see noticeable improvements month by month, and a lot of it was as a result of the work being done with them. They definitely benefitted. They liked the environment itself as well. They enjoyed going to that actual place’. (Parent 2021) Parents mentioned the fact that there was very little information available for parents/guardians of victims of CSA aside from that provided by the Lighthouse; ‘the education, the emotional and direct context education is amazing because you don’t know where to start’ (Parent 2021). They were reluctant to seek guidance from the internet because of concerns about what they might find, where their search might lead, and about the quality of the information they found; ‘I don’t want to get directed by algorithms to something I really don’t want to get directed towards’. Their engagement with the Lighthouse had provided them with ‘really good tools and things to use in certain situations’ in relation to their child’s behaviour and how to interpret and respond to it. 43 Parents particularly stressed the benefit of being able to talk with other people who had been through similar experiences to their own at the Lighthouse, and who could provide them with information about what to expect. As one remarked ‘maybe [the] initial understanding of Lighthouse’s purview wasn’t so focused on the parents, and that’s proved to be intensely valuable to the parents’, illustrated personally in his ability to talk to his daughter about what had happened to her; ‘Educating us to help ourselves and to help our children is incredibly valuable’. Emotional wellbeing outcomes Both Lighthouse and the comparison site capture data on emotional wellbeing (albeit somewhat differently). The Lighthouse staff used several means to measure changes in emotional wellbeing in the CYP. The CAMHS and LTFI teams at the Lighthouse set individual goals which were agreed with each service user and/or their parent or caregiver.49 The progress against each goal is then tracked. Between Oct 2018 and March 2021 there were 204 emotional wellbeing goals set for 88 service users (only for those who consented), with an average of 2.3 goals each.50 These goals were recorded as free text on Excelicare and the evaluation team coded each one on overall themes.51 Out of the 204 goals, 117 had an outcome recorded against them: 69 (59%) goals were achieved, 35 (30%) were partially achieved, and 13 (11%) were not achieved. For the remaining 87 there was no record of progression of the goal, which may be due to them being in progress or due to data error. Barnardo’s Tiger Light practitioners used set categories to measure CYP outcomes on (see Appendix E for list of the 18 measures and results). The CYP are marked on a score of 1 – 5 (where 1 = low risk and 5 = high risk), and there were 25 CYP who had at least two assessments completed to show progress against these. This dataset from Barnardo’s compares the difference between the first and final assessment. Comparing the proportions of positive outcomes between Lighthouse and NEL is problematic due to small samples as well as a lack of standardisation in data capture, in addition to the subjective nature of success. However, upon inspection, there is a significant difference (p<0.05) between the two groups – with 89% Lighthouse goals achieved or partially achieved – and for NEL there were improvements in 47% of outcomes (see Appendix for full breakdown of outcome results). Comparison of criminal justice outcomes and attrition This section now moves to the findings from the criminal justice analysis that was undertaken - a comparison of key outcomes through the CJ process. This analysis – and the examination into criminal justice outcomes - was viewed as one of the core metrics in terms of exploring impact. This relates to charges, convictions, and overall attrition through the system (e.g., victim withdrawal, police no further action [NFA]), in addition to some contextual factors within the investigation (such as numbers of ABE interviews completed and requests for third party material). 49 The Lighthouse also use measures such as the Trauma Symptom Checklist Children (TSCC) and Revised Child Anxiety and Depression Scale (RCADS), however the numbers of children with measured outcomes was too low (for those who consented; 8 children with two TSCC assessments, and 4 with RCADS outcomes). 50 This group was aged between 1 and 21 years old. The majority are female (n=72, 82%), and the majority (60%, n=105) are BAME. 51 To note, some goals were very specific and detailed. Where possible the evaluation aimed to categorise and simply the overall theme of the goal. 44 Examining the two key outcomes (i.e., charge and conviction) there was almost no difference between the two groups. The Lighthouse cases had 7% (n=6) cases charged by the CPS and a 5% (n=4)52 conviction rate; compared to NEL who had 6% (n=4) charged by the CPS and a 4% (n=3) conviction rate. These results were not significantly different and further illustrate the very low levels of charge and conviction in such cases potentially pointing to a wider criminal justice issue. Given the timeframes of the evaluation, there were a relatively small number of cases from both groups that were still ‘open’ in some manner (i.e., 14 Lighthouse and 7 for NEL)53 – and a longer timeframe would be beneficial to allow cases to progress. Table 8. Key stages of criminal justice attrition Criminal Justice Stage Lighthouse NEL Total cases 86 67 Police NFA 33 (38%) 25 (37%) Victim withdrawal 27 (31%) 30 (45%) Suspects identified 68 (79%) 49 (73%) Suspects arrested* 38 (44%) 18 (27%) Submitted to CPS* 21 (24%) 7 (10%) Charged by CPS 6 (7%) 4 (6%) Guilty pleas 2 2 Proceed to trial 3 1 Guilty at trial 2 1 Not guilty at trial 1 N/A *Significant difference between the two groups at 95% confidence level Looking at the wider Criminal Justice metrics, again generally, there were no significant differences between Lighthouse and NEL cases (i.e., police NFA, suspect identified, and guilty pleas were all comparable between the two groups). The Lighthouse cases had a lower proportion of victim withdrawal compared to the NEL cases (31% compared to 45%), although this was not a significant difference.54 However, the Lighthouse did have significantly higher instances of positive actions within the investigation, such as increased suspect arrests (44% vs 27%, p<0.05), and proportion of cases submitted to the CPS (24% vs 10%, p<0.05) compared to NEL (see table 8). Wider aspects of the investigation were also examined, and results indicate a significant increase in Lighthouse conducting VRIs (71%, n=61 Vs 33%, n=22, p<0.05).55 The Lighthouse cases were also significantly more likely to request early investigative advice (EIA) from the CPS (21% n=18, vs 6%, n=4, (p<.005); and significantly more likely to request third party material (most commonly social services, medical records and school records) (50%, n=43 Vs 24%, n=16, P<0.05). 52 One of the guilty outcomes includes two suspects within the same case (both convicted). 53 As of the end of May 2021, 11 cases are still being investigated by police (7 of which are working on CPS actions); 2 cases are with the CPS awaiting a charging decision and; 1 case has been charged and is awaiting a trial date. NEL on the other hand had five open cases at the end of May all of which the police were still investigating and of which two the police were working on CPS actions. 54 The reasons for victims withdrawing were broadly similar between the two groups. For 10 cases in the comparison group, the victim withdrew due to not supporting the allegation and not wanting to prosecute the suspect. In both groups, one of the most common reasons for withdrawing was due to the victim not wanting to go through the investigation or have police involvement (n=10 for the Lighthouse and n=9 for NEL). The remaining reasons for victim withdrawal for Lighthouse and NEL respectively were: Stress of exams (n=6 and n=2); general health reasons (n=2 and n=4); they had moved on and wanted to put the incident behind them (n=4 and n=2); victim became disengaged and reason was not clear (n=4 and n=2). Finally, for two of the Lighthouse cases the victims had a fear of repercussions of the allegation. 55 The main reasons for NEL not having a VRI completed was because the victim refused (n=15) or the police did not arrange (n=13). An MG11 may have completed instead, however this was difficult to accurately capture. 45 In summary, exploring impact involved triangulating across a range of data and compared to a BAU example; Lighthouse overall delivered more, had more positive emotional wellbeing outcomes, and although there is no difference so far in terms of charges or convictions, the Lighthouse had more positive investigative actions such as reaching the CPS for a charging decision. Both parents and CYP were highly positive regarding the service and spoke of a range of benefits. Table 9 below outlines the outcomes identified for the Lighthouse at the outset of this report and reflects on the extent to which these outcomes have been achieved at the time of writing. Table 9: Lighthouse outcomes and evidence of achievement Outcome measure Evidence of Achievement Enhanced referral pathways into and out of the Lighthouse. - The Lighthouse had a wider spread of referral sources in comparison to NEL. - Positive feedback from local borough Social Care Liaison Officers. - Evidence of onward referrals made by Lighthouse. Enhanced CYP, family and carer experience of support received post disclosure. - Positive feedback from interviews with CYP and parents on the support they received. Enhanced CYP experience of the criminal justice process post disclosure. - Positive feedback from interviews with CYP and parents on the support received from Advocates in navigating the criminal justice process. Enhanced mental health and well- being outcomes for CYP. - Indication of improved wellbeing through the proportions of goals achieved through the emotional wellbeing teams. Enhanced professional awareness, competence and confidence. - Findings from qualitative interviews with staff and external practitioners indicate excellent overall professional awareness and competence. Increased likelihood of charge or conviction for those cases within the Lighthouse. - Indicative impact showed there was no difference between the Lighthouse and a comparison group in terms of charges and convictions. However, further work is recommended in the future to enable a more robust analysis. Enhanced partnership working. - Findings from qualitative interviews with staff indicate partnership working within the Lighthouse is perceived as a large benefit; both to staff and the CYP. - Evidence of work carried out by the senior leadership team to improve partnership working. - Some challenges and tensions still remain regarding differences in organisational cultures and ways of working. To provide CSA victims with care and support to reduce the long-term impact of victimisation. - Positive findings from the goal-based outcomes of the emotional wellbeing teams. - Evidence from interviews with service users that the Lighthouse have supported them with many aspects of their lives including health and wellbeing, school, relationships, and self-confidence. - NB these are short term benefits and it’s not possible to assess the long-term impact at this point. 46 5. Economic analysis and the future The final aspect of the evaluation focusses upon economic evaluation of the Lighthouse Service, and also issues around sustainability (staff views) and future models. For the economic evaluation, MOPAC commissioned RedQuadrant to undertake this work.56 They produced a CBA which involved adding up the benefits of the Lighthouse and comparing them against the costs associated with setting up and delivering the Lighthouse. This was then compared against a comparison group and a difference-in-difference analytical approach was used, which studies the differential effect of the Lighthouse on its service users, versus a comparison group forecasted over a long-term period (30 years). An overview of the approach taken was outlined in the Methodology and in Appendix F. The overall costs of the Lighthouse pilot comprise the annual operating costs of £2.387m per year plus the one-off capital costs of refurbishment, installation of IT and infrastructure (£3.9m allocated to cover a period of eight years at £0.49m per year). This equates to £2.88m per year in total. Details are provided in table 10. Table 10. Elements of costs for the Lighthouse Cost/year (£000s) Operational staff (direct) 1,134 Operational staff (indirect) 277 Estates (ongoing costs) 380 Diagnostic Equipment, Drug costs, Clinical supplies, IT 154 Overheads (including management) and other 442 Total operating costs 2,387 Capital costs (refurbishment, one-off expenditure on IT and infrastructure spend – one- off expense of £3.95m amortized over 8 years) 495 Overall costs per year 2,881 With this overall cost, and 420 clients per year, the unit costs for the Lighthouse are of the order of £6,860. Costs of a Havens service were estimated at £4,925 per case (details of costs for the Haven are shown in Appendix F). Unfortunately, despite development of a comprehensive dataset for monitoring performance and utilisation rate, accurate data on many outcomes is not available across the sites used for this analysis. Outcome measurement is particularly challenging for CSA services due to the complexity of the recovery with every child’s journey likely to be different. This total cost can then be set against the wider social value identified across three broad areas: wellbeing to the client, useful savings from public sector spend and additional public sector spend on essential activity. Table 11 below shows the social costs for Lighthouse, plus social costs for a comparator based on the Havens, whose proportional impact was estimated to be 67%57 of that 56 A fuller explanation of the cost-benefit model RedQuadrant developed for the Lighthouse, and the research that underpins it, is contained in appendix F. 57 Based on a calculation that used ’referral on‘ rates taken from Harewood and Baine (2018) (assuming 18% for the Lighthouse and 40% for the Havens), together with an assumption of short-term support needs for the CYP cohort of 85%. This produced a figure of improving mental health of 67% for Lighthouse (85% baseline rate of referrals minus 18% referral on rate after Lighthouse intervention) compared to 45% for the Haven (85% baseline rate minus 40% referral on rate after Havens intervention); 67% of the Lighthouse figure. 47 of the Lighthouse over a 30-year period. Table 12 compares the costs and benefits for the Lighthouse against the Havens. These results show that there is a financial return to the public sector from the operation of the Lighthouse, with a net gain in public expenditure per client of £14,570. The ratio between future savings and cost of the intervention is 3.12 : 1.0 (calculated as £21,430 ÷ £6,860). However, the comparator scheme also shows a good financial return, with a net gain in public expenditure per client of £9,460. The ratio between future savings and cost of the intervention is 2.92 : 1.0 (calculated as £14,385 ÷ £4,925). The results suggest very substantial harm occurs for the victim and very substantial costs accrue to public services as a consequence of CSA/E. Per child or young person, public sector costs are estimated at some £122,000 (with the prosecution of crimes accounting for around £32,000), while the loss of wellbeing to them and their family, and the loss of earnings, is of the order of £98,000. Calculations suggest that the use of Lighthouse compared to the Haven costs perhaps £1,935 per case more (£6,860 - £4,925) but saves an additional £7,000 on future public expenditure (£21,430 - £14,385) (excluding costs on convictions) and improves wellbeing by additional £10,300 (£31,420 - £21,100). There are some limitations to the analytical approach – including data quality issues, as well as a lack of firm impact and attribution that precludes firm findings upon impact. None-the-less, the analysis explores the costs and potential benefits of the Lighthouse as well as comparison service. Table 11. Social costs (£ per client) for the Lighthouse versus Havens comparator Social cost without Lighthouse Social cost with Lighthouse Social cost with comparator Health & wellbeing Mental health 17,545 9,590 12,205 Sexual health, Physical health and Substance abuse 1,075 880 945 Client and family wellbeing 46,900 15,480 25,800 Children's services Child protection 17,585 17,585 17,585 School support 8,190 2,705 4,505 Criminal justice system Criminal behaviour by victim 7,100 2,340 3,905 Crimes prosecuted & convicted 32,110 33,720 33,190 Employment Employment losses individual) 50,875 44,270 46,440 Employment losses (HMRC) 10,170 8,850 9,285 System effects Closer co-ordination 28,625 26,910 27,475 48 Table 12. Overview of costs and benefits for Lighthouse versus comparator (£ per client) Difference between Lighthouse and Base case Difference between Comparator and Base case Reduction in public sector costs due to improved outcomes (1) £21,430 £14,385 Cost of operating intervention (2) £6,860 £4,925 Net gain to public sector expenditure (3) = row 1 less row 2 £14,570 £9,460 Reduction in loss of client and family wellbeing £31,420 £21,100 Reduction in loss of earnings £6,605 £4,435 The future As outlined in the previous section, both Lighthouse and the comparison yielded a positive financial return over the long term. Lighthouse was funded from October 2018, and funding has been extended until March 2022 – with a question mark over what happens when this initial pilot reaches a conclusion. Provision of support services for victims of CSA differs widely across London, as is shown by the material in Appendix G which summarises the early emotional support (EES) available to victims of CSA at the point the report was written. Clearly that available at the Lighthouse is far more comprehensive than anywhere else in London, which has raised issues about the equity of the service provision. There are discussions on-going between the various stakeholders (MOPAC, NHS England, the Home Office, HMCTS, NSPCC etc) to determine what future models of CSA support might comprise, and the level of support provided to victims. Perceptions of the future As is the case for time limited programmes, the potential ‘end point’ of the Lighthouse and sustainability were regularly on the minds of staff over the course of the evaluation and was a source of concern. Indeed, pilot schemes will often suffer from such uncertainty (Dawson, Stanko, 2011) and this was no different across the Lighthouse and is something that grows as the end point comes closer. ‘Many of the professionals who work in the Lighthouse are highly skilled, they will get job offers with other organisations rather than stay at the Lighthouse for the remainder of the pilot, without job security. Concerns about the loss of staff numbers if the pilot funding isn’t continued’. (Member of medical team, 2021) Staff, partners and parents were well aware of the cost and overall innovative behind the Lighthouse, but many tended to feel that this is the service that CYP need, and others spoke about the likely benefit of a wider rollout of the service and the possibility of financial savings this could bring if any rollout was budgeted and managed at a local level. Similarly, the young people interviewed felt there should be more Lighthouse services to make them more accessible and enable a greater number of young people to benefit from them. Many had already praised and recommended the Lighthouse to their friends. 49 ‘We wouldn’t take anything away. Do you want a Rolls Royce, or do you want a Skoda of a service? This service is what the children need, …’ (Member of medical team 2021) Whereas some staff considered a future without the Lighthouse and reflected on the wider needs any future services may need to accommodate, such as the integrated service, managing risks, and the balance of attempting to do all of this in house, or as a commissioning agency. ‘If thinking about the future you need an integrated service – either need a service that can do all of that and can think about the risk and the challenges (managing self-harm or suicidality) OR you have a service that provides consultation and support to the services out there which have more confidence and capacity to do that. Lighthouse is about integration and bringing people together’ (CAMHS staff member 2021) 50 6. Discussion The Lighthouse represents the most comprehensive attempt to implement a Barnahus in the United Kingdom to date.58 Its combination of dedicated child-friendly premises, the co-location of a range of specialist services, specific posts to encourage partnership working, and the provision of long-term support to CYP and family members surpasses anything that is available to victims of CSA elsewhere in the UK. Comparison of the Lighthouse to ‘business as usual’ elsewhere in London unsurprisingly showed that the Lighthouse reached a far larger number of clients and delivered far more outputs than practice elsewhere. As a result, there has been a great deal of interest in the emerging findings from the evaluation. For the latter, researchers from MOPAC’s Evidence and Insight unit have worked alongside those responsible for the design, implementation and delivery of services at the Lighthouse for the past three-and -a half years. The pilot used an action research rather than a summative approach for the evaluation and emerging findings have been fed back to staff at the Lighthouse, and to commissioners and stakeholders, throughout the pilot’s lifetime. In terms of the duration of the research, the number of interviews and focus groups undertaken with staff, external agencies, service users and parents, and attempts to compare the Lighthouse with practice elsewhere, this has been the most comprehensive evaluation of the implementation of a Child House in the UK to date. Findings from the performance data depict the volume of work that the Lighthouse achieved over the pilot, with 889 referrals being received between the end of October 2018 and the end of March 2021. Of those referrals, the Lighthouse saw 510 CYP for initial assessments either at the Lighthouse building in person or – since the CV-19 pandemic – virtually. For these 510 clients there was a considerable amount of work that was put in by staff, illustrated by: 4780 telephone, video or face to face sessions; 936 professional meetings; 29 psychologist-led interviews (not including the 43 police led interviews that took place at the building); 91 onward referrals into local services; 137 strategy discussions and; 118 consultations from the SCLOs. Within any given quarter there were between 489 and 695 open cases to Lighthouse worker. The evaluators were also in receipt of individual level data from the Lighthouse’s bespoke case management system (only for CYP who consented to provide their data, which was 71% out of the 510) which provided detailed insight into the backgrounds and needs of the service users. This data showed that the majority of clients were female, with over half in the age bracket of 13-17 years, and a fairly even split between BAME and non-BAME clients. It was clear from the data collected around the clients’ needs and risk assessments that this is an incredibly vulnerable group of CYP with the majority 84% reported at least one type of vulnerability, most commonly depression and/or anxiety, a history of domestic violence, and education problems. Overall, the evaluation found that the pilot was well implemented. Staff, partners, CYP and parents were positive about the general service. As expected, given the size and scale of the Lighthouse, the pilot experienced several implementation and maturation challenges – some of these were required and indeed instigated by staff, others were due to wider factors that staff had to work around, and these resulted in changes to the model and often staff not working in a way initially envisioned. However, staff responded well, and although some aspects were not totally addressed, many were resolved over the duration of the pilot. 58 Durham Constabulary’s piloting of a Child Advocacy Centre project in Durham and Darlington between June 2016 and April 2018, evaluated by the University of Durham, was much smaller in scale in terms of the services delivered, and was hampered by the failure to find suitable premises for the pilot (Hackett and Butterby, 2018). 51 Implementation of the Lighthouse was heavily disrupted by the CV-19 pandemic. In response, considerable changes were made in order to deliver the pilot – and to the credit of staff the service continued. Indeed, many of these virtual changes became embedded into the routine way of working, but overall staff felt that virtual working whilst necessary, was lacking therapeutically compared to face to face. Overall, partnership working was a resounding positive to emerge from the evaluation, appearing to bring a wide range of benefits to the service as well as the clients – something supported by specific roles (Police Liaison Officer [PLO], Social Care Liaison Officer [SCLO]) that enabled such a way of working. However, it is clear some tensions as a result of different organisational cultures and working practices were evident which were not able to be fully reconciled over the pilot duration. This serves to illustrate how such working cultures are very difficult to avoid and should be borne in mind when devising and running any similar programme. In terms of the impact of the Lighthouse, when measured against the objectives identified at its inception 59 evidence of their achievement was found for most but was largely reliant on qualitative data. Feedback from staff, from external agencies, from service users and from parents was excellent with regard to improved referral pathways, better support received post-disclosure, enhanced professional awareness, competence and confidence and enhanced partnership working. Longer term objectives (enhanced mental health and well-being outcomes for CYP for example, and the provision of care and support to CSA victims to reduce the long-term impact of victimisation) clearly could not be answered from these qualitative sources.60 However, overall it is clear that for those who took part in the interviews, focus groups and surveys undertaken by E&I throughout the duration of the evaluation, whether they be professionals, staff, parents or service users, they all spoke very positively about the impact of the service. Comparing two key outcomes of cases progressing through the Criminal Justice System (charge and conviction) there was almost no difference between cases from Lighthouse and the NEL cohort. The Lighthouse cases had 7% (n=6) cases charged by the Crown Prosecution Service (CPS) and a 5% (n=4) conviction rate; compared to NEL who had 6% (n=4) charged by the CPS and a 4% (n=3) conviction rate, although clearly the analysis is limited by the small sample sizes. However, when comparing investigative actions across the two groups, there were some positive and encouraging findings. The Lighthouse had significantly higher instances of positive investigative actions such as increased suspect arrests, and a significantly higher proportion of cases submitted to the CPS. Additionally, more early investigative advice was sought from the CPS for the Lighthouse cases compared to NEL; something that had been actively worked on between the Lighthouse and CPS during the pilot. RedQuadrant provided an indicative cost-benefit analysis (CBA) which involved exploring the benefits of the Lighthouse against the costs associated with setting up and delivering the pilot. It is clear that the Lighthouse is an expensive service. RedQuadrant calculated the annual operating costs to be £2.387m per year plus one-off capital costs of £3.9 million (covering refurbishment, installation of IT and infrastructure) which allocated over an 8-year period was £0.49m per year), equating to £2.88m per year in total. With this overall cost, and 420 clients per year, the unit costs for the Lighthouse were of the order of £6,860. Costs of the Havens service were estimated at £4,925 per case. 59 The objectives were: enhanced referral pathways into and out of the Lighthouse, enhanced CYP, family and carer experience of support received post disclosure, enhanced CYP experience of the criminal justice process post disclosure, enhanced mental health and well-being outcomes for CYP, enhanced professional awareness, competence and confidence, increased likelihood of charge or conviction for those cases within the Lighthouse, enhanced partnership working, provide CSA victims with care and support to reduce the long-term impact of victimisation. 60 A comparison of emotional wellbeing outcome data for the Lighthouse and NEL suggested that a significantly higher proportion of goals (89% to 47%) were achieved at the former, although there were differences between the two cohorts. 52 However, when these costs were set against the 3 broad areas of social value identified by RedQuadrant (wellbeing to the client, useful savings from public sector spend and additional public sector spend on essential activity) in the long term (30 years), the results suggested that there was a financial return to the public sector from the operation of the Lighthouse, with a net gain in public expenditure per client of £14,570. In comparison with the Havens the Lighthouse cost around £1,935 more per case, but saved an additional £7,000 on future public expenditure, and improves wellbeing by an additional £10,300. The future of Lighthouse and CSA/E CSA has been identified as a key priority by the government, MOPAC and other agencies. The Home Office’s CSA strategy was published in January 2021 and stated that the government would ‘help victims and survivors of recent and non-recent child sexual abuse to rebuild their lives by improving the support available and developing and embedding best practice….. we will raise awareness of support for victims and survivors and provide local commissioners with the resources to meet their needs wherever they live in the country’ (Home Office, 2021, para 20). The Lighthouse is specifically mentioned in the strategy document which states that the Home Office will build on the learning from the Lighthouse pilot phase, to publish ‘guidance for local commissioners and service providers seeking to introduce ‘Child House’ models of support to victims and survivors of child sexual abuse, and will consider how national and local funding can support the development of similar local initiatives’ (para 265, Home Office 2021). The anticipated publication of the Home Office’s ‘Child House; Local partnerships guidance’, and MOPAC’s ‘Child House in a box’ toolkit later in the summer of 2021 indicates the level of political support. In addition, MOPAC’s Victims Strategic Needs Assessment Summary Report (undated) has identified the ‘requirement for further assessment work on the levels of needs regarding CSA, alongside a review of the Lighthouse model of provision to develop a scalable operating model that can provide a consistent quality intervention across [London]’ and that MOPAC should consequently ‘prioritise the funding, commissioning and resourcing of services where violence is present (inclusive of; CSA, Sexual Violence, Criminal Exploitation, Serious Violence and Domestic Abuse)’61. Clearly the findings from this evaluation, will play a crucial role in future discussions about the delivery of CSA. Limitations / caveats Notwithstanding the above, it is important to identify the limitations of the current evaluation, and the factors that have had an impact on it. In terms of the evaluation methodology it has proved difficult to establish a satisfactory comparison site in London for the Lighthouse. The service provided at the Lighthouse for victims of CSA is different from that available anywhere else in terms of facilities, extent and duration. Attempts to compare the Lighthouse against ‘business as usual’ is problematic when the latter varies markedly across London. In terms of the performance analysis the NEL CSA hub has been used as a comparison, although clearly the two are very different in nature (partly the justification for the comparison), and in the economic analysis, a combination of data from NEL and the Havens has had to be used – something which is obviously less than ideal, but which reflects the situation on the ground, and the availability of data. 61 MOPAC Victims Strategic Needs Assessment Summary Report. Finding 1: Managing Demand - Expenditure is not aligned to victim need, and Finding 2: Violence is expected to increase and with it the profile of most vulnerable and affected victims is likely to change. 53 The case management system, whilst useful in many regards, was still unable to provide an accurate understanding of the individual Lighthouse services that were received and accessed by each client, and the duration of these services. Additionally, although the consent rate for evaluation was fairly good (71%), it should be borne in mind that this does not represent the full cohort of service users. Similarly, with regard to the analysis of criminal justice outcomes, the current delays that are affecting the progression of cases through the courts has obviously had an impact on the number of cases from the Lighthouse available for analysis, and the sample size that has had to be used (n<100) is not ideal. Consequently, the results contained in this report should be viewed as indicative rather than definitive, and there will be the need to undertake follow-up research around criminal justice outcomes once more time has elapsed, and more data becomes available (see below). With regard to the economic analysis, the timescale available for the evaluation precludes the ability to look at the actual impact of the Lighthouse on the welfare and health of service users and family members in the longer term. Hence the alternative – the use of findings from what were felt to be relevant research, and the attribution of these benefits to the Lighthouse. Clearly the findings arising from this analysis are open to critique, but this report has tried to be transparent in terms of the assumptions made. In terms of the findings around the implementation of the Lighthouse the situation has been complicated by the outbreak of the Covid-19 pandemic in Spring 2020 and the impact this had and continues to have on service delivery at the Lighthouse. In certain respects (the introduction of virtual working initially, and the development of a hybrid model consequently) Covid has led to the development of a model which was never anticipated, and elements of which are likely to be retained in the future. However, from an evaluation perspective the situation has been extremely problematic as there has been little continuity in terms of service delivery pre and post Covid. In addition, the impact of the changes on service users and on staff at the Lighthouse should not be underestimated. Final thoughts The Lighthouse evaluation has spanned three-and a half years and 3 interim products – and the current report brings the evaluation to a close. One of the aims was to deliver the largest evaluation of its type in the country covering aspects of process, performance, impact and economic value. Overall, this has been achieved and a substantial body of learning has been generated. Indeed, this manner of evidence-based working – supporting a major programme with long term and in-depth evaluation throughout its lifespan, enabling ongoing decisions to be evidence based is a positive approach to both evaluation as well as programme delivery and one that should be more frequent across criminal justice. The Lighthouse will continue to be funded in its present form until the end of March 2022, and there are specific aspects of its work which have yet to come to fruition, and with which there is likely to be continued interest. Indeed, as outlined within the report, the analysis on impact in particular would benefit from a longer analytic time period – and this is something that will be explored in terms of understanding the issue of impact more robustly. 54 References Avison, D. E., Lau, F., Myers, M., D and Nielsen, P. A. (1999) ‘Action Research’. Communications of the ACM, January 1999, Vol. 42 No. 1, Pages 94-97. Baglivio, M. T., Epps, N., Swartz, K., Huq, M. S., Sheer, A., and Hardt, N. S. (2014) ‘The prevalence of adverse childhood experiences (ACE) in the lives of juvenile offenders’. Journal of juvenile justice, 3(2), 1. Barrett, A. and Kamiya, Y. (2012) ‘Childhood Sexual Abuse and Later-Life Economic Consequences’. IZA Discussion Paper No. 6332. Beckett, H., Soares, C. and Warrington, C. (2022) ‘’There’s something there for everyone’- Learning about the Lighthouse: Young people’s perspectives on London’s Child House’. University of Bedfordshire, ISBN: 978-1-911168-15-7 Bellis, M., Lowey, H., Leckenby, N., Hughes, K. and Harrison, D. (2014) ‘Adverse childhood experiences: Retrospective study to determine their impact on adult health behaviours and health outcomes in a UK population’. Journal of Public Health, 36, 1, 88 – 91. Bethell, C.D., Carle, A., Hudziak, J., Gombojav, N., Powers, K., Wade, R., Braveman, P. (2017) ‘Methods to assess adverse childhood experiences of children and families: toward approaches to promote child well-being in policy and practice’. Blom-Hoffman, J., Leff, S. S., Franko, D.L., Weinstein, E., Beakley, K. and Power, T.J. (2009) ‘Consent Procedures and Participation Rates in School-Based Intervention and Prevention Research: Using a Multi-Component, Partnership-Based Approach to Recruit Participants’. School Mental Health, 1(1) 3-15. Brown, J. O’Donnell, T. and Erooga, M. (2011) ‘Sexual Abuse: A Public Health Challenge’. NSPCC. Children’s Commissioner for England, (2015) ‘Protecting children from harm: A critical assessment of child sexual abuse in the family network in England and priorities for action’. London. https://www.childrenscommissioner.gov.uk/wp- content/uploads/2017/06/Protectingchildren-from-harm-executive-summary_0.pdf. Children’s Commissioner for England (2016) ‘Barnahus: Improving the response to child sexual abuse in England’. London. Conti, G., Morris, S., Melnychuk, M., and Pizzo, E. (2017) ‘The Economic Costs of Child Maltreatment in the UK, a preliminary study’. NSPCC. Conroy, L., Hobson, Z., Parker, R. and Read, T. (2018) ‘The Lighthouse: London’s Child House. Initial Evaluation Report’. MOPAC Evidence & Insight. Crown Prosecution Service. Therapy: ‘Provision of Therapy for Vulnerable or Intimidated Adult Witnesses’. Available at: https://www.cps.gov.uk/legal-guidance/therapy-provision-therapy-vulnerable-or-intimidated- adult-witnesses. 55 Daro, D. (1988) ‘Confronting child abuse: research for effective programme design’. New York: Free Press. Duwe, G., and Clark, V., A. (2015) ‘Importance of Program Integrity’. Criminology & Public Policy, Vol 14 (2). Elmquist, J., Shorey, R.C., Febres, J., Zapor, H., Klostermann, K., Schratter, A., and Stuart, G.L. (2015) ‘A review of Children’s Advocacy Centers’ (CACs) response to cases of child maltreatment in the United States’. Aggression and Violent Behavior, 25, 26-34. Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A. M., Edwards, V., Koss, M.P., Marks, J. S. (1998) ‘Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study’. Fergusson, D., McLeod, G. and Horwood, J. (2013) ‘Childhood sexual abuse and adult developmental outcomes: Findings from a 30-year longitudinal study in New Zealand’. Child Abuse and Neglect 37. Goddard, A., Harewood, E. and Brennan, L. (2015) ‘Review of Pathway following sexual assault for children and young people in London’. Kings College Hospital NHS Foundation Trust, NHS England (London). Hackett, S. and Butterby, K. (2018) ‘An independent evaluation of the Child Advocacy Centre (CAC) Project in Durham and Darlington’. Durham University and Child Advocacy Project. Hailes, H., Yu, R., Danese, A. and Fazel, S. (2019) ‘Long-term outcomes of childhood sexual abuse: an umbrella review’. Lancet Psychiatry 6, pp. 830-39. Harewood, E. and Baine, H. (2018) ‘Healthy London Partnership, London Child Sexual Abuse Services Learning Report’. NHS England. Heeks, M., Reed, S., Tafsiri, M. and Prince, S. (2018) ‘The economic and social costs of Crime’. Second edition. Research Report 99. Home Office. Herbert, J L and Bromfield L (2016) ‘Evidence for the Efficacy of the Child Advocacy Center Model: A Systematic Review’. Trauma Violence and Abuse 2016 Vol 17 (3) 341-357. Herbert, J L and Bromfield L (2017) ‘Better Together? A Review of Evidence for Multi- Disciplinary Teams Responding to Physical and Sexual Child Abuse’. Trauma Violence and Abuse 2017 1-15. Home Office (2021) ‘Tackling child sexual abuse strategy’. Home Office: London. Khadr, S., Clarke, V., Wellings, K., Villalta, L., Goddard, A., Welch, J., Bewley, S., Kramer, T. and Viner, R. (2018) ‘Mental and sexual health outcomes following sexual assaults in adolescents: a prospective cohort study’, Lancet. Landberg, A, and Svedin, C G (2013) ‘Inuti ett Barnahus: a quality review of 23 Swedish barnahus’. Radda Barnen. The Lighthouse (2020) ‘The Lighthouse Annual Report 2019 – 2020’. Lighthouse. Available at 56 https://www.thelighthouse-london.org.uk/guidance-and-support-for- professionals/downloads-links-for-professionals/. McNeish, D and Scott S, (2018) ‘Key messages from research on intra-familial child sexual abuse’. DMSS Research and Centre of expertise on child sexual abuse. Newberger, C., Gremy, I., Waternaux, C., and Newberger, E. (1993) ‘Mothers of sexually abused children: Trauma and repair in longitudinal perspective’. American Journal of Orthopsychiatry, 63(1), pp. 92-102. PWC (2015) ‘The Costs of Eating Disorders. Social, Health and Economic Impacts’. PWC. Sadler, S., Tosh, J., Rawdin, A., Squires, H., Sutton, A. and Chilcott, J. (2016) ‘A model to evaluate the cost effectiveness of condom distribution schemes, developed for NICE public health guidance on condom distribution schemes and sexually transmitted infections’. PHAC 21.10 draft health economics report, University of Sheffield School of Health and Related Research. Vaswani, N. (2018) ‘Adverse Childhood Experiences in children at high risk of harm to others. A gendered perspective’. Centre for Youth & Criminal Justice. Zarse, E.M., Mallory, R.N., Yoder, R., Hulvershorn, L., Chambers, J.E. & Chambers, R.A. (2019) ‘The adverse childhood experiences questionnaire: Two decades of research on childhood trauma as a primary cause of adult mental illness, addiction and medical diseases’. 57 Appendices Appendix A – Full list of criminal Justice coding framework variables and their frequencies, comparing Lighthouse with NEL Variable Lighthouse N % of sample Comparison N % of sample Victim/survivor characteristics Age at time of offence [incl. non-crime cases] Under 10 years 45 38% 8 11% 10-15 years 63 53% 46 64% 16 years and above 11 9% 18 25% Age at time of reporting [incl. non-crime cases] Under 10 years 47 34% 8 10% 10-15 years 76 54% 49 62% 16 years and above 17 12% 22 28% Gender [incl. non-crime cases] Female 115 80% 70 88% Male 28 20% 10 13% Ethnicity White 63 50% 40 56% Black 43 34% 20 28% Asian 10 8% 11 15% Other 11 9% 0 - Has mental health issue* 58 42% 14 21% Has learning difficulty 20 14% 9 13% Previous victimisation of sexual assault/DA 19 22% 7 10% Suspect characteristics Age at time of offence [where known] Under 18 years 33 41% 27 48% 18-30 years 15 19% 16 25% 31-40 years 6 8% 5 9% 41-50 years 15 19% 5 9% 51-60 years 9 11% 3 5% 61+ years 2 3% 0 - Gender Female 5 6% 5 7% Male 84 94% 62 93% Previous history of domestic abuse 11 13% 2 3% Previous history of sexual assault/rape 9 10% 3 4% Previous history of other offending* 19 22% 5 7% Relationship to victim/survivor Stranger* 12 14% 18 27% Current or former partner 9 10% 7 10% Friend or Acquaintance [incl. family friend] 27 31% 25 37% Familial* 35 41% 13 19% Professional or Carer 1 1% 1 1% 58 Offence characteristics Location Victim’s dwelling 14 16% 6 9% Suspect’s dwelling 14 16% 10 15% Shared dwelling of victim and suspect 19 22% 8 12% Other private dwelling 13 15% 4 6% Outside 0 - 15 22% Not domestic 19 22% 16 24% Offence recorded or photographed 8 9% 9 13% Procedural characteristics Time taken to report Same day 15 17% 12 21% 1 – 7 days 15 17% 24 41% 1 week – 1 month 21 24% 6 10% 1 month – 6 months 15 17% 6 10% 6 months – 1 year 2 2% 2 3% Over 1 year 18 21% 8 14% Reported to police by Victim/survivor 13 15% 15 23% Third party 71 83% 51 77% Third party description Family 35 48% 20 38% Social Services 23 32% 10 19% Education 10 14% 13 25% Support services 0 - 7 13% Medical 4 5% 0 - Other 0 - 2 4% Number of OICs One 5 7% 12 18% Multiple 81 94% 55 82% Victim/survivor attended Havens 17 20% 12 18% Video recorded interview completed* 61 71% 22 33% Multiple video recorded interviews 13 15% 13 19% Location of 1st video recorded interview Police station 39 64% 19 86% Lighthouse 19 31% n/a - Other 2 9% Location of last video recorded interview Police station 5 38% n/a - Lighthouse 7 54% n/a - Psychologist-led interview 9 15% n/a - Police identify perpetrator 68 79% 49 73% Perpetrator arrested* 38 44% 18 27% Perpetrator interviewed under caution +3 32 37% 21 31% Early investigative advice sought from CPS* 18 21% 4 6% Early Evidence Kit administered 13 15% 12 18% No forensic opportunities 47 55% 31 46% Body Worn Video footage 12 22% 10 15% Request for Lighthouse material made 17 20% n/a - Request for other 3rd party material made* 43 50% 16 24% 59 Police reference delays in obtaining 3rd party material 16 19% 6 9% Police reference high workload 18 21% 7 10% Police reference delays due to Covid-19* 16 19% 4 6% Victim referred to support services n/a - 14 21% 60 Appendix B – Comparison of NEL criminal justice cases sample with final comparison group Table 13. Breakdown of Comparison group CRIS classifications Wider comparison sample of 1869 Final comparison sample of 80 Classification Number of cases Percentage Number of cases Percentage Sex Ass F 520 28% 21 26% Sex Ass F U 13 255 14% 9 11% Rape F U 16 223 12% 11 14% Pen F U 16 155 8% 7 9% Rape F O 15 154 8% 8 10% No Pen F U 16 94 5% 7 9% Sex Ass M U 13 81 4% 5 6% Rape F U 13 80 4% 3 4% No Pen F U 13 54 3% 2 3% Sex Ass M 50 3% 2 3% Pen F U 13 36 2% 0 0% Pen Fem 28 1% 1 1% Rape M U 13 20 1% 2 3% No Pen M U 13 18 1% 0 0% Pen M U 16 18 1% 0 0% Pen M U 13 13 1% 0 0% A/rape F U 16 11 1% 0 0% No Pen Fem 10 1% 0 0% No Pen M U 16 9 0% 0 0% A/rape F O 15 7 0% 0 0% A/rape F U 13 7 0% 0 0% Rape M U 16 6 0% 0 0% Rape M O 15 5 0% 0 0% Pen Male 4 0% 1 1% A/rape M U 13 2 0% 0 0% A/rape M U 16 2 0% 0 0% No Pen Male 2 0% 0 0% Pen Fem 1 0% 0 0% Rape MUO FU13 1 0% 0 0% Rape MUO FU16 1 0% 0 0% Sex Act U 13 1 0% 1 1% Sex Act U 16 1 0% 0 0% Total 1869 80 Table 14. Breakdown of Comparison group gender Wider comparison sample of 1869 Final comparison sample Gender Number Percentage Number Percentage Female 1624 87% 70 88% Male 240 13% 10 13% Undisclosed 2 0% 0 0% Total 1866* 80 *(NB: The base numbers for demographics vary due to data completeness in CRIS) 61 Table 15. Breakdown of Comparison group ethnicity Wider comparison sample of 1869 Final comparison sample Ethnicity Number Percentage Number Percentage White 839 51% 40 56% Black 441 27% 20 28% Asian 336 21% 11 15% Other 23 1% 0 0% Total 1639 71 Table 16. Breakdown of Comparison group age groups Wider comparison sample of 1869 Final comparison sample Age groups Number Percentage Number Percentage 12 and under 534 29% 23 29% 13-17 1311 70% 55 69% 18 and over 24 1% 2 3% Total 1869 80 62 Appendix C - Supplementary Lighthouse performance tables Table 17. Prevalence of ACEs (out of 222 CYP) Type of ACE No. of CYP Percentage Sexual abuse 183 82% Parents divorced 102 46% DV against mother or father 74 33% Emotional abuse 70 32% Emotional neglect 60 27% Parent mental illness 55 25% Physical abuse 43 19% Parents alcohol or drug issues 36 16% Physical neglect 33 15% Parent in prison 15 7% Table 18. Prevalence of disabilities among CYP Disability Service users Percentage Mild Learning difficulties 20 26% Moderate Learning difficulties 19 24% Autism 16 21% Communication speech and language under therapy 12 15% Mild physical disability 11 14% ADHD 7 9% Hearing with aids 3 4% Blind 3 4% Severe Learning difficulties 2 3% Moderate to severe physical disability 2 3% Total with disability 78 Table 19. Prevalence of vulnerabilities among CYP Type of vulnerability Service users Percentage Anxiety/depression 121 41% History of DA 110 37% Other 106 36% school/education problems 82 28% History of self-harm 78 26% Concerns over safety 55 19% CSE 50 17% Suicide risk 47 16% Drugs/alcohol 34 11% Risk of further harm 26 9% Sexualised behaviour 26 9% LA care order 25 8% Missing from home 20 7% Eating disorder 13 4% Total with vulnerability 297 63 Table 20. Onward referrals made by the Lighthouse into borough services, between Oct 2019 and March 2021 Borough Local CAMHS Local Sexual Health GP Other Paediatric Services Counselling Social Services Voluntary Sector A&E Total Barnet 3 0 1 1 2 6 7 0 20 Camden 1 1 0 0 3 2 5 0 12 Haringey 3 1 1 0 2 3 6 1 17 Enfield 4 0 1 0 0 7 8 0 20 Islington 7 0 0 1 1 6 3 0 18 Other 1 0 0 1 0 0 2 0 4 Total 19 2 3 3 8 24 31 1 91 Table 21. Quarterly levels of contact with CYP and families, between Oct 2019 and March 2021 Table 22. Quarterly Strategy discussions and signposting, by borough, between April 2020 and March 2021 Borough Apr-Jun '20 Jul-Sep '20 Oct-Dec '20 Jan - Mar '21 Total Barnet 2 1 1 5 9 Camden 15 21 14 14 64 Haringey 6 0 0 2 8 Enfield 5 4 1 1 11 Islington 6 13 10 16 45 Other 0 0 0 0 Total 34 39 26 38 137 Table 23. Quarterly consultations (delivered by SCLOs) by borough, between April 2020 March 2021 Borough Apr-Jun '20 Jul-Sep '20 Oct-Dec '20 Jan - Mar '21 Total Barnet 10 4 3 1 18 Camden 4 3 8 6 21 Haringey 9 7 4 2 22 Enfield 5 7 3 3 18 Islington 8 11 3 16 38 Other 0 1 0 0 1 Total 36 33 21 28 118 Contact type Oct - Dec 2019 Jan - March 2020 April - June 2020 Jul - Sept 2020 Oct - Dec 2020 Jan - March 2021 Total Number of face to face sessions with CYP N/K N/K 15 161 231 102 509 Number of sessions with parent/carer N/K N/K 13 14 11 16 54 Number of sessions by telephone 388 476 398 573 902 705 3442 Number of sessions by video call N/A N/A 278 149 149 253 829 Number of professional meetings (excluding strategy and consultation) 152 125 123 124 219 193 936 64 Time between reporting to police and being referred to the Lighthouse Number of cases Percentage Lighthouse referral occurred before crime reported to police 3 3% Within 1 week 30 35% Over 7 days to 1 month 28 33% Over 1 month to 6 months 16 19% Over 6 months to 1 year 6 7% Over 1 year 3 3% Total 86 65 Appendix D - Description of service provision in NEL The Northeast London CSA Hub In January 2018 a paper by the Healthy London Partnership (HLP) had been presented to the Joint Commissioning Committee established by the 7 NEL Clinical Commissioning Groups recommending the development of a Child Sexual Assault/Abuse Hub (CSA) for the NEL boroughs; Barking and Dagenham, Hackney, Havering, Newham, Redbridge, Tower Hamlets and Waltham Forest. There were two elements to the hub model. 1. an emotional support service (ESS) to work with CYP who had made disclosures of CSA that did not meet the criteria for the Haven sexual abuse referral centres. 2. CSA clinics for the medical assessment and treatment of CYP staffed by a rota of paediatricians from across NEL. Provision of early emotional support The contract for the emotional support service (ESS) was awarded to Barnardo’s to provide the ‘Tiger Light’ model of trauma informed therapy and support. The service began on the 1st April 2019, delivering across the NEL STP boroughs. Initially it provided 6–8-week sessions of support including trauma-informed early help advocacy and case management, and symptom management, with safe and appropriate onward referral, when necessary, although the number of sessions was subsequently increased to 11 weeks. The support is provided directly to the child/young person and/or their safe parent/carer as appropriate with referrals coming from local Multi-Agency Safeguarding teams. Social workers can refer directly to Barnardo’s if the young person or their family does not want a medical assessment or refer via the CSA medical Hub. The intention was for the CYP to be seen for the first ESS appointment at the same time as the medical assessment, but for logistical reasons (outlined below) this proved impossible. The emotional support service aims to provide early help, reduce the development of long-term mental health conditions such as PTSD, and thereby minimise the need for long-term support from CAMHS. The service was initially staffed with 1.7 fte practitioners and a part time team manager, although this was subsequently increased to 2.6 fte when the capacity of the service was increased by the CCG. The programme is open to any child up to their 18th birthday. Referrals have to have been notified through Childrens’ Social Care and are for abuse that has taken place outside the 3-week forensic window, the service manager described the abuse as typically being ‘much more historical than that….[happening] ‘probably six months to six years ago, it can be quite a long time’. The Barnardo's referral form asked if they could make a referral to the paediatrician, but it was not always completed, so ‘we tend to follow that up ourselves with the family’. Equally the paediatricians, if they get the referral through their route, whether it be via hospital or via a social worker, could refer CYP to Barnardo's for the early emotional support. The expectation is that a full-time worker would see 10 young people a week, their caseload might be slightly higher (possibly around 14) based on levels of anticipated non-attendance, cases starting and closing. The Tiger Light programme provides eight to eleven weeks of support – tending to be shorter for young children, but for some children, ‘can feel that you’re only getting to know them at 8 weeks, so the length of the support varies, something that the CCG recognised. While the focus of the model is on the child or young person, the parent might be offered ‘one or two support sessions as well to help them understand what's happening for their child’. The service manager described the referrals to the service as coming in ‘peaks and troughs’. The target for Tiger Light initially was to see 72 young people a year, but within the first 4-5 months 66 of operation, they had already received more referrals than that, so NEL CCG provided additional funding to increase the numbers to 120 a year. There was a recognition that the potential level of demand in NEL far exceeded the number of places available. 2018/19 MPS data suggested that there were 705 cases of historic abuse reporting in the NEL STP area. Assuming a similar need for assessment to those cases as in NCL (60%), this would have required an annual capacity of around 400 new appointments – far beyond the current availability of services to deliver. As a result, the service had received more referrals than it had capacity for and had had a waiting list from 4/5 months into the start of the project. They had also had to put people on a ‘holding list’ during covid – not being seen because of COVID. Barnardo’s run a weekly referral and allocation meeting, attended by Tiger managers. The criteria for acceptance are age, living or under the authority of one of the NEL boroughs, and a strong suspicion or disclosure of abuse. Most referrals were accepted, decisions about priorities were taken on the basis of what the immediate risks were to the child, and whether the child was engaged with other agencies. This could also affect the way of working with the child; ‘so we’ve got some examples where we work alongside CAMHS for example, we’ve just got one where we're alternating appointments for example now, so that the young person is still getting the support and we can work collaboratively’. On the course’s completion, Barnardo’s made onward referrals to schools for counselling support, they also had volunteer ‘buddies’ who might be matched to the young person to offer them longer term support, but that was mainly around supporting the young person to attain goals; ‘they want to get involved in like cadets or playing football or sports or something like that’. Generally, however ‘a lot of young people at the end of support we don't identify that need any more support that point’. The service also did not generate many returns; while the offer is there ‘to be honest with you, we’ve been operational for 2 years, we’ve only had two re-referrals and not very many young people end up coming back’. In terms of how the Barnardo's service co-ordinated with the CSA clinic, it was described as running ‘in parallel, maybe more in tandem’ (Barnardo’s staff member). The intention had been for there to be greater integration between the two services, with Tiger Light attending the clinic and making contact with CYPs there. In the event, as a result of working days that did not overlap and the impact of COVID, this had not happened. In terms of the impact of Covid on service delivery, Barnardo’s had done very little remote working pre-Covid, limited to phone contact with young people if they were unable to set up their session straight away. Otherwise, work was undertaken face to face, at a location chosen by the young person, normally school or college. When lockdown occurred, it took about 6 weeks for Barnardo’s to migrate their model online, change consent procedures, make sure platform was safe, and address technical issues. The staff also had to start thinking about how they would change the structure of their sessions. As a result, a lot of young people were initially placed on hold, via phone contact. Over time staff had become more proficient at delivering the work virtually, although, as with the Lighthouse, staff had identified the challenges of working from home. However, Barnardo’s had always continued one-to-one support where it was safe and possible to do so, and some staff had been keen to do that from the outset, so they had developed a risk assessment process to allow them to do that (PPE equipment, risk assessment of the site, ensuring the CYP was fully aware of risks), although often it depended on the school granting access to their premises. At the beginning there had been a lot of issues with digital poverty, people not having access to safe, confidential spaces, but ‘being a charity we were quite lucky. People donated like laptops and all sorts of things and we got some Vodafone, dongles, all sort of stuff that could help, but 67 it still doesn't take away the ability to have somewhere safe to sit and have a conversation without fearing that you're going to be overheard’. Most of the team were now out and about delivering sessions; ‘we're offering we can come and see you, so we’re much more out there now’. At the time of writing, like the Lighthouse, Barnardo’s were operating a hybrid model. It was accepted that remote working benefitted some young people, who would like to see elements retained (the absence of travel, and where home was a safe space) but for others, it did not work. Some children had not wanted to engage remotely during pandemic, for young children it was difficult to do the work virtually, and for children with learning needs or where the child had been abused online it did not feel appropriate. NEL CSA clinic The initial (2018) plan for the CSA clinic in NEL was that the seven NEL boroughs would contribute paediatric time to create a sector-wide paediatric rota, staffing a weekly CSA Hub clinic at two sites, 48 weeks of the year, CYP having open-access at either clinic. The two clinics would be offered on a four-week cycle; 3 weeks at the Royal London Hospital, the fourth a clinic for the outer London boroughs, probably located in Waltham Forest. Best practice standards indicated two paediatricians should be present and the intention was to offer improved referral and access to the emotional support service at the time of the medical appointment. The clinics would also enable training and support to paediatricians not currently able to provide this care and provide enough cases a year for them to maintain their competencies. However, the planned model had never been implemented. While the Royal London Hospital introduced a weekly clinic for CSA medicals from April 2019, providing a service for three boroughs (Tower Hamlets, Newham, and Hackney), it proved impossible to develop the second clinical hub, largely due to ‘very limited’ paediatric capacity and expertise in NELFT.62 As a result, the initial proposal was eventually amended to a one-site model at the Royal London Hospital of CSA medicals paid for on a block contract with Barts Health Trust with costs shared across the 7 STP boroughs, beginning in January 2020. As with Tiger-Light, the paediatrician described the number of referrals as ‘lumpy, [I] can go some time with hardly anything and then suddenly 8 referrals in a week. I can see 2 kids in an afternoon clinic. Essentially two children a week, about 80 kids a year’. The clinic usually ran once a week, which the paediatrician described as being ‘enough for current level of referrals I’m seeing’ and there was no waiting list, although, as with the Barnardo’s service there was a recognition that the level of CSA in NEL was much higher than the numbers being seen at the clinic; ‘If you look at police data for NEL for reports of historic sexual abuse to police it’s about 700 a year and I’m seeing a tiny fraction of those patients’. All referrals had to be made by social care via a standard referral form which all local authorities had been sent. Referrals came via email into generic email address, which was vetted by the clinician ‘as I’m still single handed here’. In most cases a referral would be accepted straight away, if there was not enough information on the form, or additional information required, the paediatrician would contact the social worker on the phone. Ideally the child would be seen within 2 weeks, but there was a recognition that this was not always the case, particularly if the paediatrician was on leave ‘if I go on holiday that causes a problem. I’m not meeting the two- week standard sometimes I see them really quickly and sometimes they have to wait a bit’. 62 The rota developed to support a two-site model required NELFT to provide 192 hours of paediatric time a year which equated to 18 hours per month - NELFT had a very small number of paediatricians qualified to undertake CSA medicals who did not have the capacity to support the Hubs as required. Additionally paediatricians from both Barts and NELFT raised logistical issues with a service offered across two sites operated by different organisations using different ICT systems, plus a concern about ‘chain of evidence’ arrangements which are not available in a community setting and which therefore could impact on the outcome of any criminal proceedings. 68 All appointments were undertaken face to face, something which had not changed during Covid, with the child being brought to the clinic by the social worker. Following the appointment, the paediatrician provided verbal feedback to the social worker, followed by a full written report. ‘In majority of cases I don’t need to see kid again, I may refer onto Barnardo’s and follow up any investigations I’ve done during assessment, and if I need to give any treatment, I will contact GP to arrange that to be done. Occasionally I see children for follow up if something is abnormal’. Other people involved in the clinic were the paediatrician’s PA, supported by the play therapy team in the hospital, who the paediatrician described as ‘effectively doing it free of charge, but without that support it would be really difficult for me to deliver the service as well as I do, they offer a lot of distraction and support for children during the examination. I bleep the on-call play therapist, I don’t have an allocated play therapist, I bleep them and they are on a job and they are peeling off another job to help me, and if they are too busy they may not be able to come.’ Initially it was envisaged that the weekly clinic would allow for 2 new patients or 1 new and 2 follow up patients to be seen, and that the clinic would be staffed by a rota of paediatricians who would prospectively cover annual / study leave to ensure the clinic could be delivered 48 weeks a year. However, one of the major difficulties for the clinic has been obtaining paediatric support for the lead paediatrician (something which had also prevented the two-site model being implemented in the first place). While other paediatricians had attended the clinic, this had largely been for training purposes or to get experience, they had not been released from their jobs to spend allocated time at the clinic or been paid to do so. As a result, the clinic found itself in a Catch-22 situation; ‘I can’t pay for the extra resource without having the additional referrals and I’m only having a trickle of patients through. The service is entirely reliant on me to drive the improvements. I’ve got to be adaptable and flexible to allow it to grow’. It was recognised by the CCG that there was a need to develop the service gradually to support the growing competence and recruit additional staff to meet the likely required capacity. Overall, the annual cost of the paediatric clinic at Barts is £60,390 pa (staffing costs of £59,640 as below plus £750 annually for the maintenance of the colposcope. Staffing costs include the costs of monthly clinical supervision and peer review to ensure emotional support for the team). Per Year Rate Total Consultant Paediatrician 624 hours 84 £52,416 Hospital Play Specialist 216 hours 28 £6048 Psychologist 24 hours 49 £1176 Total £59,640 69 Appendix E – Supplementary tables of comparisons between Lighthouse and NEL services Table 24. Boroughs referring to NEL CSA Hub services between January 2019 and May 2021 NEL Tiger Light NEL Paediatrician Total NEL Referring borough CYP Percentage CYP Percentage CYP Percentage Barking & Dagenham 48 34% 6 7% 54 23% Havering 30 21% 1 1% 31 13% Redbridge 25 18% 2 2% 27 12% Newham 13 9% 11 12% 24 10% Waltham Forest 12 9% 9 10% 21 9% Tower Hamlets 9 6% 37 41% 46 20% Hackney 4 3% 24 27% 28 12% Total 141 90 231 Table 25. Comparison of service user age groups63 Lighthouse NEL Barnardo’s NEL Paediatrician Total NEL Age group CYP Percentage CYP CYP CYP Percentage 12 and under 367 41% 85 74 159 60% Age 13-17 505 57% 92 15 107 40% 18 and over 17 2% 0 1 1 0% Total 889 177 90 267 Table 26. Comparison of service user gender64 Lighthouse NEL Barnardo’s NEL Paediatrician Total NEL Gender CYP Percentage CYP CYP CYP Percentage Female 730 82% 133 75 208 78% Male 155 17% 44 14 58 22% Transgender 2 0.2% 0 0 0 0% Total 887 177 89 266 63 This includes the age data on all Lighthouse referrals, not just those who consented. 64 This includes the age data on all Lighthouse referrals, not just those who consented. 70 Table 27. Barnardo’s Tiger Light wellbeing outcome measures65 Outcome Good improvement Some improvement Stayed the same Some decline Total Access to support services 1 10 11 2 24 Enhanced parent/carer/adult - child relationships 1 7 16 1 25 Increased confidence 1 10 11 3 25 Ability to express feelings 1 14 7 3 25 Improved mental health & well-being 1 16 5 3 25 Positive social/cultural/religious identity 0 9 14 0 23 Improved self esteem 1 13 9 2 25 Knowledge of sexual health strategies 0 1 1 0 2 Able to identify abusive/exploitative behaviour 0 2 0 0 2 Able to describe safety strategies 2 10 12 1 25 Reduction in level of risk/harm 0 0 2 0 2 Carers/staff aware of safety strategies 2 3 20 0 25 Reduction in impact of trauma 2 14 6 3 25 Family has access to support services 0 0 2 0 2 Link with reliable and supportive role models 0 8 15 0 23 Sustained progress on exit from service 1 7 14 0 22 Parents/carers active in working with service 0 0 2 0 2 Engaged in personal action planning 2 12 11 0 25 Grand Total 15 (5%) 136 (42%) 158 (48%) 18 (6%) 327 Table 28. Comparison of service user ethnic groups66 Lighthouse NEL Barnardo’s NEL Paediatrician Total NEL Ethnic groups CYP Percentage CYP CYP CYP Percentage White 140 47% 76 12 88 43% Asian 23 8% 34 7 41 20% Black 70 23% 34 3 37 18% Other Mixed 20 7% 12 3 15 7% Mixed White and Black 31 10% 8 4 12 6% Mixed White and Asian 7 2% 3 2 5 2% Other Ethnic group 10 3% 2 4 6 3% Total 301 169 35 204 65 ‘Good improvement’ equals a change of -3 or -4 in scores, ‘some improvement’ equals a change of -1 or -2 in scores, ‘stayed the same’ means the score remained unchanged, and ‘some decline’ equals an increase of 1 – 2 in risk. 66 Lighthouse ethnicity data is only for those who consented. 71 Table 29. Lighthouse emotional wellbeing goal setting Coded goals Achieved Partially Achieved Not Achieved Total No recorded outcome yet Support with emotions 13 10 1 46 22 Confidence & Identity 12 6 28 10 Parent support 11 2 1 25 11 Other 11 3 4 25 7 Healthy relationships 3 1 20 16 School support 5 4 1 12 2 Assessment of need 5 1 11 5 Health and wellbeing 4 4 1 14 5 Therapeutic support 4 1 1 10 4 Develop relationships 2 2 5 1 Further support 1 2 5 2 Safety advice 1 3 2 Total 69 (59%) 35 (30%) 13 (11%) 204 87 72 Appendix F – Calculations underpinning economic evaluation RedQuadrant’s (RQs) analysis deploys a Difference-in-Difference approach, a statistical technique often used in social science, that studies the differential effect of a treatment on a 'treatment group' versus a 'control group'. That meant: • Collating data on key outcomes at baseline and after treatment from the Lighthouse Annual Report; • Collating data on expected levels of improvements for these outcomes from the existing literature on the after-effects of CSAE for victim-survivors; and • Reviewing the improvement for the Lighthouse versus that identified by the literature The RQ model identifies the benefits that accrue compared against two hypotheses (a) treatment under non-specialist services; and (b) under a partly integrated approach. In drawing-up the benefits, a wide range of literature was reviewed, much of which is drawn from the international literature. In considering the benefits for inclusion, the criteria used were to consider whether they are: (a) likely to be significant in terms of impact; (b) measurable; and (c) attributable to the interventions delivered at the Lighthouse. The focus of the economic analysis is on the comparison of Lighthouse against standard forms of provision in London, in particular those considered in Harewood and Baine (editors) (2018) “London Child Sexual Abuse Learning Report. 67 This report examined three models funded by the London CSA Transformation Programme, all aiming to support CYP after experiencing CSA. These models were (1) the Children and Young People’s Haven Service (CYP Havens) based within London’s Sexual Assault Referral Centre, (2) the Lighthouse and (3) CSA hubs. As well as the insights from Harewood and Baine (2018), useful data is available in the Lighthouse Annual Report on: • Statistics on the proportion of goals met among C&YP (67% were achieved, 37% partially achieved and 4% not achieved), and • Statistics on means and standard deviations on progress made towards individual goals for the parent course (which increased from an average of 5.4 to 7.8). Attribution Just as the measurement of outcomes is not straightforward, so too is attribution of causality of improvements in outcomes to the interventions offered at the Lighthouse. There is a complex relationship between the child’s characteristics (such as their resilience and their pre-existing mental health); the family; and the quality and quantity of the therapeutic intervention. A key challenge is in establishing the counterfactual – determining what would have happened without the intervention. RQ have used two sources of information to assist in considering this point – insights into referral rates after intervention by the three different forms of support identified above; and information from the Lighthouse Annual Report and from Barnardos on the goals achieved by clients from the Lighthouse and from the NEL CSA hub respectively. Taking referral rates first, the Lighthouse Annual Report 2019-20 cites that 63% of C&YP 67 https://www.england.nhs.uk/london/wp-content/uploads/sites/8/2019/05/London-CSA-Services-Learning-Report-2018- v1.2-002.pdf 73 reported mental health conditions at initial assessment. However, this is highly likely to be an under-statement of the true rate of such conditions among the cohort, as the level of short-term support needs for the cohort is of the order of 80% to 90%. This suggests an impact of improving mental health issues of the following order: • Lighthouse – 67% (calculated as 85% baseline rate of referrals - 18% referral rate after Lighthouse intervention = 67%) • CSA South West London Hub – 54% (calculated as 85% baseline rate of referrals - 31% referral rate after CSA Hubs intervention = 54%). This represents a level that is proportionally 80% (since 54% ÷ 67% = 67% and • C&YP Haven - 45% (calculated as 85% baseline rate - 40% referral rate after Havens intervention = 45%). This represents a level that is proportionally 67% of the impact of the Lighthouse (since 45% ÷ 67% = 67%). A potentially useful cross-check is whether this pattern of referrals is upheld when examining more recent outcomes data. Ideally, this would compare Lighthouse and Havens data, as this is the scenario examined in relation to cost data. Unfortunately, however, outcomes data are only available for the Lighthouse and North East London CSA Hub. These data indicate 89% Lighthouse goals achieved or partially achieved, compared to NEL CSA Hub data showing improvements in 47% of outcomes. This suggests a relative level of improvement of 53% for North East London CSA Hub compared to the Lighthouse (since 47% ÷ 89% = 53%), considerably lower than the CSA South West London differential in relation to referrals. In choosing which set of data to underpin their estimate of impact RQ have chosen to adopt the results from Harewood and Baine (2018), as this presents consistent data across the three types of provision. This implies an estimate of impact of 67% for the Lighthouse, and 45% for the Havens counterfactual, with the Haven’s counterfactual in turn having an impact that is proportionally 67% of the Lighthouse impact but note that this estimate has to be treated with caution. Review of benefits used in the RQ analysis The long-term outcomes of child sexual abuse are known to be associated with a wide range of psychosocial and health outcomes. A recent review, Hailes et al. (2019), found the strongest links for two psychiatric disorders (post-traumatic stress disorder and schizophrenia) and one psychosocial outcome (substance misuse). Research by McNeish and Scott (2018), quoted by the UK-based Centre of expertise on child sexual abuse, finds CSA to be strongly associated with the following outcomes across the life course: - physical health problems, including immediate impacts and long-term illness and disability - poor mental health and wellbeing - externalising behaviours such as substance misuse, ‘risky’ sexual behaviours, and offending - difficulties in interpersonal relationships - socio-economic impacts, including lower levels of education and income - vulnerability to re-victimisation, both as a child and as an adult. Decisions on which benefits to examine in the economic analysis is inevitably subjective. It is also important to note that there are some benefits on which it is difficult to place a monetary value, but which nevertheless command a high value. These include the benefits to wider society of supporting victims and survivors, both immediately and in later helping them to overcome the 74 effects of their experience; and the obligation to act in the best interests of the child. RQ’s analysis incorporates three types of social value: a) wellbeing to the client (such as the wellbeing value of less trauma); b) useful savings from public sector spend (for instance, less need for mental health services); c) more public sector spend on essential activity (such as more offenders going to prison). These encompass the following domains: - Health & wellbeing – sexual health, mental health, physical health, substance abuse, and - client and family wellbeing - Children’s services and child protection – child taken into care, school support - Employment – loss of earnings for client, loss of value for public sector - Criminal justice system – improved prosecution rates - System effects – closer co-ordination In estimating social value for given categories, RQ have used two (roughly equivalent) methodologies, depending upon the data available. The first approach calculates the proportion affected by a given issue and multiplies it by the unit cost of that issue. So, for example, in relation to physical health, if it is known that 14% of child maltreatment cases require NHS treatment, and the average cost of such treatment is £850, then the unit cost is estimated at £120 (14% * £850) in respect of physical health for clients as the base case. The second approach takes the average cost of a given issue to the cohort, and reduces it in line with the impact of the Lighthouse. For instance, if the average spend on mental health services (when required) is £11,450, and the Lighthouse has an effect which reduces trauma by 25%, then the public sector saving is 25% * £11,450, which equals £2,862.50 Health and well-being Looking initially at sexual health RQ examined the proportion of cohort with a sexually transmitted infection (STI), proportions of different infections and the cost to treat. Khadr et al. (2018) found 12% of a sample of adolescents attending Sexual Assault Referral Centres had at least one sexually transmitted infection diagnosed 4 to 5 months after the assault. However, the Lighthouse Annual Report suggests that the prevalence of STIs requiring treatment is lower than this, as the Report indicates (p36) that 6% of prescriptions for the cohort involved antibiotics; further, it is highly unlikely that all of these prescriptions would be for STIs. An additional factor is the high proportion of primary age children attending the Lighthouse compared with predominantly teenagers attending the CSA Hubs after sexual assaults. RQ therefore scale down their estimate of the impact of the Lighthouse on the prevalence of STIs to be of the order of 1% of the cohort (that is, 4 cases of STI out of 400 or so clients receiving both initial assessments and onward support between October 2018 and September 2020). Turning to the issue of how much sexually transmitted infections cost to treat, they draw on Sadler et al. (2016). This provides an assessment of treatment costs for major sexual health issues – Chlamydia £120, Gonorrhoea £210, HIV £280,000 and Syphilis £210. Updating this in line with a 20% increase in the cost of GP appointments (and after equalising the unit cost so that it reflects the same amount of time per GP appointment) (PSSRU Unit Cost of Health and Social Care 2020, and PSSRU Unit Cost of Health and Social Care 2015), these rise to £144, £252, £336,500 and £252 respectively. 75 The next issue is creating a weighted average unit cost that reflects the relative extent to which these STIs occur in the age group(s) for the Lighthouse. Using the average rate on prevalence of infections from: • Sadler et al. (2016) (p29), which gives data for the prevalence of those STIs in 2015 for the 0 to 12, 13 to 14, and 15 to 19 age range [in relation to HIV, RQ have taken the mid- way point for the population as a whole, and the “high risk” population]; • Public Health England (2020) (p10 to 11) ‘STIs and screening for chlamydia in England, 2019: annual official statistics’, which directly gives data for prevalence for chlamydia, gonorrhoea and syphilis in 2019 for the 15 to 19 age range, and which they have used to scale the estimates for other age ranges in Sadler et al. (2016) to 2019 levels, and • Public Health England (2020) ‘Key population HIV data tables number 2’, (tables 1 and 4) on diagnoses of cases for those aged below 15 and those aged 15 to 24, which are used to scale the estimates for HIV by age range in Sadler et al. (2016) to 2019 levels. Calculations are shown in the table below. Age range % Lighthouse cohort Prevalence of Chlamydia by age Prevalence of Gonorrhoea by age Prevalence of HIV by age (adjusted) Prevalence Syphilis by age 0 to 4 7.6% 0.000% 0.000% 0.000% 0.000% 5 to 12 38.4% 0.000% 0.000% 0.000% 0.000% 13 to 15 30.3% 0.111% 0.016% 0.007% 0.000% 16 to 17 21.7% 2.250% 0.287% 0.032% 0.004% 18 2.0% 2.250% 0.287% 0.032% 0.004% When applied to an expected number of 4.5 among the Lighthouse cohort of 450 (1% * 450 = 4.5), these proportions imply 3.9 cases of Chlamydia, 0.5 cases of Gonorrhoea, and 0.1 cases of HIV. From these calculations, an average cost of £7,630 per client with STI then follows. Much of this cost relates to HIV treatment, and hence to NHS costs. The table below shows a difference in NHS costs of £395 per client in respect of sexually transmitted infections (STIs). It should be noted, however, that this estimate is extremely dependent upon the prevalence of HIV infections for the cohort under consideration and is highly indicative. Proportion with issue Cost per issue Cost (£) Proportion with issue Cost (£) Difference (Base case) (After) Children with STIs 1.0% 7,630 75 0% - 75 Turning to physical health and substance misuse, the table below shows no difference in respect of NHS costs for maltreatment costs, £75 improvement for obesity-related costs, and £30 improvement in respect of substance misuse treatment. 76 Proportion with issue Cost per issue Cost (£) Proportion with issue Cost (£) Difference (Base case) (After) Physical health (maltreatment) 14% 850 120 14% 120 0 Physical health (obesity) 63% 880 555 54.6% 480 75 Eating disorders 2.4% 8,850 210 1.5% 135 75 Substance misuse (alcohol) 2.3% 4,000 95 2.0% 80 15 Substance misuse (drugs) 5.8% 4,000 230 5.0% 200 30 On physical health, RQ considered maltreatment and obesity. In relation to maltreatment, Conti et al (2017) estimates the proportion of maltreated children requiring treatment in hospital and cites £850 as the national unit cost of non-elective hospital admissions for paediatric injuries. For indicative purposes, the model assumes that the Lighthouse (a) increases the rate to which treatments are identified, so increasing costs; (b) decreases the rate that future treatment is required; (c) these two effects negate each other. In relation to obesity, RQ start with estimates of the unit costs to the NHS, and then consider the likely change in the proportion of the cohort with the issue after intervention. A starting point for unit costs is the estimate that the NHS spent £6.1 billion on overweight and obesity-related ill-health in 2014 to 2015; and the calculation that in 2015, 63% of adults in England were classed as either overweight or obese. Given that the adult population in England in 2015 was 54.8m, this implies a cost per person of the order of £177 per year, since £6,100m ÷ (54.8m * 63%) = £177. Over a five-year period, this amounts to NHS costs of the order of £880. To assess the risk of adult obesity they use Hailes et al. (2019), which provides a meta-analysis of risks for a range of adverse outcomes for those who have suffered CSA, including the risk of obesity. Based on 26 studies involving more than 160,000 subjects in the UK, it identifies an Odds Ratio of 1.4 : 1.0 for those who experienced CSA. Whether or not the Lighthouse may affect such risks depends on whether it can influence the driver(s) of that difference. Hailes et al. (2019) explains that this may be possible: “The association between childhood sexual abuse and physical health problems might also be partially explained by mediating psychiatric factors. For example, the effect of childhood sexual abuse on obesity might be due to depression or certain eating disorders”. Assuming that this hypothesis is at least partly correct, RQ calculated the effect of the Lighthouse on reducing obesity as follows: • Average proportion of obesity in adults = 63% • Effect of increasing risk by factor of 40% = 0.40 * 63% = 25.2% • Proportion of this increased risk among those who have had CSA that is attributable to depression or eating disorders is assumed to be half • Impact of the Lighthouse on mental health = 67% • Impact of the Lighthouse on obesity by improving mental health = 25.2% * 50% * 67% = 8.44% Around 0.75 million people in the UK are estimated to have an eating disorder, according to PWC (2015) “The Costs of Eating Disorders. Social, Health and Economic Impacts” (p21). which represents a level of approximately 1.1% among the UK population. The report calculates an average health cost per sufferer of £8,850. RQ apply the same methodology as for obesity, using the Odds Ratio of 2.2 for eating disorders shown in Hailes et al. (2019). 77 The calculations are as follows: • Average proportion of eating disorders = 1.1% • Effect of increasing risk by factor of 120% = 1.2 * 1.1% = 1.3% increase, to 2.4% prevalence • Impact of the Lighthouse on mental health = 67% • Impact of the Lighthouse on eating disorders = 1.3% * 67% = 0.9% Turning to substance misuse, RQ apply the same methodology as for obesity, using the Odds Ratio shown in Hailes et al. (2019) since that study states that high quality studies show a strong link between substance misuse and CSA, and calculates an Odds Ratio of 1.7 : 1.0 for this issue. The calculations are as follows: • Average level of drugs dependency in adults (taking a Class A drug) = 3.4%, according to NHS Digital, Statistics on Drug Use (2019) • Average level of alcohol dependency in adults = 1.37%, according to Public Health England • Effect of increasing risk factor by 70% = 2.4% in relation to drug use, and 1.0% in relation to alcohol dependency • Impact of the Lighthouse on drug use after (1) scaling for impact on mental health and (2) assuming that proportion of this increased risk among those who have had CSA that is attributable to mental health problems is assumed to be half = 2.4% * 67% * 50% = 0.8% • Impact of the Lighthouse on alcohol dependency after (1) scaling for impact on mental health and (2) assuming that proportion of this increased risk among those who have had CSA that is attributable to mental health problems is assumed to be half = 1.0% * 67% * 50% = 0.3% The estimated annual cost to the NHS of alcohol dependency, per year per dependent drinker is some £2,000, updating NICE guidance for inflation (RQ assumed treatment applies for two years, and also assumed the cost of treatment for substance abuse is the same as for alcohol). Mental health is considered from the perspectives of (a) the NHS and local authority; (b) for the individual themselves, (c) for their family. In doing so, RQ make a distinction between costs and trauma incurred during childhood, and as an adult. Conti et al. (2017) (p31) provides an estimate of the cost of treatment per case during childhood for hyperkinetic disorders, conduct disorders and emotional disorders to be £11,450. The report further estimates a cost of around £950 in respect of anxiety and £5,145 in relation to depression as an adult (p32). These values are calculated in present value terms (costs up to 30 years from the age of 6 are discounted using an annual rate of 3.5%; and by 3% p.a. for costs incurred between 31 and 74 years). In relation to costs during childhood, RQ have scaled the estimate of cost per treatment by the impact that the Lighthouse has on the need for future referrals. As noted previously, the level of mental health disorders that the cohort has on entry into the Lighthouse is of the order of 80% to 90%; following Harewood and Baines (2018) impact is estimated to be of the order of 67%. This is in line with the outcomes reported in the Lighthouse Annual Report 2019- 20, which states (p64) that “During the year 2019/20, 276 goals were agreed with children and young people and 67% were achieved, 37% partially achieved and 4% not achieved.” 78 In the long-term, the effects of Lighthouse on adult mental health are not known, and consequently RQ have used the same methodology on the Odds Ratio of adverse effects that was applied earlier. • Average proportion of anxiety in adults (in 2014 Adult Psychiatric Morbidity Survey) = 7% in females and 5% in males. For a cohort that is 80% females and 20% males, the benchmark level is 6.6%. • Average proportion of depression in adults (in 2014 Adult Psychiatric Morbidity Survey) = 3.8% in females and 3% in males. For a cohort that is 80% females and 20% males, the benchmark level is 3.64%. • Odds Ratio of 2.7 for anxiety and depression in Hailes et al. (2018) implies that effect of CSA is to increase expected prevalence without intervention by 11.2% and 6.2% respectively. • Impact of the Lighthouse on mental health = 67%, and so expected impact of the Lighthouse on anxiety = 67%* 11.2% = 7.5%, and expected impact on depression = 67% * 6.2% = 4.1%. Based on these calculations, the table below shows a reduction of £7,670 on NHS/LA spend in respect of mental health for the client up to age of 18, and further reduction of £455 spend as an adult. Average cost per client (£ NHS/LA spend) Reduction in proportions Cost per client (£) after intervention Difference Mental health (up to 18) 11,450 67% 3,780 7,670 Adult anxiety 950 7.5% 880 70 Adult depression 5145 4.1% 4,930 215 In relation to client wellbeing, RQ have used the estimate in Home Office (2018), of a wellbeing loss from rape of £31,450. Their assumption is that there is an improvement in wellbeing (or more precisely, an avoidance of some of the loss of wellbeing) in line with the estimate of a 67% improvement in mental health. This equates to a value of £21,100 per client. Family members – parents and siblings – can also experience “profound harm” as a secondary trauma due to sexual assault on a child. The disclosure of CSA can be a major life crisis for the non-abusing parent or carer. This possibility is exacerbated if the parent has experienced abuse in childhood themselves. Research findings are limited, however. Newberger et al. (1993) in a study of 44 mothers and 2 maternal care givers found 55% of them initially in the clinical range of traumatisation, with symptoms diminishing over a year but still present for many. RQ have assumed that the extent of this trauma is on a par with violence with injury, at a level of £11,200 (according to Home Office 2018). They further assume that this applies to, on average, one parent and one sibling, so making a potential loss of wellbeing of 2 * £11,200 * 69% rate of secondary trauma as a proportion of primary trauma = £15,400. The impact of group-based interventions for non-abusing parents have been found to be significant in achieving positive outcomes for children and in reducing carers’ stress. Evidence shows that parents who participate in such groups report increased wellbeing and confidence, reduced stress, and greater ability to care for their child and deal with professionals. In particular, the Lighthouse Annual Report 2019-20 cites outcomes in the parents’ psychoeducation course that attendees made progress in 69% of their goals and 31% stayed the same (see page 10). 79 Using the previous estimate of impact of 67%, RQ calculate a wellbeing improvement of £10,320 in relation to family wellbeing (on the basis of the calculation 67% * £15,400 = £10,320). Average adverse effect (£) Reduction in effect (%) Adverse effect after intervention Difference in adverse effect Client wellbeing 31,500 28% 22,640 -8,860 Family member wellbeing 15,400 28% 11,070 -4,330 Children’s services and child protection In examining Children’s Services RQ have looked at: • Formal child protection action due to Lighthouse investigations • School support (child behaviour and/or learning difficulties). The table below shows no identifiable saving for local authorities in respect of children taken into care per Lighthouse client, but a £1,200 saving for them in respect of additional school support. Proportion with issue Cost per issue Cost (£) Proportion with issue Cost (£) Difference (Base case) (After) Formal child protection action 8% 217,000 17,585 8% 17,585 0 School support 22% 32,200 8,190 16% 5,880 -2,300 The Annual Report (p47) cites that 16 out of 198 clients were subject to a Local Authority Care Order, a rate of 8%. In addition, 12 cases were referred onward to Local Authority children’s services (p62 – although this may be an under-estimate, but does not affect the results for reasons explained below). Since the average age of the Lighthouse cohort is 11.8 years, this would imply additional protection action for 6.2 years; either in the form of a 6 month Child Protection Plan (at an indicative cost of £922 according to Holmes (2021) (p41) ‘Children’s social care cost pressures and variations in unit costs’) and/or the cost of placement, which PSSRU unit costs data imply would be around £26,000 per year for foster care, and £86,000 per year for residential care, with an overall cost of £217,000 per case over 6.2 years for the proportions of support outlined in DfE statistics on children in care. However, RQ feel that volumes are unlikely to be reduced by the Lighthouse (indeed, some CAC research suggests an increase in the numbers of children taken into care), and in line with the view that spending that is required but not previously undertaken should not be included in the analysis, and have set the difference in costs between the Lighthouse and the alternative to zero. Turning to school support, Conti et al (2017) (p35) provides an estimate of the cost of additional school support as some £32,200 over a ten-year period, with the proportion of children requiring such support rising by 22% due to maltreatment. RQ assume that the Lighthouse proportionally reduces this cost by 67% - that is, from 22% to 7.3% - as per previous assessments. Employment The table below suggests that the Lighthouse is able to reduce losses of earnings by £6,600 per client, and loss of taxes / increases in benefits by £1,300 per client. 80 Proportion with issue Cost per issue Cost (£) Proportion with issue Cost (£) Difference (Base case) (After) Employment losses (client wellbeing) 100% 50,875 50,875 87% 44,270 -6,605 Employment losses (DWP/HMRC) 100% 10,170 10,170 87% 8,850 -1,320 Direct ONS data on the issue is not available, but survey results for adult victims of crime suggests that the effect of sexual assault in hampering employment can be substantial. Drawing on data from the 2020 Crime Survey, analysis set out in ‘Nature of sexual assault by rape or penetration, England and Wales: year ending March 2020’ (see table above) cites 5.6% of respondents “lost job or gave up work”, and 33.2% of respondents took a month or more off work. Conti et al. (2017) (p36) calculates loss of earnings into adulthood for those affected by child maltreatment based on a 5% effect on wages, drawing on Daro (1988). RQ have adopted a 13% figure, taking the average of this study, Barrett & Kamiya (2012) (p18), which indicates a 12% effect, and Fergusson et al. (2013) (p670), which suggests a 22% figure. Note also that a 13% figure is close to the expected impact of a 12% reduction in the level of adult anxiety among the cohort. Over a 30-year period, set against average earnings for a group that is 80% female and 20% male, this amounts to some £50,875 in terms of lifetime loss of earnings. The DWP / HMRC loss is assumed to be equal to 20% of this loss of earnings. The savings to these Departments due to the Lighthouse vary in line with changes in earnings, which also are assumed to vary in line with changes in mental health. Criminal justice Criminal justice issues relate to observed patterns of increased criminal activity by child abuse victims when they reach adulthood. Other possibilities that were considered were a reduction in cracked trials, and effects on reductions in reoffending by the perpetrator. Both of these were excluded on the grounds of lack of good data. RQ have also excluded costs in relation to charitable funds of ISVAs, as broadly speaking the same amount of money is spent on per ISVA whether they are funded by the public sector or by charity. The table below shows an estimate of a reduction in CJS spend due to future criminal behaviour by the victim of some £1,775; and an increase in CJS spend on crimes prosecuted and convicted due to an increased willingness to testify by victims. Average cost per client (£) Change in proportions Cost per client after intervention (£) Difference Criminal behaviour by victim when an adult 7,100 -67% 2,340 -4,760 Crimes prosecuted and convicted 32,110 +5% 33,715 +1,605 Conti et al. (2017) (p32) cites average costs of criminal behaviour by victims as £7,100, and (p34) cites average costs per case of £32,110 (£11,750 on courts, £20,360 on convictions). Then the assumption is that the Lighthouse reduces additional criminal behaviour of victim in line with their mental health improvement of 67%; and that a small effect (5%) relates to extra cases brought to light; a small proportion is used given the mixed research evidence, and the strong 81 possibility that there would be a reduction in reoffending by the perpetrator which would lower the costs for the CJS. System effects Turning to the last domain of social value, the table below shows an estimated productivity improvement due to better co-ordination between agencies of some £1,720 per client. System cost per case (£) Improved productivity Difference (£) Co-ordination effect 28,625 6% -1,720 The source for the scale of improved productivity is Nat Cen (2012), a study of an integrated approach to drugs offenders, which estimates improved co-ordination improving efficiency by 6%. Clearly, this is only one example and not directly equivalent, so can only represent an indicative figure. The system cost per case derives from: • court costs of £11,750 in respect of maltreated children); • police & CPS costs (at least 50% of court costs, according to NAO data on cost of young offenders); • Local Authority investigation costs £10,400 based on PSSRU unit costs data; and (d) NHS costs (based on consultant paediatrician and psychotherapy appointments). 82 Appendix G - Summary of Early Emotional Support (EES) Services commissioned in London (FY 2020/21). STP Boroughs ESS service Provider NEL Barking & Dagenham, City & Hackney, Havering, Newham, Redbridge, Tower Hamlets, Waltham Forest The CSA Hub service (TIGER Light) offers early emotional support to children, young people and their families being seen for a CSA medical examination at the newly established CSA Hub in Royal London. The paediatricians providing CSA paediatric assessments in the hub work with the emotional support practitioners, providing a holistic health review focused on the needs of the child/young person and their family. Barnardo’s NWL Brent, Ealing, Hammersmith & Fulham, Harrow, Hillingdon, Hounslow, Kensington & Chelsea, *Westminster TIGER Light is a trauma informed early emotional wellbeing service provided by Barnardo’s. TIGER means Trauma Informed Growth and Empowered Recovery. The approach builds on the expertise Barnardo’s has developed over the past 25 years to improve the lives of sexually exploited children Barnardo’s SEL (LSL) Lambeth Lewisham Southwark Safer London offer dedicated support to children, young people and their families who have experienced CSA. They champion their voices to ensure they receive the best possible support, and offer specialist consultations and advocacy. The paediatricians providing CSA paediatric assessments in the LSL boroughs will work with the Safer London practitioners, providing a holistic health review focused on the needs of the child/young person and their family. Safer London SEL (BBG) Bexley, Bromley, Greenwich Family matters confirmed providing a service in Bexley but no further information was received. No information received for Bromley and Greenwich Family Matters (Bexley) SWL Croydon, Kingston, Merton, Richmond, Sutton, Wandsworth The NSPCC provides up to 6 sessions to the child or young person and their family in order to support them through the initial difficult stages following a disclosure. During the sessions the worker will get to know the child, and together with the child and family make an assessment of their needs, which will inform the focus of the support sessions. In addition, an assessment of longer-term therapeutic need will be completed. NSPCC 83 NCL Barnet, Camden, Enfield, Haringey, Islington The Lighthouse offers early and long-term support is offered through CAMHS, the NSPCC’s LTFI and Protect and Respect services. This additional emotional support and therapy helps CYP to recover. The Lighthouse model is different to the rest of London. It offers emotional support as part and parcel of its CSA services and does not offer EES as standalone service. NSPCC, CAMHS All Across London The CYP Havens is a specialist sexual assault referral centre (SARCs) that provides immediate services to reduce the risk of long term problems. In addition to providing FMEs and post assault follow-up care they provide counselling therapy and clinical psychology treatment for up to one year post-assault (Up to 20 sessions of EES post abuse for under 18s; 16-18 yrs. access to SARCs for up to one-year therapeutic support post abuse). CYP Havens also make referrals for EES support to the services described above. The CYP Havens 1 The Lighthouse: 2-year interim evaluation report November 2020 Rachael Parker and Tim Read MOPAC Evidence & Insight 2 Contents Executive Summary ...................................................................................................... 3 1. The Lighthouse evaluation ......................................................................................... 7 Methodology ............................................................................................................... 7 Future plans for analysis of impact ................................................................................ 8 2. Performance review ................................................................................................ 10 Performance data sources ........................................................................................... 10 Referrals to the Lighthouse ......................................................................................... 10 Initial assessments at the Lighthouse ........................................................................... 12 Background of clients who consented to evaluation ...................................................... 13 Offence details ........................................................................................................... 15 Advocacy goal setting ................................................................................................. 16 Summary of performance review................................................................................. 17 3. Learning from implementation ................................................................................ 18 Consistency to the developed model ........................................................................... 18 Ways of working......................................................................................................... 23 Partnership working ................................................................................................... 26 Looking ahead ............................................................................................................ 29 4. Discussion .............................................................................................................. 31 Next steps .................................................................................................................. 32 References ................................................................................................................. 33 Appendices ................................................................................................................ 34 3 Executive Summary The Lighthouse, London’s Child House, opened at the end of October 2018 initially as part of a two-year pilot, although now funded until September 2021. Bringing together a range of organisations under one roof, the Lighthouse intends to be a child friendly, multidisciplinary service for victims1 of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden, it serves the five surrounding North Central London boroughs of Barnet, Camden, Enfield, Haringey and Islington. The Evidence and Insight (E&I) Unit are MOPAC’s in-house social research and analytical team and were commissioned to evaluate the Lighthouse. The E&I evaluation focuses on four distinct areas for analysis; a performance review; a process evaluation; impact evaluation and an economic evaluation. This report concentrates on the first two areas, looking at the first 21 months of operation of the Lighthouse.2 Findings to date Process: Summary of performance review The data used for the performance review came from two sources. The first was provided by the Lighthouse data officer and included aggregate data around referral month, borough and referral source, in addition to age categories and gender of the referrals. The second, and more comprehensive data source was individual-level data produced from Excelicare, the case management system for the Lighthouse. This data was only for clients who specifically consented to have their data included in the MOPAC evaluation. Between the end of October 2018, when the Lighthouse launched, and the end of July 2020 there were a total of 639 referrals to the service, which works out as an average of around 29 referrals per month, which is lower than the estimated demand of 700 referrals per year that was projected at the start of the pilot.3 The highest number of referrals in one month (with 42 referrals) was December 2018, and after that the number varied between 23 to 35 referrals each month. However, there was a large drop in referrals from March 2020 to April 2020 (34 to 14) which coincided with the Covid-19 lockdown. Referrals returned to the ‘normal’ range from June 2020. Comparing the volume of referrals between boroughs, there have been some notable changes since the last interim report. Barnet had previously been one of the lowest referring boroughs but is now the highest referring borough having referred 23% (n=146) of all the referrals within the reporting period. The next borough being Enfield (which has always been a high referring borough) which referred 20% (n=130); this is not surprising given they have the largest child populations. 1 Referred to as victims throughout the remainder of the report. 2 Details of the rationale for the establishment of the Lighthouse, and E&I’s overall evaluation approach are contained in E&I’s first evaluation report, published in April 2019. A subsequent interim evaluation report, describing the first nine months of Lighthouse service provision was published in 2020. Both are available on MOPAC’s website. 3 Conroy et al. (2018) used police data to estimate the potential demand 4 As found in the previous evaluation report, children’s social care remains the largest referring organisation, making over half of the referrals to the service (52%, n=333), Police made 10% (n=61), self-referrals made up 6% (n=38) and medical sources (i.e., GPs, Hospital and sexual health clinics) made up only 6% (n=41) of referrals. Most children and young people (CYP) referred were female (81%, n=520), and the majority were in the age bracket 13-17 years (n=340, 53%). These demographics coincide with findings from other studies which report that girls and older children are more likely to experience sexual abuse.4 There were also 49 referrals for survivors aged 18-25 years (with learning difficulties), where there were none reported in the previous report. Overall, there were no notable differences across boroughs or organisation in terms of who they were referring (such as gender and age variation). Between the end of October 2018 and the end of July 2020 the service carried out 392 Initial Assessments (IAs). Discounting October 2018 (where only 1 IA was conducted)5, the number of IAs averaged 18 per month (varying from 13 to 25). However, there was a large drop in initial assessments in April and May 2020 (to 9 and 7 IAs respectively), due to Covid-19 lockdown and the closure of the Lighthouse building. During this time, practitioners at the Lighthouse conducted IAs virtually. Subsequently the number of IAs has recovered to pre- Covid-19 levels. The conversion rate of referral to IA overall sits at 61% (very similar to the proportion (59%) observed at the point the previous evaluation report was written). Of the 392 referrals that reached an initial assessment, 279 individuals consented to take part in the evaluation – an overall consent rate of 71% (up from 56% at the time of the previous evaluation report). However, it was noticeable that, despite an increase in referrals and initial assessments from June 2020 to July 2020, the consent rate dropped markedly to 27% during this period. An explanation for this drop in consent is that the Lighthouse’s change to virtual working led to shorter appointments and more focus on therapeutic needs in that shorter time. Therefore, consent for MOPAC evaluation was not taken at IA and was sometimes delayed but the service are reviewing whether all outstanding CYP have been asked for consent; this will be continue to be monitored. Examination of the characteristics of the consenting individuals showed the majority were female (83%, n=231), the most common age range was between 13 and 17 (n=126, 46%), the average age was 12 and the modal age 14 (n=30). Ethnicity was recorded for 250 service users and there was an almost equal split between BAME and non-BAME clients (n=130 and n=120 respectively); a comparable proportion to the previous report. As evidenced in the previous evaluation report, referrals are a highly vulnerable group of CYP with complex needs. A large proportion of service users (87%, n=234) were assessed as having at least one type of vulnerability, and 157 (67%) of these service users have at least 2 types of vulnerability. Among the most frequently identified types were anxiety and/or depression (n=96), followed by history of domestic violence (DV) (n=86) and education problems (n=61). A fifth of service users (n=55) had a history of self-harm. Twenty-eight percent of service users were recorded as having a disability (n=65/231), with 22 having more than one disability. Mild (n=16) or moderate (n=13) learning difficulties were the most common forms of disability. 4 NSPCC, 2019 5 The service opened to referrals on 23rd October, therefore it was unlikely to start seeing children until after at least a week in order to give them enough notice to attend. 5 The Adverse Childhood Experience Questionnaire (ACE-Q) was completed at initial assessment with 162 service users. Of these, the majority (n=147/162) had a score of at least 1, and two in five (n=64) had a score of 4 or more.6 The most common ACE for the Lighthouse service users was sexual abuse (n=139, 86%7), followed by parents divorced/separated (n=77, 48%). Comparing the Lighthouse service users’ ACE-Q results to other populations, the results are consistent with previous findings: the Lighthouse ACE scores are above non-clinical national populations8, but closer to other at-risk youth populations. 9 Process: Summary of learning from early implementation Findings related to practice at the Lighthouse were drawn from several qualitative sources including: interviews and focus groups with Lighthouse staff, telephone interviews with police officers who have worked with the Lighthouse, police liaison officers (PLOs), social care liaison officers (SCLOs) and a focus group with borough representatives working with the SCLOs. This interim report focusses upon these roles due to their uniqueness within the Lighthouse. Focusing on the PLO and SCLO roles, for the former, their role in liaising with external police officers and the CPS around the progression of investigations, and the facilitation of police and psychologist led ABE (Achieving Best Evidence) interviews at the Lighthouse was stressed as particularly important by respondents. For the latter, the part they played in ensuring appropriate referrals to the Lighthouse was emphasised, as well as providing guidance and advice around safe-guarding and social care pathways to Lighthouse and local authority staff. Both roles were reported as essential to the Lighthouse model. As in the last evaluation report, staff remain concerned about elements of the Lighthouse’s infrastructure/estate although staff were aware of attempts to rectify these. Even with specific work conducted, soundproofing and telephony remain issues for staff, and there are on-going changes being made to the case management system to try and improve it. The quality of some referrals and caseloads across the various services within the Lighthouse also remain an issue for some staff. Similarly, there remain tensions within the Lighthouse arising from the different disciplines working there, and the perception on the part of some staff that their part of the service feels undervalued. Attempts by senior management to address these concerns have had limited success to date but are on-going. 6 The ACE-Q is an internationally validated self-report tool encompassing 10-items across 10 areas which cover household dysfunction (parental separation/divorce, parental domestic violence, parental substance misuse and mental illness, and parent incarceration), child abuse (sexual and physical), and child neglect (emotional and physical). The more events that a person experienced before the age of 18, the higher their ACE ‘score’ will be, and literature demonstrates that the higher the score (a maximum of 10) the greater the risk of health issues (i.e., mental or physical), substance misuse, victimisation and offending in adulthood.6 This emphasises the importance of providing holistic, integrated support to these young people to mitigate the risk of health and lifestyle problems in adulthood. 7 Lighthouse staff suggest this percentage must arise as a result of incomplete data as clearly all the children that attend the Lighthouse have experienced sexual abuse. 8 As a baseline, a nationally representative survey of adults in the UK found that 46% of respondents reported at least 1 ACE, and 8% reported at least 4. This was undertaken by Bellis et al. (2014) with 3885 18-69-year olds in the UK. There have also been many other studies, in various populations and nationalities, which have also shown that most adults (between 52%- 75%) have experienced at least one ACE (Zarse et al, 2019). 9 In a study that looked at vulnerable young people with mental health problems in Scotland (who present serious harm to others), there was a much higher prevalence of ACEs; 93% (out of 130) had experienced at least 1 ACE, and 59% had experienced at least 4 (Vaswani, 2018). 6 There have been developments in terms of service delivery since the last report. Psychologist led ABEs have now begun at the Lighthouse and twice-weekly intake meetings in place of daily allocation meetings. However, the major operational changes have been necessitated by the demands of the Covid-19 pandemic, and the Lighthouse staff have shown a remarkable degree of adaptability and flexibility in continuing to offer a service to young people under lockdown initially, and subsequently as the service has moved to a hybrid model of virtual and face-to-face working. Interestingly, some of these changes, particularly the use of technology to facilitate virtual meetings between members of staff, external agencies, and CYP and parents in certain circumstances, appear likely to be retained once the situation returns to normal. However, staff also emphasised that certain aspects of the Lighthouse’s work require face to face working, and the ability of the organisation to deliver the intended service has been disrupted since March 2020. In addition, certain expected practices are still not yet operational; the use of the Live Link facility, while now installed and functional, has still not started. The main challenges anticipated by Lighthouse staff for the future continue to be uncertainty about ongoing funding and sustainability, and concerns about the ability of the service to respond to the anticipated increase in demand once the schools had re-opened, although at the time of writing, this had not materialised. This is the third in a series of four MOPAC E&I Lighthouse evaluation reports to be released, enabling learning both internally as a catalyst for improvement, and externally to advance the evidence base. A final evaluation report will be produced in Summer 2021 which will aim to explore the effectiveness of the Lighthouse initiative against criminal justice, health and wellbeing outcomes. 7 1. The Lighthouse evaluation The Lighthouse, London’s Child House, opened in October 2018 initially as part of a two-year pilot, although funding has subsequently been extended until September 2021. Bringing together a range of services (medical, social care, police, advocacy and therapeutic support) under one roof, the Lighthouse intends to be a child friendly, multidisciplinary service for victims and survivors of Child Sexual Abuse and Exploitation (CSA/E). Based in Camden it replaces the existing services 10 and serves the five surrounding North Central London boroughs of Barnet; Camden; Enfield; Haringey; and Islington. Further details of the Lighthouse’s objectives and structures and the services it offers can be found in E&I’s previous evaluation reports.11 Methodology The Evidence and Insight (E&I) Unit is MOPAC’s in-house social research and analytical team which has been commissioned to undertake an evaluation of the Lighthouse. The evaluation will cover the processes of the Lighthouse (from design through implementation), monitor routine performance, as well as seek to explore impact and cost benefit. The evaluation uses a mixed methodology approach – balancing qualitative context from staff, stakeholder or client feedback, particularly in the shorter-term, with the ‘harder’ performance figures indicating how the service is running on a day-today basis. It focuses on four distinct areas; performance monitoring; process; impact and economic analysis. The ability to successfully complete each element will depend on the quality and quantity of data and will be reviewed throughout the life of the research, as it is subject to change. Over the course of the three-year pilot, four E&I evaluation reports are planned, of which this is the third. Taking a broad action research perspective - findings from the evaluation are routinely fed back to the programme teams, the academic advisory group set up to advise the evaluation, to update partners at the official Partnership Oversight Board and other relevant meetings to ensure learning is continually shared within an active feedback loop. Like the previous evaluation report, this report focuses upon the performance monitoring and process aspects of the evaluation, exploring the first 21 months of the Lighthouse’s operation and draws from the following data: • Performance management data. This includes aggregate performance data provided by the Lighthouse as part of its reporting to the service commissioners. This provides an overview of service delivery between go-live (October 2018) and July 2020 (numbers of referrals, sources of referrals, number of assessments etc). In addition, data taken from Excelicare (Lighthouse’s case management system (CMS)) provides details of the service received by individuals at the Lighthouse, and about the individuals themselves. However, it should be noted that these data only relate to individuals who have consented to participate in the evaluation, a subset of those who 10 NB CYP Havens continue to provide the acute/Forensic Medical Exam (FME) service. 11 https://www.london.gov.uk/what-we-do/mayors-office-policing-and-crime-mopac/data-and-statistics/academic- research 8 have received the service overall. The number of individuals to consent is a potential evaluation risk and is explored in greater detail in the results. • Focus groups and interviews held between January and August 2020 to explore perceptions of the early implementation of the Lighthouse: o A mixture of face to face (before Covid-19) and virtual interviews and focus groups with 14 members of Lighthouse staff from several teams including senior management, Letting the Future In (LTFI), PLOs, SCLOs and health staff12; o Virtual focus group with 4 borough SCLOs; and o Telephone interviews with 13 police officers and written email feedback from a further 2 officers. A final Lighthouse evaluation report is planned for the Summer of 2021. In addition to looking at process and performance issues this report will also examine the impact of the Lighthouse in terms of criminal justice, health and welfare outcomes, comparing the Lighthouse to a comparison site, and examining the perceptions of service users. Future plans for analysis of impact To date, E&I’s evaluation of the Lighthouse pilot has focussed on process and performance issues. However, the final report due in 2021 also aims to undertake an impact evaluation, examining if Child House has delivered its desired outcomes and if and how much it has impacted upon those who are involved. To robustly evaluate impact and seek to explore which key aspects or ‘ingredients’ of the service provided by the Child House has had an effect, a comparison or counterfactual is essential (i.e., this is a group of similar individuals, in a similar location, but who did not receive the Child House services), against which the outcomes and experiences of those who did receive the Child House services can be compared. The development of the counterfactual is one of the most critical aspects within any evaluation seeking to explore impact. There are many different techniques to draw from, from the ‘gold standard’ randomised control trial, to quasi-experimental approaches (such as generating comparison groups) or lesser quality approaches that only look at before and after an initiative. The stronger the design the more confident one can be in the findings and any differences observed are associated with the treatment. The Lighthouse roll-out presents several challenges when deciding upon a counterfactual. The strongest design (i.e., randomly allocating to the service) is not feasible on ethical grounds. There is no ideal ‘like for like’ comparison, given the unique status of the Lighthouse service. The approach to be adopted will be quasi-experimental and will focus on identifying a comparable area within London, and then within this area identify a comparable collection of clients (i.e., CSA caseload) to then compare to a Lighthouse cohort on key outputs (i.e., 12 Virtual working during Covid-19 lockdown made it difficult to make contact with other members of staff. Seeking feedback from the remaining teams will be prioritised for the next report (e.g. Advocacy, CAMHS, and the admin team). 9 referrals); outcomes (i.e., health, wellbeing, criminal justice outcomes) subsequent to interventions; as well as staff and service users’ perceptions13. This is essentially comparing the Lighthouse service to a business as normal service. The final evaluation will report on this impact analysis and is due mid-2021. 13 See the first evaluation report for the full list of outcomes that are being measured. Measuring criminal justice outcomes takes time to allow for cases to move through the criminal justice system. Equally, measuring health and wellbeing outcomes requires use of psychometric instruments which may not be sensitive enough to pick up changes after just a few months. These outcomes therefore have been chosen to be analysed at the final stage to assess the impact of the service. 10 2. Performance review Performance data sources This section presents a picture of the Lighthouse service in terms of its internal workings and processes, throughput, activities, and the demographics of its clients using data such as: - Numbers of referrals - Referrals by month - Number of initial assessments - Referring borough - Referral source - Service allocations - Client demographics - Client vulnerabilities and disabilities This relatively basic information is key in understanding how the service is operating and what has delivered. The evaluation team had two main data sources for this report. The first provided by the Lighthouse data officer and which included aggregate data around referral month, referring borough and referral source, as well as the age categories and gender of the CYP. The second, and more comprehensive data source was individual-level data produced from Excelicare. These data were for all clients who consented to have their data used in the MOPAC evaluation. At the point of writing this report, this was 279 individuals (71%). These data were much more detailed and provided over 300 variables, although at this stage of the evaluation the level of completion of the various fields varies widely. Referrals to the Lighthouse Between the end of October 2018, when the Lighthouse launched, and the end of July 2020 there were a total of 639 referrals to the service, which works out as an average of around 29 referrals per month (see Figure 1). The highest number of referrals in one month was (with 42 referrals) December 2018, and after that the referrals varied between 23 to 35 referrals each month. There was a large drop in referrals from March 2020 to April 2020 (34 to 14) which coincided with the Covid-19 lockdown. However, the referrals were then back up to ‘normal’ range from June 2020. 11 Figure 1. Total number of referrals, IAs, number of services users who consented to the evaluation. The volume of referrals continues to remain below the original estimation of 700 per year14. When comparing the volume of referrals between boroughs, there have been some changes since the last report. Barnet had previously been one of the lowest referring boroughs, but is now the highest referring borough and referred 23% (n=146) of all the referrals within the reporting period, and Enfield (which has always been a high referring borough) referred 20% (n=130) (more details of referrals by borough are provided in Appendix A). Children’s social care are still the largest referring organisation, making over half of the referrals to the service (52%, n=333)15, police made 10% (n=61), self-referrals made up 6% (n=38) and medical sources (GPs, Hospital and sexual health clinics) made up only 6% (n=41) of referrals. The majority of CYP referred were female (81%, n=520), and the majority were in 14 Conroy et al (2018) It is unclear why the number of referrals is lower than initially predicted. This may be because the original estimation was too high, that the original police data was not the most suitable to base predictions upon or could be linked to wider factors such as the communication across referrals agencies or awareness of the service. 15 It is understood that in most CSA cases the police and social care work together. The social worker and police officer will decide between them who is best placed to complete the Lighthouse referral and in practical terms the social worker is best placed to complete the form. 12 the age bracket 13-17 years (n=340, 53%). There were also 49 referrals for survivors aged 18- 25 years (with learning difficulties), where there were none reported in the previous report. These demographics coincide with findings from other studies which report that girls and older children are more likely to experience sexual abuse.16 Overall, there were no notable differences across boroughs or organisation in terms of who they were referring (such as gender and age variation). A further exploration of police reported CSA, referrals and attrition in the criminal justice system will be addressed in the final report. Initial assessments at the Lighthouse Between the end of October 2018 and the end of July 2020 the service carried out 392 Initial Assessments (IAs). After October (where 1 IA was conducted)17, the level of IAs is at an average of 18 per month, varying from 13 to 25. However, there was a large drop in initial assessments in April and May 2020 (9 and 7 IAs respectively), due to Covid-19 lockdown. During this time, practitioners at the Lighthouse conducted IAs virtually using an NHS system called Attend Anywhere. Since then the number of IAs have increased to normal range. The overall number of clients engaged with is far below the expected number that the service was commissioned to support (544 in one year) 18. The conversion rate of referral to IA sits at 61% and is consistent with the conversion rate recorded in the previous report (60%). For the remaining 247 referrals that did not reach an IA yet, there could be several reasons including: they are pending further information; the CYP and/or parents/carers did not want the service; they did not meet the criteria; they received a consultation to the professional network; they attended for a video recorded interview; they may be pending CYP availability, or the CYP are not feeling ready yet. The final evaluation report will aim to further uncover the reasons for referrals not reaching an IA and the other ways in which they might be supported by the Lighthouse. At each IA there may be several practitioners in the room,19 and there was an average of three professionals present at each IA (this data only relates to those service users who consented to providing their data for the evaluation) - this is no change since the last report. Most frequently this profession was a Paediatrician (91%, n=254), followed by an Advocate (44%, n=122). A category of ‘other’ was present for 134 IAs, and this includes individuals such as Social Workers, foster carers, or family members. As a result of the changes to the service delivery model necessitated by the Covid-19 pandemic (described in the ‘Consistency to the developed model’ section) the number of virtual IA’s increased after late March 2020. The Lighthouse’s own figures show that between October and December 2019 there were 50 face to face IAs and no virtual IAs. Between January & March 2020, there were 55 IAs overall (50 of which were face to face while 5 were 5 virtual). However, between April and June 2020 the pattern changed quite dramatically – the number of referrals decreased (to 46), as did the number of IAs overall (to 33), but of these, while 14 were face to face, 19 were virtual. 16 NSPCC, 2019 17 The service opened to referrals on 23rd October, therefore it was unlikely to start seeing children until after at least a week in order to give them enough notice to attend. 18 As outlined earlier, this difference could be related to a variety of factors. 19 These may include a paediatrician, play therapist, advocate, sexual health nurse and clinical psychologist. 13 Achieving Best Evidence (ABE) interviews The Lighthouse is contracted to offer 3 psychologist-led ABEs a month. In the nine months before June 2020 there were 13 psychologist-led interviews undertaken at the Lighthouse (4 between April-June 202020, 5 between January and March 2020, and 4 between October and December 2019), therefore about half of the number of contracted interviews completed. Over the same period there were 19 police led ABEs at the Lighthouse (11, 6 and 2 for the same quarters – so the number of police led ABEs at the Lighthouse has increased during Lockdown)21. The rest of the performance analysis will only relate to service users who have consented to providing their data for the evaluation. The overall consent rate has improved since the previous report, and the current consent rate is 71% which equates to 279 clients out of the 392 who reached initial assessment. However, despite an increase in referrals and initial assessments conducted from June 2020 to July 2020, the consent rate dropped markedly reaching a low point of 27% (n=6)22. The evaluation will continue to monitor the consent rate. Background of clients who consented to evaluation This section will provide a summary of the demographics, backgrounds, and needs for the service users who consented for their data to be used in the evaluation. Demographics For the 279 service users who consented to sharing their data for the evaluation, their demographics were very similar to the overall cohort: the majority were female (83%, n=231). The ages ranged from 0-25; the most common age range was between 13 and 17 (n=126, 46%). The average age is 12 and the most common is 14 (n=30). Ethnicity data was recorded for 250 service users and there was an almost equal split between BAME and non-BAME clients (n=130 and n=120 respectively); this is a comparable proportion from the previous report. The majority of service users are in full-time education (89%, n=233/263). Vulnerability Lighthouse staff conduct several different assessments with the Lighthouse service users, often at the initial assessment stage, and as such collect a wide range of data relating to service users’ vulnerabilities and needs. For example, staff record Adverse Childhood Experiences Questionnaire (ACE-Q) scores during the IA based on past experiences. They will also assess and record current vulnerabilities of the children (such as depression, anxiety, eating disorders, drugs and alcohol), taking into account any history of self-harm or previous 20There were more psychology led ABEs planned between April and June 2020, but a number were cancelled due to family anxiety about travelling in for an interview during COVID lockdown. 21 Whilst the rest of the Lighthouse building was closed during lockdown, the ABE suite was still made available for police to use during that time. 22 An overview of the work that went into improving the consent rate is provided in the previous evaluation report. 14 attempted suicide. Practitioners will also assess any future risk for the children and young people such as risk of suicide, self-harm, or risk to others. Looking across these datasets in summation we see (as expected) a highly vulnerable client group. The Adverse Childhood Experience Questionnaire (ACE-Q)23 was completed for 162 clients24, whose ages ranged from 3 to 21. For these service users, the ACE scores ranged from 0-9, with an average of 3. The majority (n=147/162) of service users had a score of at least 1, and two in five (n=64) had a score of 4 or more. The most common ACE for the Lighthouse service users was sexual abuse (n=139, 86%25), followed by parents divorced/separated (n=77, 48%; see Appendix A for the prevalence of all ACEs). Putting these figures into context, comparing the Lighthouse service users’ ACE-Q results to other populations, the results are consistent with wider evidence: the Lighthouse ACE scores are above non-clinical national populations26, but closer to other at-risk youth populations.27 Information is also recorded on the service users’ background including vulnerabilities, disabilities and medical history.28 A large proportion of service users have such vulnerabilities (87%, n=234 out of 269). These 234 service users presented a total of 618 vulnerabilities between them (at an average of 2.7 each). 157 (67%) of these service users have at least 2 types of vulnerability. Among the most frequent categories were anxiety and/or depression (n=96), followed by history of domestic violence (DV, n=86) and education problems (n=61). A fifth of service users (n=55) had a history of self-harm. 28% of clients were recorded as having a disability (n=65/231)29, and 22 had more than one disability. For the 65 who had a recorded disability, there were a total of 93 disabilities between them. Mild (n=16) or moderate (n=13) learning difficulties were the most common forms of disability. A full breakdown of service user vulnerabilities is presented in Appendix A. 23 The ACE-Q is an internationally validated self-report tool encompassing 10-items across 10 areas which cover household dysfunction (parental separation/divorce, parental domestic violence, parental substance misuse and mental illness, and parent incarceration), child abuse (sexual and physical), and child neglect (emotional and physical). The ACE questionnaire has been used internationally and the original ACE study used a 10-question tool (Felitti et al. 1998), however the ACE-Q has sometimes been adapted by other organisations or researchers and has either been shortened or lengthened in terms of the number of items (Bethell et al. 2017)The more events that a person experienced before the age of 18, the higher their ACE ‘score’ will be, and literature demonstrates that the higher the score (a maximum of 10) the greater the risk of health issues (i.e., mental or physical), substance misuse, victimisation and offending in adulthood. This emphasises the importance of providing holistic, integrated support to these young people to mitigate the risk of health and lifestyle problems in adulthood. 24 ACE-Qs were not completed for all clients due to an issue at the start of the pilot where there were issues as the service adjusted to using a new patient record system and ways of working. 25 NB previous caveat in footnote 7 refers 26 As a baseline, a nationally representative survey of adults in the UK found that 46% of respondents reported at least 1 ACE, and 8% reported at least 4. This was undertaken by Bellis et al. (2014) with 3885 18-69-year olds in the UK. There have also been many other studies, in various populations and nationalities, which have also shown that most adults (between 52%-75%) have experienced at least one ACE (Zarse et al, 2019). 27 In a study that looked at vulnerable young people with mental health problems in Scotland (who present serious harm to others), there was a much higher prevalence of ACEs; 93% (out of 130) had experienced at least 1 ACE, and 59% had experienced at least 4 (Vaswani, 2018). 28 In most cases, the information at IA is recorded by the Doctor or Health and Wellbeing practitioner. 29 Disability information was not recorded for the whole sample. 15 Through risk assessments conducted by staff, there is a variety of valuable information captured to further the knowledge on vulnerabilities, including whether a CYP is currently known to children’s social services (n=166/235), or whether they have previously been known to children’s social care or early intervention support services (n=125/217). 30 The assessments also show that 33 of 220 service users are subject to care order or child arrangement orders. A key part of the risk assessment is also for the worker to assess whether there is further risk to the CYP or others (see table 1 below). This type of information on the risk assessment was only 184 service users. 31 It is apparent that many of the service users are still at risk, particularly of further abuse where there are at least ‘some’ risks or concerns for 61% (n=112) of service users. Table 1. Risk assessment details Risk Ongoing and potential immediate risk Some risks/concerns None Total Further abuse 15 97 72 184 Self-Harm 18 63 103 184 Suicide 8 50 126 184 Risk to others 2 17 160 179 Offence details The Police Liaison Officers and children and young people’s advocates at the Lighthouse record data on Excelicare around criminal justice details, including characteristics of the offence, and key dates in the criminal justice processes (i.e., police actions and decisions, CPS actions). As outlined earlier the impact of the Lighthouse on criminal justice processes and attrition will be analysed for the final evaluation report. At this stage, it is too early to examine criminal justice processes and timeliness through the system, however some analysis has been undertaken to understand the nature and types of offences. The most frequent offence or incident was intra-familial sexual abuse where almost half of all service users experienced this (48%, n=134), followed by peer-on-peer sexual abuse (n=75).32 See table 2 for a full breakdown of types of incident. Twenty-one service users experience at least two types of violence. Most offences (64%, n=179) occurred within the five Lighthouse boroughs, 45 were outside of London, and 28 were within another London borough. For two clients there were two locations recorded.33 30 Please note that the base size differs due to gaps in data collection. 31 As mentioned previously, this appears to be due to data quality issues in the early stages as the service adjusted to a new case management system and ways of working. 32 No distinction was made between rape cases or other child sexual abuse cases. 33 For the remaining 29 cases the location was not recorded. 16 Table 2. Type of sexual abuse incident Incident description Number of service users (out of 279) Proportion Intrafamilial sexual abuse 134 48% Peer on peer 75 27% Extra family abuse 38 14% Assault by unknown 17 6% Violence unknown 15 5% Child Sexual Exploitation 9 3% Sexually harmful behaviour 7 3% Violence other 6 2% In 241 cases the CYP made a disclosure of the offence, most commonly to someone in their family (n=119). Other individuals disclosed to were someone in their school (n=23), a health service (n=14), police (n=12), social services (n=12), peers (n=6), or a community service (n=3). For the remainder it was unknown who the disclosure was made to. The analysis also shows that a large number of service users were recorded as being repeat victims (n=87 out of 197), 42 service users were victims of multiple perpetrators, and violence was a feature of the incident for 31 service users. Further highlighting the level of risk and vulnerability that this group of service users faced. Advocacy goal setting The children and young people’s Advocates support the service users throughout their time with The Lighthouse. The Advocates ensure the voice of the child is heard by professionals within and outside of the Lighthouse. They can be involved at any stage of the pathway including: a show around prior to first appointment, work with the young person outside of the Lighthouse before attending, support on the day of their evidential interview following ABE guidance, at the initial assessment, advocating with school, social care and the police; and finally a key role in support throughout the police investigation, court preparation and support during the trial. When a child engages with the Advocate at the Lighthouse, together they set goals for what the child would like to achieve in the support they receive from the service and these goals are recorded on the case management system34. The goals are recorded as free text but were coded by the evaluation to get an understanding of the types of goals being set and the progress against them. There were 206 different goals detailed in this dataset for a total of 87 clients, with an average of 2.3 goals each and a range of 1-10. Among these 87 clients: 51 are BAME and 36 are non- BAME; minimum age is 5 but the majority were in the higher age bracket 13-20 (n=70); the majority were female (n=77). The most common type of goal was relating to criminal justice support (n=59) which could mean that the service user wanted to be kept updated on the proceedings or wanted to be supported by the Lighthouse throughout the investigation which 34 Goals are also set with the CYP by other services such as CAMHS and LTFI. 17 may include help in understanding their options or to demystify the criminal justice processes. The second most common type of goal was around education and support in school (n=33), which may include getting help getting back into school, or for the Lighthouse to liaise with the school around a support plan, or help communicating the service users’ needs to the school. A full breakdown of the types of goals in place are outlined in Appendix A. Against these goals, the Lighthouse staff indicate whether the goals have been achieved or not. At the stage of analysis: 85 goals were achieved (indicating that action against the goal had been taken), 43 were partially achieved, and 14 were not achieved (it is not recorded why the goal was not achieved). For the remainder there was no indication of the progression of the goal although it may be because they are still in progress or data error. Summary of performance review Over the first 21 months, the Lighthouse has received 639 referrals and delivered 392 IAs (a 61% conversion rate). The rate of referrals and IAs has remained fairly consistent over the lifetime of the pilot so far, however there was a noticeable drop in the number of referrals from March to April 2020 when the Lighthouse was forced to close due to the Covid-19 lockdown. There have been some key changes in referrals between boroughs since the previous report; notably Barnet is now the highest referring borough, having previously been one of the lowest. Enfield also remains one of the highest referring boroughs. Interestingly the two most outer boroughs have referred the most. Consistent with previous findings, Children’s Social Care remains the highest referring agency. The nature of the offences that brought the CYP to the Lighthouse show that most of the abuse was intra-familial. This reinforces the lack of safe and stable household that some service users are in. Notably, through analysis of the advocacy-based goals, the majority of the service users and their families appear to particularly value support through the criminal justice system and their respective investigations, something which has also emerged from the qualitative interviews, particularly with the PLOs, that have been undertaken. The consent rate for the evaluation has improved markedly since the previous report, to 71%, which means that 279 clients are included in the more in-depth analysis. As was found in the previous report, the theme of vulnerability was clearly seen in the data from numerous assessments, and information regarding their victimisation. This group of children and young have clearly experienced a disproportionate amount of adversity in their lives so far. Such a wide palette of observed vulnerability requires a service that can call on and work with multiple partners and strands of support to best support. 18 3. Learning from implementation This section explores learning from a range of sources, notably interviews and focus groups carried out with Lighthouse staff, telephone interviews with police officers, and a focus group with borough SCLOs; with the aim of presenting learning related to the set-up and continued implementation of the Lighthouse. As the evaluation progresses, the aim will be to track these issues over the course of the Lighthouse. The previous evaluation report identified six themes; four of these form the basis for this report: • Consistency to the developed model; • Ways of working; • Partnership working; and • Challenges ahead.35 In the main, these themes continue to provide the framework for the findings that follow. There is also one additional theme that emerges in this report - the impact of the Covid-19 virus – but this has been discussed in terms of its impact on the subject areas above. Consistency to the developed model Since the previous evaluation report the Lighthouse has begun to undertake psychologist as well as police-led interviews following Achieving Best Evidence (ABE) guidelines (which are also referred to by police as Video Recorded interviews (VRI)) on the premises. The ABE interview process at the Lighthouse forms a significant part of the PLOs’ roles, who acknowledged that the ABE process could be ‘labour intensive’. The PLOs describe that the process takes up a full eight-hour day, from the practical requirements, to final planning with the psychologist around the evidential points to prove and how challenges to the victim’s account should be presented without “shaming” them. In the early months of the Lighthouse service, one of the PLOs had spent much of their time training the psychologists. The PLO took a creative approach, including practising with them in role play exercises and arranging trips to Crown Court to observe trials of sexual abuse cases. It was suggested that this process had been more difficult and taken longer than anticipated, as in practice it wasn’t a simple transfer of skills. A contributory factor was the psychologists clinical work and “full caseloads” reducing their availability to conduct or practice ABE interviews. The PLO’s acknowledged that the psychologist’s strength is their understanding of the impact of trauma and their skill in adapting and managing this within the evidential interview; combined with the PLO’S 35 The other two themes were Perceptions of the Lighthouse and Referrals and demand. Information relating to the first is now in the Ways of Working and Partnership sections and the second in the Performance Analysis and Partnership Working sections. 19 expertise means best practice is followed and the victim is given the best opportunity to provide a detailed evidential account and valuable experience for any investigating officer. Another operational change was an alteration to the frequency of the allocation meeting – which changed in early March 2020 from a daily meeting to a twice-weekly intake meeting. While the same range of people attended (an individual from LTFI, CAMHS, advocacy, SCLO. PLO, paediatrician) rather than have a various people chairing, the meeting had moved to one consistent chair - the Consultant Psychiatrist). A variety of explanations were given for the change; more time to ensure that information required for the children being discussed had been collected, greater continuity overall, the time/resource implications of the daily meeting, and the ability to provide better administrative support for more infrequent meetings. Although it was recognised that the change might lead to a longer delay in allocation for the child it was stressed ‘We’re not an emergency service, so to be able to review and make a decision within 72 hours is perfectly fine’. The COVID-19 pandemic had a major impact on the service delivery model at the Lighthouse. Following the introduction of lockdown (March 23rd), the Lighthouse site was closed, apart from allowing occasional urgent ABE interviews36 and medicals (which were also available at UCLH), although it was stressed that ‘most people chose to wait’ until after lockdown. Intake meetings continued but moved to video conference with triage and consultation for new referrals. Similarly, rather than being undertaken in person, initial assessments were offered as ‘Virtual IAs’ with the CYP and family meeting the full Lighthouse IA team on an NHS video consultation facility called ‘Attend Anywhere’; ‘you can have multiple professionals in the same meeting so for a child and family they were seeing a multidisciplinary team’, including social workers who attended the first part of the virtual IA to update the medical staff (developments demonstrated by the increase in the number of virtual rather than face to face IAs shown in the performance data). Medical history details were taken but examinations were delayed until the Lighthouse reopened and it was safe to do so. Unsurprisingly, as shown in the performance analysis, the number of referrals to the Lighthouse and IAs declined during Lockdown. Interestingly, staff felt there was better attendance at appointments by CYPs, who can be hard to reach, although the average time from referral to IAs increased to 39-49 days in lockdown with many children and families opting to wait for an IA once the Lighthouse building has opened. During this time there was an increase in telephone contact and ad hoc virtual appointments to keep in contact with families. Meetings between professionals took place on Microsoft Teams – and attendance was felt to be good, with GPs and CAMHS attending. Generally, under lockdown there was an increased frequency in the number of virtual team meetings, although the number has subsequently decreased. CAMHS had set up a daily meeting with the SCLOs to provide safe-guarding advice 36 The criteria for psychologist led ABEs was changed to: those with an urgent safeguarding risk, young child where there was a risk of memory loss or a child the psychologist had already worked with ‘I think it’s brilliant, and having worked in a police team of child abuse investigation and knowing the service they get, often not great police interview rooms when they come to police stations, so the children here get sort of top service, and it really strikes me how much they are thought about’ (PLO) 20 because of concerns about the impact of Covid-19 restrictions on the CYP they were managing in terms of the mental health, isolation and increased anxiety. Also discussed was the morning IA to discuss and reflect on it as a team. Heads of service met initially 3 times a week, although the frequency of the meetings had subsequently declined to twice a week and (currently) once a week. At the point of writing (September 2020) the Lighthouse operates a ‘hybrid’ model with some activities being undertaken at the Lighthouse as before, and others being offered virtually, or in combination. Similarly, some members of staff are back working at the Lighthouse, some are working at home, and some are working alternative weeks at home and in the Lighthouse, providing services face to face and virtually as circumstances demand. The Lighthouse re- opened for face-to-face IAs from 8th June onwards. ABEs and IAs take place at the Lighthouse regularly, together with limited therapeutic sessions (the latter tends to be a mix of virtual and face to face work). Some CYP remain on remote appointments only. In addition, a ‘virtual’ parent psychoeducation course began in July with seven parents attending. There are elements of the Lighthouse model that it was anticipated would form part of the service when it was initially implemented which are still yet to be introduced. As outlined in the previous two evaluation reports, Section 28 (s.28) of the Youth Justice and Criminal Evidence Act 1999 sets out a range of special measures which should be available to help vulnerable and intimidated witnesses be cross examined at a criminal trial, including pre- recording cross-examination. It was hoped that the pre-recorded cross-examination would take place at the Child House rather than a child having to give evidence at court. While this remains the long-term ambition, the delayed implementation of s.28 nationally has prevented this happening in the Lighthouse. The technology to allow this to happen has been installed, albeit not until October 2020, but approval from the judiciary/HMCTS has still to be obtained. Similarly, the use of the Lighthouse as a Live Link location, allowing children to give evidence remotely without attending court was approved in Spring 2020, although the installation was delayed until October 2020. The Police Liaison Officer and Social Care Liaison Officer roles within the Lighthouse For this report, it was decided to gain more in-depth insight into the roles of the Police Liaison Officers (PLOs) and Social Care Liaison Officers (SCLOs) at the Lighthouse. The reason for looking in-depth at these roles is due to their uniqueness; as there had previously been no such roles for child sexual abuse in the UK.37 The two PLOs at the Lighthouse were interviewed face to face in January 2020, and the two SCLOs were interviewed virtually in July 2020. The PLO The PLO role is unique to the Lighthouse; there is currently no such role within other support services for child sexual abuse in the UK. This section presents feedback on the role from both the PLOs themselves, and officers in the boroughs. 37 There has since been recruitment for an SCLO role at the Havens. 21 Both PLOs at the Lighthouse are experienced front line investigators in child abuse cases; one officer has been with the Lighthouse service from the beginning, and one officer joined in July 2019. They explained their key motivation for taking on the PLO role at the Lighthouse was the model’s victim-centred approach, where they hoped they would have opportunity to “do a really good job for the child” at the Lighthouse. The PLOs described their role as ‘busy’ and ‘varied’; with a key aspect being to demystify criminal justice processes both for practitioners within the Lighthouse and the service user. An example of this is providing training on how to respond to a child that makes a new disclosure and how to record it. It was reported to be particularly important for the Advocates to have assistance and “help them understand why the police or courts or the CPS [Crown Prosecution Service] are doing what they’re doing” or “what’s going on in the case and why it’s so slow”, acknowledging that it can be ‘distressing’ for families when they do not understand the process, or are not kept up to date on the investigation’. Whilst the PLOs do not routinely attend appointments with children at the Lighthouse, they explained that they did have opportunities to interact with the children and young people (aside from the ABE interviews). These are opportunities for the children or their parents/guardians to ask questions around their concerns of the criminal justice process. Since the PLOs were interviewed, in July 2020 their role changed to include a crime recording and case holding element. Due to the unique nature of The Lighthouse the PLOs have identified numerous cases whereby a child has disclosed sexual abuse, but this had not been shared with police contrary to S 2.4.18 Children Act 2004. The PLO will now record rape and sexual assault allegations whilst protecting the best interest of the child and optimising the likelihood of them feeling able to engage with a criminal justice investigation at their pace. When asked about what would be different at the Lighthouse if the PLO role did not exist; both PLOs suggested that it would have a detrimental impact on the victim’s experience and support, and the flow of communication between the police and the Lighthouse. 22 The SCLO Like the PLOs, the SCLO role is a unique feature of the Lighthouse and is a role that was described as a key role in providing expert advice and support in safeguarding to Lighthouse staff and local social care teams. They have extended this role during the pilot to include extensive training of professionals and schools, as well as consultation for social workers and other referrers prior to the Lighthouse accepting the referral. This can sometimes mean deferring a referral until the child and family are ready. They can also escalate issues in safeguarding practice and pathway processes within and outside of the Lighthouse’. The SCLO is not a case holder and is not responsible for individual children or young people38. The importance of the triage role, its evolution, and the failure to anticipate the centrality of this function at the outset of the pilot, was stressed by the SCLOs themselves (and which had been identified in the previous evaluation report). They noted the particular importance of triage where inappropriate referrals to the Lighthouse had been made (i.e. where there had been an element of sexual abuse but where support from the Lighthouse was not appropriate). All referrals come to the SCLOs for triage. They suggested that part of their role was challenging local authorities around decision making. The SCLOs suggested that their role was critical in preventing referrals to the Lighthouse being made too early and stressed that it was the quality of referrals that was important rather than the number. Decreases in referrals from boroughs (particularly Camden and Islington) might reflect the fact that the SCLOs had more involvement at an earlier stage with these boroughs. They also stressed the importance 38 Each child and family has a social worker in their borough Children’s Social Care (CSC) team. ‘Our role breaks down barriers and misunderstandings and if we weren’t here that would continue and be confusing for the victim and their family, that the police officer was saying one thing but the Lighthouse worker was saying something different, there wouldn’t be working together in the best interest of the child’ - PL0 23 of their role in providing safe-guarding advice, particularly to the CAMHs teams, who held what were termed ‘risky cases’ but who were not social workers. As one SCLO said; ‘if someone has an underlying mental health [issue], the Lighthouse is probably not the place even if sexual abuse is the underlying [factor], it’s probably not the place for that complex young person. So, what happens with that referral is doing support work for the team, professionals meeting, supporting CAMHS locally, that’s quite a bit of work, it takes a lot of work to think about timing’. Similarly, with schools the SCLOs wanted to ensure that they did not divert away from standard reporting processes by going straight to the Lighthouse but went to social care first. A key part of the SCLOs role is also to provide consultations, attend strategy meetings and offer signposting .39 In terms of the sources of the consultation, the greatest number came from social care, followed by Lighthouse staff, and mental health, as well as the police, support workers and GPs. There were a variety of reasons cited as to why the consultation had taken place, reflecting the multiple roles played by the SCLO. Amongst the specific reasons given were: advice to CAMHS (around referral to local authority, need for an immediate MASH referral); consultation with Lighthouse staff about liaison with social worker around the young person’s entitlements; consultation and written support in drafting a letter to the local authority setting out safe-guarding concerns around insufficient safety for therapeutic work; advice to Lighthouse staff around local authorities’ duties around accommodation and support for a child under the Children’s Act 1989 (due to concern about the closure of police investigation, and the suspect being at home with the victim); advice to Lighthouse staff and written support in drafting a letter to mother concerning limits of advocacy, and signposting in relation to complex housing situation. Ways of working It was suggested by some staff that the pressure to take referrals at the Lighthouse, to ‘keep the momentum and the numbers up’, meant that there were occasions where the information required by the Lighthouse was missing from the referral at the point that it was initially received, leading to the information having to be sought retrospectively - linking back to the SCLOs’ point about the quality rather than the number of referrals being critical. This did not only impact on the SCLOs. Because the latter did not talk to families, it was up to practitioners to do this, seeking consent from families for example, and who could be left holding cases until the information was forthcoming. It was recognised that there remained different levels of awareness about the Lighthouse in the various boroughs, linked to pre-existing networks of contact, but also the fact that there was a large turnover of staff in local authorities and the police, which pointed to the need for continued external engagement by the Lighthouse. The Department for Education (DFE) funded work for increased outreach work with schools 39 The Lighthouse’s own monitoring data suggests that the SCLOs are involved with around 30 strategy /signposting discussions and 39 consultations a quarter, although the variation quarter on quarter is much more marked for the latter (ranging from 28 to 56 over the last 3 quarters). However, looking at more detailed figures E&I obtained directly from the SCLOs, albeit only for one month, this appears to be something of an undercount. These figures indicated that in July 2020 the two SCLOs undertook 51 consultations during the 23 working days during the month, an average of 2 per day, with a maximum of 7 per day (twice), and there only being two working-days when either of the SCLOs were not involved in a consultation. 24 that the Lighthouse is about to undertake, and the fact that existing members of Lighthouse staff will be involved in this outreach work, will clearly be relevant in this regard. Some remedial soundproofing had been undertaken at the Lighthouse to address the issues of sound carrying between rooms/floors raised in the previous evaluation. In addition the decision to add more desks to the mezzanine level at the Lighthouse, sacrificing the 1st floor staff rest-area40, meant noise carrying from the staff kitchen down to the Lighthouse’s reception on the ground floor was no longer a problem, and had increased desk capacity at the same time (another problem that had been previously identified).41 However, from staff comments, soundproofing still remained an issue, and the proposal to use ‘white noise’ to combat noise pollution in the Lighthouse had not yet come to fruition. Similarly, while there had been improvements to the Excelicare system, and more were on-going, it’s limitations still emerged as a matter of concern. In addition, there had been delays in the delivery of 9 laptops (due since March) to assist in homeworking during Covid-19. Frustrations arising from delays in the implementation of the Live Link system were identified too. In terms of the impact of Covid-19 on ways of working it was recognised that the focus under lockdown had increasingly been on staying connected with the child or young person rather than delivering therapeutic work. There had also been a reduction in the use of measures of stress and anxiety (for example the Trauma Symptom Checklist for Children (TSCC) and the Revised Children's Anxiety and Depression Scale (RCADS)) due to difficulties completing detailed evaluation questions over video link due to reduced ability to engage with the child, issues around confidentiality in their own home for CYP when answering, and shorter appointments generally. Staff had tried to arrange more frequent but shorter contacts with children and young people – often twice a week using multi-media approaches including telephone, text, Attend Anywhere, Zoom, voice memos or video messages, and WhatsApp. Clearly, in these circumstances there was the need to consider confidentiality for CYP and families in sessions during lockdown. The latter had meant, for example, arranging simultaneous calls for the parent and young person with two separate practitioners so conversations could not be overheard, or young people talking on the phone during their daily exercise outside the home, or text chat instead of talking so they cannot be overheard. There have been obvious repercussions arising from the social distancing requirements at the Lighthouse. While the IA is offered daily face to face, some practitioners, particularly those who are self-isolating or shielding, attend virtually. In the medical rooms on the fourth floor 40 Staff were now sharing the NSPCC’s canteen facilities on a different floor 41 Pre-existing concerns about the lack of desk-space at the Lighthouse have currently been rendered redundant by the restrictions on the numbers of staff who are able to attend the building due to social-distancing restrictions under Covid- 19. ‘We had to do thorough risk assessment and, following UCLH advice limited to 50% desk occupancy and a maximum of 5 people in a meeting and everyone else joins virtually. We usually have 24desk spaces and now we have 15. It means we have a rota of one or two people from each team. Depends on who is in for IA or has children to see’. - Lighthouse manager 25 of the Lighthouse staff are having to wear full PPE because of the limited space and the fact that family members and professionals are in attendance. Children are given a choice about whether to wear masks. From conversations with staff it appeared that, even if a vaccine was available, it was likely that some of the working arrangements that had been introduced because of Covid-19 would continue longer term. It was recognised that lockdown had had a particular impact on the ability to undertake specific elements of the Lighthouse’s work; routine medical examinations, therapeutic work (Letting The Future In (LTFI) for example)42 and the impact and delays in the criminal justice process had impacted on advocacy’s ability to move cases. There had also been problems with IT, losing the signal, children not wishing to go on video. As one member of staff said: ‘I’ve used social worker direct tools and do some exploring that way but that’s the extent of what we can do virtually’. Staff mentioned that although they had been able to undertake virtual IAs they were preferable face to face; ‘I did one virtual IA and found it quite difficult, classic connection issues, not feeling as able to create that safe space for the young person to talk and less ability to be part of wider discussions. So, I had conversation with young person and thought through recommendations and then when I did manage to speak to rest of the team there was so much more information. It’s much easier to have those conversations as you’re going in the same building and it’s more difficult virtually’. The difficulty of undertaking work from home was also identified – both for the child and for the practitioner. One of the foundations of the LTFI model, for example, was described as providing a ‘safe space’ ‘which not always possible with them at home. And bringing that into our home as practitioners, and I live in a house share. In terms of boundaries [it’s] quite difficult’. As a result, work under lockdown was described as ‘checking in’ rather than therapeutic work. Much of the LTFI work, for example, was described as ‘play based’ – ‘it’s 42 LTFI is a service to help children who have been sexually abused rebuild their lives, through activities such as play- therapy https://learning.nspcc.org.uk/services-children-families/letting-the-future-in ‘I think IAs, follow up and therapeutic work and play needs to be done face to face but we will continue to use virtual methods for virtual meetings, and consultations we offer. Might start with a virtual appointment first, a few weeks before IA. Once young people have come face to face, they might want to continue virtually. A lot of our therapeutic work needs to be done face to face’. - Lighthouse manager ‘In early Covid most of support was more stabilisation and coping with Covid changes so lots of therapeutic work was put on hold. Contacting children more frequently than before to help them cope with Covid and other underlying issues, then offered more therapeutic work after that’ - Lighthouse staff member 26 not so difficult with older children but with younger children it’s so play based which is not possible virtually’. In addition, the detrimental impact of Covid-19 and the changes in working conditions on staff well-being was noted, as well as the problems it presented for new- starters in settling into the team and getting to know people. The perception from staff was that Covid-19 had not made a great deal of difference in terms of contact with external agencies, a view echoed by the borough liaison officers’ perception of the impact of Covid-19 on the work of the SCLOs. While face to face work had proved to be impossible, the SCLOs had still been available via Teams/telephone – ‘what [the SCLOs] have managed to do is offer social workers more frequent consultations to upskill the social worker when speaking to the young person. Giving the social workers prep work and the conversations she can have to make the young person ready when Covid-19 is more lifted. During Covid-19 I’ve had no issue with them attending any meeting. And they’ve been really keen to keep comms up on cases which are on hold a little bit’. Partnership working Experiences of partnership working within the Lighthouse The previous evaluation report identified the benefits that practitioners expressed about the partnership model that is central to the Lighthouse model, both for the service user in terms of the co-location of different disciplines under one roof, and for practitioners, in having timely access to the expertise possessed by other members of staff, although both have been clearly impacted by the introduction of remote-working as a result of the Covid-19 pandemic. Equally, it was suggested that, having been previously co-located in the Lighthouse made it easier to maintain links between different disciplines when working from home. However, the previous evaluation report also identified the challenges that arose from bringing individuals from a range of disparate organisations together (differences in language, terminology or organisational cultures that could result in working tensions). At the time this was described as being part of the process of developing a ‘Lighthouse way’ that sought to address these tensions. However, in speaking to staff, it was clear that this process is still on- going, and that frustrations remain; ‘it feels like a rollercoaster, you think you’ve resolved something and worked through a process and something’s happened, and we’ve come out at a better place only for something random to come along and unsettle all that. The cultural thing has been more significant than anyone could have imagined, and it’s a jostling between cultures’. There remained a feeling on the part of some staff that the Lighthouse was very much a ‘medical’ model and that there was a lack of understanding of the social work role. Some members of staff felt that there was a hierarchy at the Lighthouse with CAMHS at the top and LTFI and advocacy at the bottom. Concerns were expressed about how senior management interacted and friction identified between NSPCC and CAMHS in terms of the way they operated, with claims of there being inflexibility (clearly linked to the discussions about caseload and capacity outlined below). Levels of staff dissatisfaction had been reflected in the high levels of staff turnover particularly in the advocacy team. This is something that will be reflected upon in the final evaluation. 27 The capacity of the different services working at the Lighthouse had been identified as an area of concern in the previous evaluation report, and this remained an issue43. Different services had different agreed caseloads per member of staff, which led to claims of inflexibility, and, in addition, the ability for individuals within the Lighthouse to allocate cases to other teams led to difficulties in planning resources – as one member of staff remarked ‘I’ve found cases allocated to our team and letters already gone out’… my capacity has been impacted by infiltration’ (meaning Concern about the caseload levels of LTFI workers at the Lighthouse had led to senior staff removing their workers from the IA meetings in order to preserve their time for therapeutic work. Ironically, this led to concerns on the part of other LTFI staff that the service was becoming overly CAMHS led, because of the lack of an LTFI presence at the IA. It was suggested that it was difficult to talk about these issues openly at the Lighthouse; ‘not allowed to mention capacity, as that would mean we’re ‘managing’ and we’re not’. Problems arising from cases not being closed were also raised – ‘deal with front door by looking at back door and making sure cases are closed’, but large numbers of open cases remained, as shown by the figures in the appendices of this report. Interestingly it was suggested that working online might help to address the capacity issue, by increasing throughput, and make medicals more efficient. The need for better feedback and communication from the Senior Leadership Team was identified as an issue in the earlier evaluation report, and a programme of work had been implemented by management to address staff concerns. In the late Summer of 2019 work had been undertaken with the Lighthouse team to develop visions and values for the Lighthouse, and, in November and December 2019 an external consultant had been brought in to work with staff. As one member of management remarked; ‘It would be fair to say it opened up lots of issue and lots of concerns, some which were painful for some staff, it didn’t successfully resolve them, being a group of 30 or more staff it wasn’t the right forum or the right way or place to resolve those issues’. There had subsequently been a change of approach, and a change of facilitator running events and workshops with senior leadership (delivery board, senior leadership team, clinical leads and heads of service), the rationale being that ‘the senior team needed to be a cohesive strong unit before we could expect to model that with the whole team’. However, it was apparent from the staff interviews that issues around diversity and inclusion remained, which will be picked up in the final round of staff interviews. What had emerged from this work was a change to the management structure of the Lighthouse, involving the addition of an inward facing Lighthouse Service Manager, responsible for the day-to-day management of the Lighthouse, with the current Delivery and Service manager retaining the outward facing Strategic Lead role, sitting on the Lighthouse Delivery Board. As well as changes to the Lighthouse leadership to create Heads of Service: office manager, an SCLO, PLO, health team lead, CAMHS lead, LTFI/P&R lead, and an advocacy lead (details of the proposed structure are provided in Appendix C). Clearly there is an expectation on the part of Lighthouse senior management that the proposed structural changes, the addition of a service manager responsible for day to day management of the Lighthouse, (to whom everybody reports and to oversee every part of the service), and the inclusivity of the expanded senior team, will diminish the perception of some staff that they 43 Figures for the Lighthouse for the 3 quarters from October 2019 to June 2020 in relation to open cases by service, caseload by service and vacancy levels by service are provided in Appendix C of this report. 28 are not being listened to. The new structure and clearer roles and responsibilities will provide better access to and knowledge of decision-making processes within the Lighthouse. The final evaluation report will examine whether this reorganisation does indeed address the concerns that have previously been expressed by staff. Perceptions from external agencies The previous evaluation report contained feedback collected by the Lighthouse about CYP’s and parents’ perceptions of and experiences at the Lighthouse, as well as the findings from a survey of professionals indicating their level of knowledge about the Lighthouse. In general, this feedback was positive around the general perceptions and experiences, with parents and CYP feeling that they had been listened to, that their views and worries had been taken seriously, and that it was easy to talk to Lighthouse staff. In terms of the experiences of service users, the final evaluation report will contain detailed findings from interviews with service users. This interim report focuses on collecting feedback on the role of the Police Liaison Officers (PLOs) and Social Care Liaison Officers (SCLOs) at the Lighthouse, to gain a deeper understanding of the two positions. Data were gathered from police officers on their perceptions of the Lighthouse and the role of the PLOs there, and about the Lighthouse SCLOs via a focus group with social care staff from the local boroughs who work with the SCLOs. The police officers spoken to had heard about the Lighthouse from a variety of sources; word of mouth, email from supervisor to the unit, from colleagues working in the child sexual abuse arena and via promotional material /visits by Lighthouse staff. Most of the contact the police officers had had with the Lighthouse was either in relation to the police undertaking ABE interviews at the Lighthouse, or where their victims had been referred to the Lighthouse for ongoing support, but where the initial ABE had taken place at a police station. The police officers were asked for their perceptions of the psychologist led ABEs offered at the Lighthouse. While the majority did not offer an opinion, of those that did, more were supportive than not; ‘Obviously, every investigation is different, and every child is different and what we found is that in some investigations the child would have a better relationship with the social worker, sometimes they would have a better relationship with the allocated police officer, and investigations would progress differently within each investigation. So, for example, if the young person didn’t want to speak directly to the police officer now about what happened, but would be happy speaking with the psychologist, surely that is the best thing to do to get the best evidence from the child?’ The one individual who was opposed did so on the basis that it slowed the process and was concerned that it wouldn’t focus on evidential needs. Similarly, with one exception (where the officer concerned felt more comfortable interviewing in his own police station and was concerned about time delays in getting the victim to the Lighthouse), the police officers interviewed were extremely complimentary about the facilities offered at the Lighthouse, commenting on their quality and its ‘child- oriented’ nature, contrasting it with police VRI suites. The PLOs were commended for their work in arranging VRIs at the Lighthouse and for ‘chasing up’ and facilitating the cases. More widely the officers also commented on the level of support and assistance provided for the 29 children and families by Lighthouse staff during the VRI; ‘I have no words, they are fantastic. They are just an absolutely fantastic resource to have’. The responses obtained from the focus group undertaken with the liaison officers from the local authorities involved in the pilot stressed the importance of the Lighthouse as being a ‘valuable resource’, but also the centrality of the SCLO to the Lighthouse where they were described as ‘the glue between social care and other professionals, paediatricians, psychologists within the Lighthouse’. Interestingly, when asked if the SCLO role could exist independently of the Lighthouse, the feeling was that this was not the case; ‘only exists if Lighthouse is there, personally. If it was case consultation then that is about a specific case, that would be a specialism in a local authority, we have a lead on sexual exploitation. What [the SCLOs do] is bring in the whole multi agency network and liaise between the two which is connected to the LH’. A further respondent commented ‘I can’t imagine life without the Lighthouse, the thought of [the SCLOs] being in a meeting is the same as having NSPCC medical CAMHS in the meeting, we wouldn’t be able to pull those people together in a timely way to have a meeting like that. Terrifies me the thought of it not being there’. Factors emphasised were the knowledge, flexibility (‘they’ve never put up a wall and said ‘that’s not in our job description’) and speed of response of the SCLOs (‘they answer the phone and email back within a few minutes’). The respondents stressed their usefulness in offering suggestions and guidance about how to navigate referral pathways, and their role in terms of quality assurance; ‘they know the policy the processes so well that I’ve had many useful conversations where they say ‘you haven’t’ done these stages’ and actually they are always right. They do tell us what we should have done and what we should be doing, so I’ve seen an improvement across the service’. The importance of the level of experience that the two SCLO’s possessed was also stressed; ‘they need to [be experienced] as there will be clashes of opinion, liaising with NSPCC, we rarely agree on a case and they were the ones who were having to mediate and work through it with all of us. The consultations are really high quality it’s really thinking about stuff, they need to have a sense of authority and the knowledge that comes with that’ and there was a recognition that there was not the resource to undertake this role ‘in-house’. Looking ahead In terms of the development of their role, the PLOs believe they would start to have less involvement in training the psychologists on the basics for ABE interview, and focus on the research and further developing interview best practice and seeing it implemented. Additionally, the impact of S28 and how to continue to seek to improve the criminal justice experience for the victim. They would also see benefit in being more involved with the young people at the initial assessments, so that children have the opportunity to speak to a police officer outside of their case and improving the overall perception of police and police officers. This also links into the perceived benefits of conducting ABE interviews at the Lighthouse “rather than a police station”. 30 As with the previous evaluation report, when staff were asked to identify the main challenges faced by the Lighthouse in the future the most popular responses were the uncertainly about future funding/sustainability, concerns about the ability of the service to respond to the anticipated increase in demand post-Lockdown once the schools had re-opened44 and the extent to which the reorganisation at the Lighthouse could address staff concerns around the ‘Lighthouse way’. The final evaluation will seek to explore many of these issues, as well as consider the strengths and weaknesses of different models of working for the delivery of CSA services. 44 Although at the time of writing (October 2020) the Lighthouse was reporting that the anticipated increase in referrals had not taken place. The trend in referrals in the period after the return to school of children in Autumn 2020 will be examined in detail in the final evaluation report. 31 4. Discussion The focus of this report has been to highlight performance and process findings relating to the first 21 months of the Lighthouse’s implementation (end of October 2018 to August 2020). Data for this report was gathered from several sources, including the Lighthouse’s performance management data, case management system, qualitative interviews and focus groups with Lighthouse staff, police officers and social care workers. Consistent with findings from the previous evaluation report, it is evident from the variety of data collected (medical histories, risk assessments and ACE-Q scores) that this is a highly vulnerable group of children and young people with complex needs, reinforcing the need for a service that integrates multiple strands of support to appropriately respond to the specific requirements of the CYP. The evaluation (and of course the service) have been heavily impacted as a result of Covid-19 and the national lockdown that was imposed in March 2020. Such a pandemic presents substantial challenges for both service delivery and any subsequent evaluation. The initial service delivery model of the Lighthouse was not designed for delivery in such an environment and this evaluation report has sought to address service implementation both before and after Covid-19. Due to the impact of the Covid-19 pandemic, staff have had to make large changes to their ways of working and how they interacted with and supported the CYP, by conducting sessions virtually and staff arranging more frequent but shorter contacts with children and young people, often twice a week. Additionally, staff had to consider the issue of confidentiality for CYP and families in virtual sessions. The Lighthouse staff have shown a remarkable degree of adaptability and flexibility in continuing to offer a service to young people under lockdown initially, and subsequently as the service has moved to a hybrid model. Interestingly, some of these changes, particularly the use of technology to facilitate virtual meetings between members of staff, external agencies, and CYP and parents in certain circumstances, appear likely to be retained once the situation returns to normal. How these have been maintained will be included in the final report. However, it is also clear that certain aspects of the Lighthouse’s work require face to face working, and the ability of the organisation to deliver the intended service has been disrupted since March 2020. Aside from the Covid-19 changes, since the previous evaluation report there have been some other progressions made, and as such the psychologist-led ABE interviews at the Lighthouse have begun (13 in the last 9 months). The use of the Live Link facility has still not started, however positive steps have been made in this area. Additionally, other operational practices at the Lighthouse have continued to evolve such as reducing the number of allocation meetings (previously held daily). In terms of the operational model, this report focused on the potential value of two operational roles unique to the Lighthouse; the PLO and SCLO. For the former, their role in liaising with external police officers and the CPS around the progression of investigations, and the facilitation of police and psychologist led ABE interviews at the Lighthouse was highlighted by respondents. For the latter, the part they played in ensuring appropriate referrals to the Lighthouse was emphasised, as well as the importance of the guidance and advice around safe-guarding and social care pathways provided to Lighthouse and local authority staff. Both roles were seen as essential to the Lighthouse model and received positive feedback externally. 32 The levels of referrals to the service have continued to reflect the pattern observed in the last evaluation report. Since the project began, there has been an average of 29 referrals to the Lighthouse per month, while the number of IAs carried out has averaged at 18 per month, although there was a decline in referrals and IAs in the early stages of Covid-19, for understandable reasons although these numbers have recovered since the early days of lockdown. Most referrals continue to be received from social care, and the number of self- referrals remains relatively small. It will be interesting to see if the number of referrals received from schools increases as a result of the increased liaison work about to start between the Lighthouse and local schools with DfE funding (which will be evaluated separately). Staff are concerned about the impact of increased referrals as a result of schools re-opening after Lockdown releasing previously suppressed demand, although at the time of writing, the anticipated increase had not materialised. There remain tensions in partnership working at the Lighthouse, from the different disciplines working there, and the perception on the part of some staff that the medical model predominates, and social care is undervalued, continues. Attempts by senior management to address these concerns have had limited success to date but are on-going, with a restructure of the Lighthouse in progress. The final evaluation report will examine the impact of these changes on staff perceptions. While there has been some remedial work since the last evaluation report to improve aspects of the Lighthouse’s infrastructure (soundproofing and the case management system for example) it is clear that these remain an issue of concern but will continue to be monitored. Interestingly some of the changes in working practices arising from Covid-19 (particularly social distancing requirements) have changed the nature of some of these concerns, however this situation will be temporary. Consistent with the previous findings, staff also remain concerned about issues of sustainability and the future of the service beyond the end of September 2021. Next steps The current document is the third published report within the MOPAC Evidence and Insight evaluation of the Lighthouse. A final evaluation report is planned for the summer of 2021, at which point it is anticipated that details of many more service users will be available for analysis, as well as ‘impact’ data on the criminal justice and health/well-being outcomes of service users, and the perceptions of service users themselves. 33 References Bellis, M.A., Hughes, K., Leckenby, N., Hardcastle K.A., Perkins, C., & Lowey, H. (2014) Measuring mortality and the burden of adult disease associated with adverse childhood experiences in England: a national survey. Conroy, L., Hobson, Z., Parker, R., & Read, T. (2018) The Lighthouse: London’s Child House Initial Evaluation Report. MOPAC Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A. M., Edwards, V., Koss, M.P., Marks, J. S. (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) Study. NSPCC (2019) Statistics briefing: child sexual abuse. Parker, R., Read, T. & Dawson, P. (2020) The Lighthouse: 9-month Interim Evaluation Report. MOPAC Vaswani, N. (2018) Adverse Childhood Experiences in children at high risk of harm to others. A gendered perspective. Centre for Youth & Criminal Justice. Zarse, E.M., Mallory, R.N., Yoder, R., Hulvershorn, L., Chambers, J.E. & Chambers, R.A. (2019) The adverse childhood experiences questionnaire: Two decades of research on childhood trauma as a primary cause of adult mental illness, addiction and medical diseases. 34 Appendices Appendix A: Total number of referrals by borough, between October 2018 and July 2020 Monthly Referrals by Borough Year Month Barnet Camden Enfield Haringey Islington Other London Borough Outside of London Total 2018 Oct 3 7 2 3 15 Nov 7 9 4 10 10 1 41 Dec 4 9 10 10 7 2 42 2019 Jan 4 9 6 6 5 30 Feb 5 3 6 2 7 23 Mar 3 6 6 8 7 2 32 Apr 5 7 7 6 5 5 35 May 9 4 9 5 8 35 Jun 1 5 8 3 6 1 24 Jul 4 8 6 8 4 30 Aug 8 5 4 6 3 26 Sep 9 5 9 3 6 1 33 Oct 15 7 5 4 2 1 34 Nov 12 3 3 6 4 28 Dec 5 4 3 4 7 23 2020 Jan 13 6 4 4 1 28 Feb 10 6 5 5 4 1 3 34 Mar 10 3 8 8 5 34 Apr 6 2 3 3 14 May 5 2 5 3 4 19 Jun 4 3 9 6 7 29 Jul 4 4 9 7 5 1 30 Total 146 111 130 117 116 16 3 639 35 Service user vulnerabilities Type of Vulnerability Number of service users Anxiety/Depression 96 Other vulnerability 89 History of DV 86 Education problems 61 History of self-harm 55 CSE 43 Concerns over safety 42 Suicide risk 34 Drugs/alcohol 26 Sexualised behaviour 24 LA care order 22 Risk of further harm 16 Eating disorder 11 Missing from home 13 Repeat attender 0 ACE scores ACE category Number of service users Sexual abuse 139 Parents divorced 77 DV against mother or father 54 Emotional abuse 54 Emotional neglect 44 Parent mental illness 40 Physical abuse 33 Parents alcohol or drug issues 30 Physical neglect 27 Parent in prison 11 Total clients with ACE assessment 162 36 Advocacy Goals full breakdown Type of goal Achieved Not Achieved Partially Achieved Total Criminal justice 19 1 16 59 School support 14 5 7 33 Other 11 0 4 16 Healthy relationships 5 1 1 13 Build confidence 5 10 Advocacy support 5 1 8 Health and wellbeing support 4 8 Safety advice 4 4 8 Emotional need support 2 4 7 Accommodation 1 2 1 6 General support 2 1 6 Medical support 4 1 1 6 To attend support sessions 2 1 2 6 Support for referral to CAMHS 4 5 Support for parents 3 1 4 Help to engage in community activities 1 1 3 Communication skills 2 To engage with the Lighthouse 2 To talk to someone 1 2 Total 85 14 43 204 37 Appendix B – Lighthouse open cases, caseload, and vacancy rates (Oct 2019 – June 2020) Number of open cases each quarter for each Lighthouse service. The number of open cases for the various services provided at the Lighthouse obviously reflects the nature of the work done by the various services, and the nature of the contact with the CYP. The figures above are taken from the Lighthouse’s monitoring returns. Overall, paediatricians and CAMHS practitioners have had the highest numbers of open cases. Most notably, the number of open cases in CAMHS saw a 70% increase in cases (n=60) between the first and final quarters to June 2020, with a particular increase in the time during the Covid-19 lockdown. Similarly, the open cases for the paediatricians have increased over the nine months (albeit peaking in the middle quarter) with similar numbers to CAMHS. On the other hand, the figures for advocacy have remained consistent across the three quarters at 90+ cases. The sexual health nurse saw a 40% increase in open cases from the first quarter to the second, but the figure has now fallen by 36% in following quarter to 65 cases (which represents a 10% decrease overall). Conversely the figures for LTFI/P&R staff have consistently decreased over the 9 months (by 35%) down to 34 cases, while that for the play specialist has decreased (by 61%, particularly in most recent quarter) down to 14 open cases45. Caseload and vacancy rates The tables below, also taken from the Lighthouse’s monitoring returns, provide the average caseload and the vacancy rates for the various services over the same 9 months. 45 With all these figures Lighthouse staff felt that the numbers reflected issues related to data collection, particularly delays in closing cases on the system. 0 20 40 60 80 100 120 140 160 Advocacy CAMHS LTFI & P& R Paediatrician Play Specialist Sexual Health Nurse Oct '19 to Dec '19 Jan '20 to Mar '20 Apr '20 to Jun '20 38 Average Caseload/WTE in period Service Oct - Dec 2019 Jan - Mar 2020 Apr - Jun 2020 Paediatricians (1WTE) 113 149 126 Sexual health nurse (1WTE) 54 96 56 Advocacy (6.8WTE) 13 12 14 CAMHS (5.6WTE) 18 21 23 LTFI/P&R (5WTE) 11 10 10 Play support (0.6WTE) 52 62 23 Vacancy rates Service Oct - Dec 2019 Jan - Mar 2020 Apr - Jun 2020 Medical/nursing 0% 8% 5% Advocacy 0% 14% 0% CAMHS/LTFI/P&R 15% 0% 33% Management and Business Support 0% 0% 0% 39 Appendix C – Restructure of Lighthouse, September 2020 Lightouse Clinical Lead - VACANT Lighthouse Service manager - NEW POST Office Manager - Meta Madden Admin team Data Officer Social Care Liaison Officer - Martin Slack Social Care Liaison Officer - Eimear Timmons Police Liaison Officer - Lisa Isaacson Police Liaison Officer - Jon Guy Health team lead - Deborah Hodes Consultant paediatricians Clinical nurse specialist - SH Play specialist CAMHS team lead - Victoria Mattison Consultant Psychiatrist - Rob Senior Clinical psychologists CAMHS practitioners CAMHS trainees LTFI and P&R team lead - Rachel McPate LTFI practitioners P&R practitioner Advocacy team lead - Margaret Galloway Advocacy team (NSPCC and Solace Women's Aid) Solace service manager – Liz Brailsford Head of SG, Camden LA – Sophie Kershaw Assoc. Clinical Director, TPFT – Steve Bambrough NSPCC service manager – Marian Moore Det Insp, Met Police Service - Tracy Cormack Lighthouse Delivery Board Lighthouse Senior Leadership Team Lighthouse Senior Leadership Team Lighthouse teams Lighthous e Strategic lead – Emma Harewood UCLH Divisional Manager – Felicity Hunter 1 The Croydon Drive Project: A 2-year Evaluation Final Report Sophie Geoghegan-Fittall, Charlotte Keeble & Daniela Wunsch MOPAC Evidence and Insight September 2020 2 Contents Acknowledgments ........................................................................................................... 3 Executive Summary ......................................................................................................... 4 Introduction .................................................................................................................... 8 Methodology ................................................................................................................. 15 Results: Learning from year two of the Croydon Drive Project ................................... 16 Project throughput and activity ................................................................................................ 16 Understanding the cost of domestic abuse and Drive ............................................................... 24 Further embedding Drive in Croydon: Year two ........................................................................ 25 Working with perpetrators and victim-survivors ........................................................................ 29 Looking ahead – The future of Drive ........................................................................................ 34 Conclusion ..................................................................................................................... 37 References .................................................................................................................... 40 Appendices .................................................................................................................... 44 Appendix A: Drive Model ....................................................................................................... 44 Appendix B: Drive Case Manager training summary .............................................................. 45 Appendix C: Drive referral guidance ...................................................................................... 46 Appendix D: Croydon Drive interim evaluation: Learning from year one of the Drive Project in Croydon ............................................................................................................................. 47 Appendix E: Fieldwork conducted to date ............................................................................ 48 Appendix F: Drive case manager reported outcomes ............................................................. 49 Appendix G: Initial typology of Drive cases (based on one Case Manager interview) ............ 52 Appendix H: Perpetrator ‘Footprint’ (based on a dip sample of twenty-five cases) ............... 53 3 Acknowledgments MOPAC Evidence & Insight would like to thank Laura Duckworth for her research support, Dr Paul Dawson for his advice throughout the project and Lisa Le Masson for her support and assistance in undertaking the evaluation. Many thanks also to Kyla Kirkpatrick, Emma Vallis, Martha Tomlinson and colleagues from the Drive Partnership for their support and assistance. MOPAC would also like to thank Steven Kelly and Alison Kennedy for their support in undertaking this evaluation. Finally, MOPAC would like to thank the practitioners, stakeholders and service users who gave up their time to support the Croydon Drive project evaluation. 4 Executive Summary The Drive Project went live in Croydon in July 20181. The project addresses the behaviour of high harm domestic abuse perpetrators’ - often those who would likely not have been engaged with before or excluded from alternative perpetrator responses - to improve outcomes for adult and child victims, and perpetrators. Initially delivered as a pilot in Essex, South Wales, and West Sussex between 2016 and 2019, in 2018 the project combined funding from the Home Office Police Transformation Fund and Police and Crime Commissioners (PCCs) to roll out to Croydon (London), Cardiff (South Wales), Birmingham and Sandwell (West Midlands) and Worcestershire (West Mercia) for an additional two-year replication phase until March 20202. Drive is currently the only high-harm domestic abuse perpetrator scheme commissioned by MOPAC. The Evidence and Insight (E&I) Unit - MOPAC’s in-house social research and analytical team - were commissioned to undertake an evaluation of Drive in Croydon. The two-year evaluation of the Croydon site examines two distinct areas: monitoring the performance of the service through the routine capture of core project metrics; and generating in-depth understanding of the processes - from design through implementation of the service - of those working to deliver the Drive project. This final evaluation report presents the second year of performance data and process learning and should be read in conjunction with the Drive Project Croydon interim report3. Key Findings Project throughput and activity • Over the duration of the project (July 2018 and March 2020), Drive in Croydon has worked with a total of 1704 perpetrators, 203 associated victim-survivors and 309 associated children and young people. This means Croydon delivered to 83%5 of the target set for this time period and is in line with the other Phase 2 replication sites6. This is likely due to a slower initial delivery pace during implementation, commonly expected during the set-up phase of projects. • Additionally, 799 interventions were recorded in relation to cases in the Drive service, including ‘support’ (n=135), ‘disrupt’ (n=100) and a combination of the two approaches (n=564). • In year two, 99 cases were accepted onto the programme - an increase of 26 cases compared to year one. These cases relate to 99 perpetrators, 114 survivors (1 in 10 cases were associated with multiple victim-survivors) and 184 children or young people. • Data collected in relation to all Drive service users since the project inception (n=170) indicates that three out of five (61%, n=103) presented with three or more needs. This compares to 44% of service users in the year one cohort only, which may be indicative of an increase in prevalence of multiple needs amongst service users, the result of improved multi-agency information sharing, and/or improved data recording practices in year two. During phase 2 of the pilot, replication sites began recording information relating to Adverse Childhood Experiences (ACEs) in the Drive Cohort7. 1 Croydon was selected to be part of the Drive project following a successful expression of interest that was sent to London boroughs with the highest police recorded levels of domestic abuse. 2 Referred to as the Phase 2 replication sites later in the report. 3 MOPAC (2019) Drive Project: An Interim Evaluation report from the Croydon site. 4 The two-year total number of cases may differ. It is possible that cases that were moved to referral did not progress. Any case that is classed as unsuitable is removed from SafeLives and Drive Partnership reports. 5 204 service users. 6 84% nationally across all Phase 2 sites. 7 Growing up with adverse childhood experiences (ACEs) such as abuse, neglect, community violence, homelessness or growing up in a household where adults are experiencing mental health issues or harmful alcohol or drug use, can have a long-lasting effect on people's lives. 5 Although the picture is incomplete due to missing data, in closed contact cases where more reliable data is available (n=30), one in three service users (33%, n=10) in Croydon had at least one ACE recorded. • ‘Children, families and parenting’ was identified as the highest level of perpetrator need and the most commonly identified need across all cases (high need = 52%, medium need = 15%, low need = 3%). This was followed by ‘Substance Misuse-Drugs’ (high need = 18%, medium need = 26%, low need = 4%), ‘Mental Health and Psychological Wellbeing’ (high need = 19%, medium need = 19%, low need = 8%), and ‘Housing’ (high need = 12%, medium need = 16%, low need = 14%). • Given its focus on high-risk, high-harm perpetrators who are harder to engage, Drive anticipates that direct contact is made in approximately half of cases, with non-direct work taking place in the remaining half. This is reflected in Croydon, where at the time of reporting 96 cases were open8 and in half of these (49%, n=47/96), Drive Case Managers had been able to make direct contact with the perpetrator to coordinate and deliver interventions. This is also in line with figures reported for Croydon in year one (46%) and with those collated from across all Drive phase 2 replication sites 9. • Sixty-eight cases were closed during year two, which compared to just four cases closed during year one, depictive of the longer-term nature of work associated with Drive. • During year two, a total of 694 multi-agency interventions were made in relation to open cases in the Drive service (n=96), a substantial increase in activity levels in Croydon compared to year one in which 105 interventions were made. Most interventions (73%, n=509) included a combination of both ‘support’10 and ‘disrupt’11, again, a considerable increase compared to year one where just over half (52%) of cases used the combined approach. This may be due to changes in local recording practices; improved understanding of the model; improved agency buy-in; or improved multi-agency working during year two. Further exploration of the increased number of interventions and use of both ‘disrupt’ and ‘support’ should be undertaken to develop the learning from Drive. • Year one performance data indicated that there were challenges in moving towards meaningful behavioural change work with perpetrators to address their attitudes and behaviour to domestic abuse - eleven of all open cases had received behavioural change support. By the end of year two, just under half (n=41/96) of all open cases had received behavioural change support, which points towards progress in this area. • Case manager reported outcome data suggests that Drive in Croydon may reduce physical, sexual, Harassment & Stalking (H&S) and Jealous, Coercive & Controlling (JCC) abuse, the effect of which is greater in cases where direct work with perpetrators is undertaken. Whilst based on a relatively small sample of cases (n=68), these reductions are consistent with those seen across other Drive replication testing sites and are also following the same downward trend found in the original pilot sites. • Given the relative infancy of Drive in Croydon, it has not been possible to undertake cost analysis at this stage in the programme. It is suggested that in order to support future cost analysis, opportunities to identify relevant data sources, collect quality data and improve current data recording practices are explored in Croydon. Further embedding Drive in Croydon: Year two • Overall, the Drive model continues to be viewed positively by practitioners and stakeholders in Croydon and was commended for filling a gap in support for high-harm perpetrators. 8 Data relates to all cases that were open for all or part of the reporting period, including any cases opened prior to the reporting period, and any cases that were closed during the reporting period. 9 As of 31st March 2020, 446 open cases across all Phase 2 replication sites, 45% of which had direct contact with service users. 10 This is defined as; ‘addressing perpetrator need with the aim of removing barriers to the change process’. 11 This is defined as: ‘putting barriers and obstacles in place to prevent abuse’. 6 • Although Drive was considered expensive and time-consuming by some participants - particularly by practitioners and those delivering the programme - most felt that it is worth continuing with if it can help stop the cycle of abuse and reduce risk. • Whilst most agencies and practitioners were supportive of the Drive model and working with offenders, some practitioners - particularly those working to support victims - remained sceptical, voicing concerns that Drive exposed an ‘imbalance’ between long-term perpetrator interventions and a more generic, short-term victim support provision. Although such scepticism is unsurprising, reflecting other research in this area, findings from this evaluation suggest that work still needs to be done to ensure all staff are aligned to the benefits of working with offenders – if not addressed such working tensions will continue. • Securing and maintaining buy-in is therefore a key enabler of the project and continuing to build evidence and learning will be vital for Drive in this respect, as well as other perpetrator programmes. • Whilst the Drive Central team have continued to provide a key support function to Croydon, during the second year of delivery there is evidence of movement towards more local ownership of the model, with Croydon reportedly leading and formulating solutions to problems and implementing local processes. • A multi-agency partnership approach continues to be at the ‘heart’ of Drive. Practitioners and stakeholders reported a greater understanding of each other’s roles and the levers they can bring and were particularly positive about the cooperative partnership which has developed with the police. • Challenges were recognised around stakeholder engagement and keeping up attendance levels at the Domestic Abuse Perpetrator Panel (DAPP). Practitioners reflected concerns that some agencies, particularly mental health and substance misuse, are becoming increasingly disconnected with the project as workload pressures seemingly result in a de-prioritisation of Drive - a concern given the prevalence of substance misuse and mental health associated needs within the Drive cohort. • Ways in which attendance at the DAPP could be improved were shared and learning from recent virtual meetings held during the COVID-19 pandemic could provide useful learning in this respect. • Issues around information sharing between agencies have seemingly persisted in year two. Thought should be given to how these issues should be addressed, such as reviewing local governance structures, prioritising the review, implementation of and training in information sharing guidance and processes and consulting an information sharing expert to support this work. Working with perpetrators and victim-survivors • Drive’s cohort is complex and challenging in terms of needs, with some service users presenting long, intergenerational histories of abuse and control. Further work to develop a more in-depth understanding of the Drive cohort in Croydon - and high-harm domestic abuse perpetrators more broadly - should be considered to further learning around the role of effective interventions for perpetrators of domestic abuse. This should include a thorough understanding of the histories and needs of perpetrators, as well as factors in relation to motivation, willingness and ability to engage with an intervention. • Whilst Drive was seen as most effective when it was using its armoury collectively, with ‘change, support and disrupt working together, not separately’, the ‘disrupt’ tool continues to cause some uncertainty for staff delivering Drive and further awareness raising, training, and documented guidance would be beneficial. • The role of the IDVA service in supporting victims was considered an essential part of the delivery of Drive. Despite positive elements of joint working between IDVAs and Drive case managers, some of the challenges of supporting victims as part of a perpetrator focused programme were highlighted. 7 • Going forward, thought should be given to further developing the eligibility and selection criteria for Drive in Croydon, to provide clarity for both new and existing staff and agencies, and to ensure and maintain project integrity. Looking ahead – The future of Drive • Uncertainty around the future of Drive in Croydon was expressed by participants, signalling the requirement for early sustainability planning and communication with those delivering Drive in Croydon. • There is a clear role for evidence and generating learning. Promoting success and demonstrating the effectiveness of Drive was seen as key to sustaining commitment from agencies and staff delivering Drive in Croydon - particularly those not working directly with perpetrators - as well as raising awareness of the project amongst new partners, new areas, and across agencies not directly associated with Drive, including: gangs units; Community Safety Partnerships; Housing departments (including housing associations); adult and children’s social care; as well as police working in non-domestic abuse areas. • The Drive replication pilot in Croydon has generated key learning for sustaining the project in Croydon, as well as future expansion and replication of the project in other areas. Several building blocks for future project planning are set out in this report and centre on: buy-in; evidence and learning; a multi- agency approach; a model tailored to the local landscape; and, skilled and dedicated operational staff. 8 Introduction Background According to the Crime Survey for England and Wales12, an estimated 2.4 million adults aged 16 to 74 years experienced domestic abuse (1.6 million women and 786,000 men) in the last year. During this time, the police recorded 746,219 domestic abuse-related crimes - an increase of 24 per cent from the previous year (ONS, 2019). Turning to London, in the same period to March 2019 there were 88,278 recorded domestic abuse offences across the Metropolitan Police Force area, an increase of 12 per cent (+9,177) from the previous year. A Home Office report published in January 2019 estimated that the cost of domestic abuse for victims in England and Wales for the year ending 31st March 2017, was approximately £66billion; the largest component of this estimated cost was the physical and emotional harm experienced by victims, which accounted for an estimated £47million of the cost13. Tackling violence against women and girls (VAWG) is one of the key priority areas in the Mayor’s Police and Crime Plan 2017-202114. The Violence Against Women and Girls Strategy 2018-202115 sets out a commitment to embed the Drive project; challenging perpetrators of domestic abuse to change their behaviour. The Drive Project went live in Croydon in July 201816. The project addresses the behaviour of high-harm domestic abuse perpetrators’ - often those who would likely not have been engaged with before or excluded from alternative perpetrator responses - to improve outcomes for adult and child victims, and perpetrators. Initially delivered as a pilot in Essex, South Wales, and West Sussex between 2016 and 2019, in 2018 the project combined funding from the Home Office Police Transformation Fund and Police and Crime Commissioners (PCCs) to roll out to Croydon (London), Cardiff (South Wales), Birmingham and Sandwell (West Midlands), and Worcestershire (West Mercia) for an additional two-year replication phase until March 2020. Funding for Drive in Croydon has now been extended to March 2021, supporting delivery of the project for a further 12 months. This time will be utilised to explore both the sustainability of Drive in Croydon beyond current funding, as well as scoping and planning any further expansion of Drive within London. The Evidence and Insight (E&I) Unit - MOPAC’s in-house social research and analytical team - were commissioned to undertake an evaluation of Drive in Croydon. The two-year evaluation of the Croydon site only, examines two distinct areas: monitoring of the performance of the service through the routine capture of core project metrics; and, generating in-depth understanding of the processes- from design through implementation of the service – of those working to deliver the Drive project. The process- based element aims to generate findings around roll out and delivery of the project in Croydon (rather than the more outcome focused aims of the project set out above which were explored robustly in the Gold standard Bristol study - see below), to inform learning around replicability of the Drive model going forward. This final evaluation report focuses on the second year of performance data and process learning, outlining where there have been key changes and developments compared to year one. A summary of the findings of year one of the evaluation - which explored the initial set-up and implementation of Drive in Croydon - can be found in Appendix D17. 12 Data for the year to 31 March 2019. 13 HORR107: Oliver et al, 2019: The economic and social cost of domestic abuse 2019. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/772180/h orr107.pdf. 14 https://www.london.gov.uk/sites/default/files/mopac_police_and_crime_plan_2017-2021.pdf. 15 https://www.london.gov.uk/sites/default/files/vawg_strategy_2018-21.pdf. 16 Croydon was selected to be part of the Drive project following a successful expression of interest that was sent to London boroughs with the highest police recorded levels of domestic abuse. 17 MOPAC (2019) Drive Project: An Interim Evaluation report from the Croydon site. 9 Domestic Abuse Perpetrator Programmes: A Brief Review of the Literature Preventing and addressing the harm caused by domestic abuse is a focus of political and academic interest, both in terms of supporting and empowering victim-survivors and reducing the likelihood of further abuse. Domestic abuse perpetrator approaches can form part of the response to addressing domestic abuse, with flexible, trauma informed programmes that can adapt to individual needs and circumstances recognised within the Government’s recent Domestic Abuse Bill (HM Government, 2019). This requires a well- developed evidence base; however, research in this area is in its infancy, particularly in a UK context. Domestic abuse perpetrator programmes emerged in the UK in the late 1980s and early 1990s as part of the response to domestic abuse. Programmes are characterised by a range of approaches, often drawing on cognitive-behavioural and psychoeducational models, to hold perpetrators to account and offer opportunity to change, usually in a group work setting (Lilley et al., 2018). As programmes have developed, different approaches have been adopted, although theoretical frameworks remain similar, with the aim of reducing violence and abuse against women and children at the forefront (Kelly and Westmarland, 2015; Bates et al., 2017). The UK approach is largely influenced by the Duluth Model (Pence and Paymar, 1993), characterised as psycho-educational in nature, challenging men’s use of power and control over women; however, many programmes have now broadened their approach to recognise shortcomings of the model, incorporating elements of cognitive-behavioural therapy (CBT) and targeting the factors that impact on domestic abuse risk (e.g., emotional and relationship skills deficits, substance abuse, mental health, housing), and delivering programmes to both male and female perpetrators. Results from studies into domestic abuse perpetrator programmes that adopt experimental or quasi- experimental research methodologies have reported mixed findings, with interventions just as likely to prove effective in achieving their aims as not to18. For example, experimental studies by Dunford (2000), Feder and Dugan (2002), and Labriola et al. (2008) observed no significant difference in outcomes between treatment and control groups, while Taylor et al. (2001) showed a lower rate of recidivism for the treatment group than the control group, and fewer incidents of recidivism based on official criminal justice records; however, reported no significant difference when using victim self-report as the outcome measure. Likewise, while some have found no significant treatment impact (e.g., Gordon and Moriarty, 2003; Haggard et al., 2017; Jenkins and Menton, 2003; Mennicke et al., 2015; Rempel et al., 2008), others have reported positive – albeit sometimes small - effects (e.g., Blatch et al., 2016; Boots et al., 2015; Dobash et al., 1999; Hasisi et al., 2016). For example, Bloomfield and Dixon’s (2015) study evaluated two domestic violence interventions delivered by the National Probation Service: The Integrated Domestic Abuse Programme (IDAP) and the Community Domestic Violence Programme (CDVP). The authors found that both the IDAP and CDVP were effective in reducing domestic violence reoffending and any reoffending in the two-year follow-up period, with small but significant effects. For participants who did go on to reoffend, those who received treatment took significantly longer than the control group. 18 Domestic abuse perpetrator programmes have proved challenging to evaluate with difficulties defining success, capturing good quality monitoring information, inconsistent implementation, limited follow up periods for data collection, and small sample sizes, often leading to inconclusive results (HMIC, 2014; Bullock et al., 2010; Stanley, 2011; Price et al., 2009; Akoensi et al., 2012; Vigurs et al., undated; Lilley et al., 2018; Bates et al., 2017; Ali et al., 2017). 10 In January 2020, the University of Bristol (UoB) finalised its evaluation of the initial Drive pilot sites (Essex, South Wales and West Sussex) (Hester et al., 2019). The evaluation adopted a multi-method approach, with a random control trial design to assess outcomes, along with qualitative interviews to provide deeper understanding of Drive’s processes and practices. Findings highlighted the positive impact of the project; Drive was found to reduce the use of abusive behaviours by service users – particularly serial perpetrators, those using severe domestic violence and abuse, and those service users with a variety of needs. Furthermore, Drive helped to reduce high-risk perpetration, including by serial perpetrators, and this was sustained for a year after the cases were closed – longer than those in the control group not receiving the Drive intervention. Providing support to victims forms a key aspect of the Drive model; findings indicated that Drive increased safety for adult and child victims, doing so to a greater degree than in cases where only support to the victim is being provided. This gold standard evaluation is a valuable, recent contribution to the evidence base for the Drive model and, more broadly, perpetrator programmes. In the US, Cox & Rivolta (2019) employed a quasi-experimental research design with a propensity- matched comparison group to test the effectiveness of Connecticut’s batterers’ programme for serious, male, family violence offenders. Findings showed that the programme participation group (n=185) had significantly lower one-year arrest rates than the comparison group (n=185). While the effect size for any type of arrest was moderate, it remained when controlling for other variables known to influence recidivism such as age, race/ethnicity, family violence risk, and criminal history. These differences, however, applied to any type of new arrest but not specifically to family violence offenses. For family violence offences, the programme group had lower arrest rates, but these were not statistically significant. Overall, findings show that court-mandated batterer programs can be effective in reducing general recidivism but is inconclusive with regards to reducing battering violence. Looking more broadly across the literature, there is much to be learnt from single-sample studies, systematic reviews, and exploratory pieces of qualitative research. Project Mirabel, a programme of research exploring the delivery of domestic violence perpetrator initiatives and the experiences of survivors, children, and perpetrators involved in them reported positive changes in a range of measures: survivor perceptions of perpetrator respectful communication; expanded ‘space for action’ empowering victim-survivors through restoring their voice and ability to make choices; enhanced awareness of self and others for men on programmes (including an understanding of the impact that domestic violence has had on their partner and children); shared parenting; and healthier childhoods in which children feel heard and cared about. The authors noted marginal changes in some cases, although more significant positive shifts in relation to safety and freedom from violence and abuse, particularly physical and sexual violence (Kelly and Westmarland, 2015). In their evaluation of the Doncaster Domestic Abuse Programme, Ali et al. (2017) also reported positive changes for some perpetrator clients including improved relationships, feeling calmer and better able to cope with stress, and changes in alcohol consumption. In an American study assessing attrition in ‘batterer intervention treatment programs’ (BIPS), the influence of 25 risk and protective factors on treatment “no shows,” “dropouts,” and “completers” were examined for a large sample of male and female intimate partner violence offenders (n = 1,553) (Richards et al (2019a). Multinomial regression analysis demonstrated that the relationship between risk/protective factors and treatment engagement was different across most factors. ‘No shows’ were less likely to have education qualifications, be employed, or to be on probation, and more likely to report a mental health problem, or have a history of drug crimes than completers (but not dropouts), while dropouts were more likely to have a history of general violence or property crimes than completers (but not no-shows). These distinctions can inform efforts to improve intake procedures and engagement strategies and enhance recognition of “red flags” for early treatment disengagement. 11 Survivors of domestic abuse in McGinn et al’s. (2016) US based systematic review on interventions with perpetrators of intimate partner violence reported perceptions of positive changes in partners or ex- partners who took part in a programme, both in terms of their partners’ belief systems and their own feelings of safety. However, the authors also highlighted cases in which survivors felt that the programme was an additional source of family stress, that the perpetrator became more skilled in manipulation and abuse as a result of their involvement, or that the programme had potential to create a false sense of security. Victim-survivors and Independent Domestic Violence Advisors (IDVAs) who took part in AVA’s (2018) consultation commissioned by the Drive Project to better understand victim-survivor ‘voice’ (although not all respondents necessarily had direct experience with the Drive Project) shared similarly mixed experiences of perpetrator responses to domestic abuse. Some expressed concern that perpetrators were resistant to change, would not fully engage with programmes, or would use their engagement for further manipulation and control. However, some respondents did note positive changes in perpetrator behaviour following involvement in a programme and recognised the need for greater focus in this area – alongside adequately resourced services to protect and support victim-survivors and children, and improved responses from police and wider criminal justice practitioners. Studies have pointed to a range of issues that may be important when addressing domestic abuse through perpetrator programmes including: co-ordinated multi-agency partnership and information sharing arrangements, skilled professional workers, robust mechanisms of accountability, and an overall approach which has victim-survivor safety at the core (Bullock et al., 2010; Ali et al., 2017; McGinn et al., 2016). Furthermore, programmes should recognise and manage the complexities of perpetrator intervention work and the possible negative effects for survivors, addressing poor communication and conflict resolution skills and attitudes that were supportive of violence, together with deep rooted barriers to change which some perpetrators can present including mental health support needs, substance misuse, and histories of abuse and poor family relationships from childhood (Bates et al., 2017; Vigurs et al., undated; McGinn et al., 2016). The process of change can differ for each perpetrator and can occur as a ‘series of sparks’, rather than a single ‘light bulb moment’ transformation (Kelly and Westmarland, 2015: 34). Work with perpetrators of domestic abuse is a contested field, often linked to fears about further pressures on already-stretched resources and concerns about redirecting victim-survivor budgets to perpetrator responses (Phillips et al., 2013; AVA, 2018). However, managing domestic abuse offenders with the aim of reducing either harm or offending behaviours as well as addressing the wider risks and needs will not only benefit the offender, but the victims and wider families in ways that stretch beyond current criminal justice responses. While perpetrator programmes should not be assessed in terms of a ‘miracle cure’, Kelly and Westmarland (2015) encourage some optimism about their place as part of broader responses. Domestic abuse perpetrators are a heterogeneous group who require tailored, rather than ‘one size fits all’, approaches to address their behaviour (Vigurs et al., undated). This points to the need for a more developed evidence base and a better link between research, practice, and policy (Bates et al., 2017). The Drive Model Drive addresses the behaviour of high harm domestic abuse perpetrators to target the root cause of the abuse and improve outcomes for adult and child victims. The key outcomes of the Drive response are to: • Reduce the number of serial perpetrators of domestic abuse • Reduce the number of repeat and new victims • Reduce the harm caused to adult and child victims • Intervene earlier to safeguard families living with high harm domestic abuse 12 The project shifts the focus to addressing the behaviour of the perpetrator and aims to knit together services and complement and enhance existing interventions – including crucial protective work with victim-survivors. Addressing the issues that underlie domestic abuse is at the heart of the Drive Project which combines intensive one-to-one case management (differing from the traditional group work settings of some of the earlier cited studies) and delivery of multi-agency services to challenge and change perpetrator behaviour, and address support needs to remove obstacles to the change process. Disruption activity is also a central feature of the Drive model – ensuring that the behaviour of perpetrators who choose not to engage in the process is addressed. The Drive Partnership is made up of: • Respect (the main UK charity working with domestic abuse perpetrators, male victim-survivors, and young people); • SafeLives (a national charity dedicated to ending domestic abuse); and • Social Finance (a not-for-profit organisation that works with government, social sector, and financial bodies to tackle social problems). The Drive Project in Croydon The Drive Project in Croydon went live in July 2018. Domestic abuse cases are identified via the police or Multi Agency Risk Assessment Conference (MARAC) and referred to a fortnightly Drive Domestic Abuse Perpetrator Panel (DAPP) which is made up of professionals across a range of local services including police, mental health, adult and children’s social care, housing, and National Probation Service (NPS)/Community Rehabilitation Companies (CRCs). Four Drive Case Managers (CMs), supported by a Service Manager, employed by Rise Mutual (an organisation specialising in the design and delivery of programmes to challenge and change the attitudes and behaviour of domestic abuse perpetrators) and IDVAs (who hold both Drive and non-Drive caseloads) also sit on the panel. The panel is jointly chaired by a police officer and the Operations Manager of the Family Justice Centre (FJC) where Drive in Croydon is based and CMs and IDVAs are co-located. A diagram outlining the Drive Model is included in Appendix A and a Drive Case Manager training summary is included in Appendix B. Details of each referred case (e.g., perpetrator name, offending history, victim-survivor(s) details) are presented by the RISE Service Manager, and discussions then turn to DAPP members to ascertain eligibility for the Drive Project. Decisions are made based on professional judgment and risk to the victim-survivor, together with the availability of sufficient information to develop strategies using one-to-one intensive case management to respond to abuse and offending behaviour. DAPP members seek to identify ‘levers’ for engagement with the perpetrator via agencies that they may be in contact with including - but not limited to - those that sit on the panel, and opportunities for disruption via criminal justice measures where appropriate (e.g., arrest, non-molestation orders, Domestic Violence Prevention Orders, restraining orders), and other routes to disrupt (e.g., legal proceedings, housing, social care and information sharing). Cases that are deemed suitable for the Drive Project are referred to a CM and IDVA to begin a programme of intensive case management and behaviour change with perpetrators, and support to victim-survivors. Details of Drive referral guidance are included in Appendix C. Work with perpetrators can include both direct (where it is safe and possible to do so) and indirect (e.g., multi-agency disruption and risk management) interventions, and/or disrupt (putting barriers and obstacles in place to prevent abuse) activities (see table 1 and 2 for further details of disrupt and support interventions). Some interventions may take place singularly, whereas others may be delivered consecutively or concurrently and alongside support, disrupt or behaviour change work (or a combination), depending on the risk and needs of the perpetrator, victim-survivor and any children and young people 13 (CYP). Drive exists to target high-risk, high-harm perpetrators who are harder to engage, as such, it does not aim to work directly with all service users; guidance indicates that sites should aim for direct contact with approximately half of service users, and non-direct work with the remaining half. The Drive Case manager will coordinate any support/disrupt interventions alongside other agencies and deliver any behaviour change work with the perpetrator. For example, where the decision has been made not to evict a perpetrator as a risk management strategy, the Drive case manager may also work with the substance misuse team to address their needs as well as building a relationship with the service user that will form the basis of behaviour change work (to be delivered by the case manager on a 1-1 basis when it is appropriate to do so i.e., when their other needs, housing and substance misuse have stabilised, and they are able to engage effectively). Table 1. Areas of support and disruption interventions Areas of support: Physical Health Socialisation and Community Psychological wellbeing – Mental Health Purposeful Activities Homelessness/Housing Education, Employment and Training Children and Families Benefits, Debt Management and Finances Substance Misuse Behavioural and Attitudinal Change Identity Areas of disruption: Children and Families Availability of Victims Evidence gathering for prosecution Notification Use of breach of orders Housing Legal proceedings Benefits, debt and finances Information sharing with police The Drive Model is underpinned by a strong partnership approach and in Croydon this includes a range of agencies involved in the DAPP, CMs, IDVAs, the FJC, and a Drive Fellow - a senior ranking (DCI level or equivalent) police officer who brings local expertise and knowledge to ensure that Drive adapts effectively to the local context. It is intended that Drive Fellows provide leadership and influence throughout set up and delivery of the project, including multi-agency working practices, information sharing protocols, and securing buy-in from key stakeholders. The Croydon project is also supported by the central Drive team at SafeLives and the Drive Partnership, and a designated practice advisor from Respect. At any time between three- and twelve-months cases will be considered for closure on Drive. Numerous factors are considered when deciding whether and when to close a case including, but not exhaustively: all possible actions have been undertaken; the service user has disengaged; a timepoint has been reached which means closure would be considered. Three months is the earliest point at which a case would be considered for closure and only in cases where the service user has not engaged and all possible actions available to the panel and Drive are considered complete. Cases can be extended beyond 12 months where the risk is still a concern. In such cases, where there are still opportunities for meaningful activities clear goals and actions are set out prior to case closure, and/or, additional time is used to secure an appropriate onward referral to other relevant services for oversight. Decisions to close cases involve agreement by the Drive case manager and IDVA, before they are returned to the panel for multi-agency agreement. Source: SafeLives 14 In year two of the project there were no significant changes to the model itself or its delivery in Croydon. Relevant cases continued to be referred to a fortnightly Drive Domestic Abuse Perpetrator Panel (DAPP) via the police or Multi Agency Risk Assessment Conference (MARAC); the frequency, core membership and referral guidelines of the panel remained unchanged (see Appendix C). Additional money was made available to Croydon in year two and has been used to introduce additional therapeutic support for CYP through both one-to-one and group work. 15 Methodology The final evaluation report builds on interim findings (see Appendix D), which focused on the initial learning generated during year one of the Drive Project in Croydon (1st July 2018 to 8th May 2019). Year two learning is based on: • Performance data capturing the activity of the Croydon Drive project between 9th May 2019 and 31st March 2020 (provided by colleagues from Croydon Drive, SafeLives and the Drive Partnership); and • Fieldwork conducted between November 2019 and July 2020. This includes: • Six semi-structured face-to-face interviews with key stakeholders involved in set up and delivery of Drive both centrally and at local level in Croydon (conducted by E&I researchers), including one Case file interview with a Drive Case Manager (CM). • Four depth telephone interviews with DAPP members (conducted by Opinion Research Services (ORS) on behalf of MOPAC E&I). • Two focus groups - one with Drive CMs and one with IDVAs (conducted by ORS). • A dip-sample of a small number of cases (n=25) to illustrate the ‘footprint’ of perpetrators across a range of agencies (provided by Croydon Drive colleagues). • A search of the Metropolitan Police Service (MPS) gangs matrix for the Drive Croydon cohort (provided by colleagues from the MPS and Croydon Drive). • One Drive service user interview (conducted by E&I researchers). • Two Drive service user case studies (provided by Croydon Drive colleagues). Appendix E sets out all fieldwork conducted to date. The UK Government introduced ‘lockdown’ on Monday 23rd March in response to the global COVID-19 pandemic. Whilst delivery of Drive has continued in Croydon, there has been an impact on staff delivering frontline services, as well as the wider Domestic abuse landscape within which Drive sits. Despite this, the performance data which underpins this evaluation - up until 31st March 2020 - was largely unaffected. Planned activity to conduct additional service user interviews as well as victim-survivor interviews have been affected by COVID-19 and resultant pressures on staff delivering front line services and the effects on the general public. Capturing the views and experiences of victim-survivors and perpetrators receiving support through the project is a key element in understanding how a service works. SafeLives are conducting service user interviews with perpetrators and victim-survivors across all Drive sites. This work is due to be published later in the year and will provide valuable insight and learning for Croydon and other Drive sites. 16 Results: Learning from year two of the Croydon Drive Project Project throughput and activity This section reports on year two Croydon performance data from 9th May 2019 to 31st March 2020 and comprises three sections: the first section presents an overview of Drive’s service users, as well as the victim-survivors and CYP associated with cases; the second section outlines the project activity during the second year of Drive in Croydon; and the third section provides insight into how cases have progressed through the service in year two, as well as case manager-reported outcome data in relation to perpetrators, victim-survivors, and changing levels of risk and harm. Section 1: Drive Service users, victim-survivors and CYP Year two cases • In year two, 99 cases were accepted onto the programme - an increase of 26 cases compared to year one. These cases relate to: 99 perpetrators; 114 survivors (1 in 10 cases were associated with multiple victim-survivors); and 184 children or young people. Service Users19 Victim-survivors • Data estimates that 41% of Croydon’s population is from a Black, Asian and Minority Ethnic (BAME) background and 59% is white20. More than half of Drive’s service users in year 2 where BAME (54%) which points towards disproportionality within the cohort. Just over 40% of victim-survivors associated with cases were BAME (41%), representative of Croydon’s population. Further work to better understand how Drive may impact on some people disproportionately should be considered. 19 ‘Serial’ defined as someone who has perpetrated against 2 or more victims. 20 ONS Annual Population Survey (2018). 17 Children and Young People Service User Needs • Using data collected in relation to all Drive service users since the project inception (n=170)2122, demonstrates the complex nature of this client group. Three out of five Drive service users (61%, n=103) presented with three or more needs. This compares to forty-four percent of service users in the year one cohort only (n=32/73), which may be indicative of an increase in prevalence of multiple needs amongst service users, or as a result of improved multi-agency information sharing, and/or improved data recording practices in year two. • During phase 2 of the pilot, replication sites began recording information relating to Adverse Childhood Experiences (ACEs) in the Drive Cohort23. Although the picture is incomplete due to missing data, in closed contact cases where more reliable data is available (n=30), one in three service users (33%, n=10) in Croydon had at least one ACE recorded. • Areas of need were rated ‘low’, ‘medium’, ‘high’ or ‘none’. ‘Children, families and parenting’24 was identified as the highest level of perpetrator need and the most commonly identified need across all cases (high need = 52%, medium need = 15%, low need = 3%). This was followed by ‘Substance Misuse-Drugs’ (high need = 18%, medium need = 26%, low need = 4%), ‘Mental Health and Psychological Wellbeing’ (high need = 19%, medium need = 19%, low need = 8%), and ‘Housing’ (high need = 12%, medium need = 16%, low need = 14%) (see Figure 1). These findings are in line with those reported in year one indicative of consistency of the needs identified throughout the project lifetime. • The UoB evaluation also showed that the Drive cohort had a high degree of needs; 63% reported having one or more needs and 34% reported having three or more needs25. Housing, unemployment and alcohol misuse were the three most prevalent service user needs, closely followed by having a mental health issue26. The evaluation found that a greater degree of needs was found in the Drive cohort than usually seen in voluntary sector Domestic Violence Perpetrator Programmes (DVPPs), where those with needs involving mental health, alcohol and/or drugs may be excluded (Lilley et al, 21 The two-year total number of cases may differ from individual year totals. It is likely that following allocation to Drive cases were categorised as ‘referral not progressed’. Any case that is classed as unsuitable is removed from SafeLives reports and dashboards. 22 It should be noted that information in relation to needs may not be available for all service users as Case Managers are not in direct contact with all service users. 23 Growing up with adverse childhood experiences (ACEs) such as abuse, neglect, community violence, homelessness or growing up in a household where adults are experiencing mental health issues or harmful alcohol or drug use, can have a long-lasting effect on people's lives. 24 This refers to any needs in relation to the service user’s family and / or children, as identified by the Drive Case Manager, e.g. relationship with children or family members, a need to engage with relevant agencies (e.g. children social care), improved understand of the impact of the abuse on their family, or support to engage in safe contact. 25 Service user needs reported at mid-point (n=487). 26 Differences in prevalence of need between the Croydon site and the pilot are likely to be the result of differences in how needs were categorised and recorded. Whilst the pilot captured these needs in three separate categories: relationship with children, relationship with family members & parenting capacity, in the Croydon site they were grouped together as one area of ‘Children and families‘. 18 2018). Further work to develop understanding of Drive’s cohort in Croydon - and high-harm perpetrators more broadly - is suggested (see ‘Working with perpetrators and victim-survivors). Figure One: Service user need Section 2: Year two project activity This section presents data in relation to Drive project activity between 9th May 2019 to 31st March 2020. The data relates to all cases that were open for all or part of the reporting period, including any cases opened prior to the reporting period, and any cases that were closed during the reporting period27. Open Cases • A total of 96 Drive cases were open in Croydon at the time of reporting28. • In half of these (49%, n=47/96), a Case Manager had been able to make direct contact with the perpetrator to coordinate and deliver interventions. This is similar to figures reported in year one (46%) and in line with figures collated from the Phase 2 replication sites (45%)29. Drive exists to target high-harm, high-risk perpetrators, and therefore does not anticipate that all perpetrators will engage with Case managers directly, utilising disrupt opportunities to affect those not in direct contact with Case managers. Guidance from the Drive partnership suggest that sites should aim for between 50 - 60% of cases with direct contact. 27 Although the data in this section may refer to service users previously reported on in the year 1 interim report (covering July 2018 to May 2019), the data will only cover project activity occurring during the reporting period (project activity figures will not be duplicated across reports). 28 As of 31st March 2020. 29 As of 31st March 2020, 446 open cases across all Phase 2 replication sites. Source: Drive Performance Management Croydon dashboard, 9th May 2019 – 31st March 2020 19 • The status of the other open cases (n=49) continues to vary: undergoing additional background research (which may include disruption and risk management activity) (n=22); at pre-closure stage (n=15); on hold (n=11); or, the service user was deemed not safe to contact at that time (n=1). • Of those where direct contact was made, 60% of perpetrators were fully engaging (n=28), 10 partially engaging (21%) and five not engaging (11%) with the Drive service (engagement status was missing for the remaining four cases). It is difficult to compare this data to year one as some of the year two cases may have been opened in year one and as such formed part of the year one data, however, there appears to have been an improvement in service user engagement levels in year two, with increases in the number of perpetrators fully and partially engaging (by 16 and 8 respectively). Although levels of full engagement from perpetrators is lower for Croydon than that found across all Phase 2 replication sites - 71% compared to 60% - Croydon had a higher proportion of partially engaging service users - 9% compared to 21%. When comparing cases in which there was any level of engagement30, Croydon is comparable to the level found across all replication sites with 81% (n=38) of service users in Croydon engaging, and 80% (n=159) of service users across all replication sites engaging with the service. Interventions • During the reporting period, a total of 694 multi-agency interventions31 were made in relation to cases in the Drive service (including both open and closed cases – many cases involved multiple intervention attempts conducted alongside other agencies). This demonstrates a substantial increase in activity levels in Croydon compared to year one, in which 105 interventions were made (see Table 2 for examples of interventions used in the work of Drive). The evaluation was not able to explore in more detail the nature of the interventions or analyse these further, such as the average number of interventions per client or how they were matched to client need. It would be recommended that future work in the area explores this issue. • Most multi-agency interventions (73%, n=509/694) included a combination of both ‘support’32 and ‘disrupt’33. Interventions focused on ‘disrupt’ or ‘support’ only were considerably less frequent (11%, n=73/694 and 16%, n=112/694 respectively). Again, the proportion of interventions which combine both ‘disrupt’ and ‘support’ in year two constitute a considerable increase compared to year one (just over half, 52%) of cases used the combined approach. This may be due to a number of reasons: changes in local recording practices; improved understanding of the model; improved agency buy-in; and/or, improved Multi-agency working during year two. Further exploration of the increased number of interventions and use of both ‘disrupt’ and ‘support’ should be undertaken to develop Drive’s learning. 30 Both full and partial engagement. 31 This figure does not include behaviour change sessions delivered by Case Managers alone. 32 This is defined as ‘addressing perpetrator need with the aim of removing barriers to the change process’. 33 This is defined as ‘putting barriers and obstacles in place to prevent abuse’. 694 Interventions 11% ‘Support’ 16% ‘Disrupt’ 73% Combination of ‘support’ & ‘disrupt’ 20 • Most interventions were jointly conducted by the Drive Case Manager and a range of agencies; most frequently Children’s Social Services (29%, n=201/694), IDVAs (16%, n=109/694) and Police (14%, n=95/694)34. Given the increase in the number of interventions used by Croydon in year two, unsurprisingly, all agencies have increased in the number of interventions they have jointly delivered. In year one, Children’s Social Services jointly conducted 22% of cases with Drive Case Managers, which may point towards an increased and welcomed prominence in their role in year two, given the number of CYP associated with Drive cases in Croydon. • Year one performance data pointed towards the challenges of moving towards meaningful behavioural change work with perpetrators to address perpetrators’ attitudes and behaviour to domestic abuse - eleven of all open cases at the time of reporting had received behavioural change support. Behaviour change work is one tool made available by the model, which does not anticipate that all perpetrators will engage with Case Managers – indeed, as the data shows - disrupt and risk management activities are the primary interventions in some cases. In year two, 43% (n=41/96) of all open cases had received behavioural change support at the time of reporting, with an average of 5 sessions per case; slightly higher than figures for all Phase 2 replication sites (38%). It is evident that there has been progress in delivering this area of the model with increasing opportunities to engage perpetrators in behavioural change work. Given the longer-term nature of such interventions, as cases continue to progress through the service the number of closed cases which have received behavioural change support will likely increase. Table 2. Examples of support and disrupt interventions Examples of support interventions: Homelessness: Drive supported a perpetrator with unsuitable living situation and substance misuse needs to attend an open day and to become engaged with a service working with unemployed, homeless young people to help them lead positive futures. Benefits, Debt Management and Finances: Drive supported perpetrator to access Citizens Advice Bureau for support with debt issues. Drive assisted with setting up payment plans with water and electricity providers. Mental Health: Drive working with a perpetrator in prison for possessing offensive weapon. He has previous convictions for domestic abuse and is not permitted to have contact with any of his children. Service User disclosed he was the victim of childhood sexual abuse and whilst in prison tried to take his life after being told he would not be able to see his son. Service User disclosed that feeling like he had support from the case manager was keeping him going and that he wants help and wants to change. The Case Manager is now liaising with sexual abuse services to find out if support can be provided in prison and to find out what mental health support is available. Examples of disrupt interventions: Police Action: Timely information shared with Police identifying whereabouts of perpetrator, leading to arrest and bail conditions not to contact victim or enter her area of residence. Legal Proceedings: Case Manager supported the victim of the perpetrator in seeking legal support to facilitate the sale of their marital home which he had been obstructive in selling. Actions are currently ongoing. 34 Interventions were also carried out by Prison (12%, n=80/694), CRC (9%, n=63/694) and a range of ‘other’ agencies (21%, n=146/694). 21 Use of Breach of Orders: The Case Manager identified that the perpetrator and victim were in contact. This was a direct breach of his order. The victim was 11 weeks pregnant. The Case Manager notified Community Rehabilitation Companies (CRC) and provided evidence and provided a statement to the police. The victim had already previously had one child removed due to her vulnerability. The Case Manager reported this to Children’s Social Care (CSC), who prioritised an assessment for the victim. The Case Manager is continuing to work with CSC, CRC and the police. As a result, the SU received a 2-year community order, Building Better Relationships rehabilitation programme, 100 hours community service and 20 rehabilitation activity requirement (RAR) days. Examples of support and disrupt interventions: Housing: Case Manager supported a service user who has insecure housing to find permanent accommodation which: a) supported them to get secure housing, as well as b) aimed to decrease harassment/abuse by reducing likelihood of the service user insisting on staying at the victim-survivor’s home. Section 3: Case closure and Case Manager-reported outcomes Closed cases • In year two, 68 Drive cases were closed. The average length prior to closure was 244 days. In year one, just four cases were closed, depictive of the longer-term nature of work associated with Drive. • The primary reason recorded for closure varied: the most commonly reported reason was service user disengagement (44%, n=30/68); followed by limited scope of work (18%, n=12/68) and the risk to victims/CYP deemed too high (16%, n=11/68). Of the cases closed, direct contact was made in 30 cases prior to their closure; of which two thirds (67%, n=20/30) of perpetrators were recorded as ‘non-engaging’ at the point of case closure35. • Figures in relation to case closure reason need to be considered within the context of recording practices: Recording of reasons for case closure began during phase 2 replication of the pilot and during this time understanding of the Drive model, what constitutes an intervention and case closure procedures has evolved in Croydon and in other sites. As such, findings and data recording practices relating to case closure should be reviewed to ensure that they appropriately reflect operational activity, and that staff are recording their work consistently and accurately. Robust data relating to case closure will deepen understanding around service user engagement and motivations for taking part in the Drive project. Abuse Profile The Severity of Abuse Grid (SOAG) is an assessment36 of the changing risk levels to victim-survivors made by Case Managers in relation to the type of abuse (Physical, Sexual, Harassment & Stalking (H&S), Jealous, Coercive & Controlling (JCC)) and severity of abuse (classified as either standard, moderate or high, none or unknown). Cases were assessed at intake and exit for all victim-survivors associated to all closed cases 35 Four were fully engaged and two partially engaged. In the remaining four closed cases, contact had ceased. The definition of ‘non-engaging’ is a service user who is not communicating with Drive and is consistently not meeting commitments or communicating when commitments will not be met. 36 The Severity of Abuse Grid (SOAG) allows Case Managers to draw together information from all available sources, including that from other agencies involved, the IDVA, and their conversations with the service user. Case Managers complete assessments at intake and exit, and at regular intervals (at least every 3 months) whilst a case is open. Assessments at intake and exit are used to monitor progression against outcomes. Source: SafeLives 22 (n=83)37 (see Appendix F, Figures 2-5). At the point of case closure, Case Managers are asked to consider the level of abuse experienced in the previous three months. It is important to note this data relates to a relatively small sample of cases (n=68), has no control condition, and although multiple agencies feed into this assessment, it is based on Case Managers’ own assessments of a case. • A reduction in Physical, Sexual and H&S, JCC abuse was reported by Case Managers across all closed cases for this in which direct and non-direct contact was made with perpetrators (See Appendix F, Figures 2-5). • Although an increase the number of cases classified as ‘standard’ at exit was seen for H&S abuse, it is not necessarily indicative of an increase in overall levels of abuse in this area, for example, service users who were deemed moderate/high risk at intake may leave the programme as ‘standard’. Further work to explore such patterns with a larger sample would be beneficial. • A greater reduction in Physical, Sexual and JCC abuse was reported by Case Managers in cases where direct contact was made with the perpetrator. Given the nature of working face-to-face with service users and having the opportunity to undertake Behaviour change work, this is an expected outcome of Drive work. • Levels of reduction in Case Manager reported abuse levels in Croydon, are consistent with those seen across other Drive replication testing sites and are also following the same downward trend found in the original pilot sites. Summary of performance data Over the duration of the project (July 2018 and March 2020), Drive has worked with a total of 170 perpetrators, 203 associated victim-survivors and 309 associated CYP, this means Croydon delivered to 83%38 of the target set for this time period and is in line with the other Phase 2 replication sites39. This is likely due to a slower initial delivery pace during implementation, commonly expected during the set-up phase of projects. In addition, 799 multi-agency interventions were recorded in relation to cases in the Drive service, including ‘support’ (n=135), ‘disrupt’ (n=100) and a combination of the two approaches (n=564). A substantial increase in the number of interventions recorded may be due to changes in local recording practices; improved understanding of the model; improved agency buy-in; and/or, improved Multi-agency working during year two. Further exploration should be undertaken to develop Drive’s learning. Performance data provides valuable insight into how the project has been implemented, delivered and embedded within Croydon. In year two, project activity in Croydon has increased. A higher number of new cases were allocated via the DAPP, adding to the existing year one caseload, and there was a substantial increase in the number of interventions made, particularly those combining ‘support’ and ‘disrupt’. This may signal increased confidence and understanding of the model leading to greater creativity, although further work to understand local recording practices should be undertaken. In year two, more cases have progressed through the service with an increase in the number of cases closed; indicative of the longer-term nature of the work of Drive. Although service user disengagement is the most common reason for case closure, work is recommended data to review how data is recorded to ensure consistent and robust capture which accurately reflects the work of Drive. The Drive model does not anticipate that all perpetrators will engage with Case Managers – indeed, disrupt and risk management 37 Please note that more than one victim-survivor may be associated with a case. 38 204 service users. 39 84% nationally across all Phase 2 sites. 23 activities are the primary interventions in some cases reflecting Drive’s aim to target those who are harder to engage. Data points towards an increase in the level of behaviour change support being undertaken in year 2, a promising development and indicative of the longer-term nature of direct work with perpetrators. Furthermore, case manager reported outcomes suggest that Drive in Croydon may reduce physical, sexual, H&S and JCC abuse, the effect of which is greater in cases where direct work with perpetrators is undertaken. Performance data, as well as staff feedback (see below), highlights the longer-term nature of working with perpetrators as part of the Drive project. As more cases progress through the service, understanding of what works, for whom and in what ways will develop, as well as understanding of any longer-term benefits of the project. Whilst assessing the impact of the Drive project in Croydon does not form part of this evaluation, work undertaken centrally indicates that reductions in abuse in the replication testing sites are following the same downward trend as found in the initial pilot sites (evaluated by UoB). Information from such pilot sites further along in their journey will be useful - for both Croydon and any new sites - to understand whether the project is working as anticipated. 24 Understanding the cost of domestic abuse and the contribution of Drive Understanding the fiscal benefits of programmes such as Drive is challenging, given its multi-agency approach to working with complex service users. The evaluation of Drive in Croydon did not set out to undertake an analysis of the associated costs of the programme. As well as understanding the impact of Drive on service users, victims, CYP, and society more broadly, understanding the potential cost savings associated with such programmes, is key for commissioners and policy makers building the case for investment in Drive. A Home Office report published in January 2019 estimated that the cost of domestic abuse for victims in England and Wales for the year ending 31st March 2017, was approximately £66billion; the largest component of this estimated cost was the physical and emotional harm experienced by victims, which accounted for an estimated £47million of the cost40. The report goes on to estimate the cost for a single victim of domestic abuse in this period as approximately £34,01541 and acknowledges that these costs likely under-estimate the true cost. The UoB assessed the costs associated with high-risk, high-harm domestic abuse as part of their evaluation of Drive42. Using Drive data to determine the costs relating to victims (and their children), and perpetrators, the analysis estimated the state funded costs associated with perpetrators identified as high- risk via the MARAC referral pathway, as approximately £63,400 per case43. Although it was not possible to assess overall fiscal benefits of the programme as part of their work, using the financial analysis carried out by the UoB as a foundation, the Drive Partnership have since undertaken an assessment of the potential cost savings as a result of the Drive intervention, with a focus on serial perpetrators of domestic abuse44. In their analysis of MARAC data for one Drive site, they observed the reappearance of serial perpetrators (who had been involved in the Drive programme) at the MARAC between 6- and 12-months following completion of the Drive intervention. This preliminary work points towards potential cost savings of Drive and contributes to the ongoing efforts to understand the economic benefits of the programme. It has so far not been possible to assess the economic benefits of the Drive programme in Croydon. In addition to the difficulties associated with understanding the cost of delivering such programmes, given its relative infancy in Croydon, the number of cases that have progressed through the service is not yet sufficient to replicate the cost savings work that the Drive Partnership have conducted which requires sufficient post-intervention time to have lapsed. Furthermore, challenges around the suitability of using locally collected MARAC data were encountered. It is suggested that in order to support future cost analysis following a longer delivery period, opportunities to identify relevant data sources, collect quality data and improve current data recording practices should be explored in Croydon. 40 Oliver et al., 2019 41 1,946,000 victims in England and Wales in the year ending 31 March 2017. 42 Hester et al. (2019) Evaluation of the Drive Project – A Three-year Pilot to Address High-risk, High-harm Perpetrators of Domestic Abuse. 43 The costs for loss of quality of life and economic output were not included. 44 Those perpetrators who have repeatedly used abuse in multiple relationships. 25 Further embedding Drive in Croydon: Year two General perceptions of Drive in Croydon and working with perpetrators Overall, the Drive model continues to be viewed positively by practitioners and stakeholders in Croydon and was commended for filling a gap in support for high-harm perpetrators, as reflected upon by one practitioner: ‘…before there wasn’t anything for them and now there is. And you do have to provide something, … they are the ones wreaking the most havoc, and causing most harm, so you’ve got to fill that gap’. Although Drive was considered expensive and time consuming - particularly by practitioners and those delivering the programme - most participants felt that it is worth continuing with if it can help stop the cycle of abuse and reduce risk. Although participants reflected that despite initial scepticism, most agencies and practitioners were now supportive of the Drive model and working with offenders, it was evident that some - particularly those working to support victims - remained sceptical. Although they acknowledged the need to work more directly with entrenched, persistent offenders, IDVAs in particular felt that Drive exposed an ‘imbalance’ between the tailored, long-term interventions offered to perpetrators and the more generic, short-term support available to victims. Such scepticism is not necessarily a surprise as previous research has also highlighted this issue45. However, there is growing awareness that working with Domestic abuse perpetrators is core to delivery46 – given that they are the source of the problem and tackling offending behaviours can avoid future harm, preventing both future victimisation and new victims. Findings from this evaluation suggest that work still needs to be done to ensure all staff are aligned to the benefits of working with offenders – if not addressed such working tensions will continue. Achieving, widening and sustaining buy-in Buy-in to the Drive model and support for working with perpetrators, was considered a key enabler, something that required both initial and ongoing awareness-raising, but also somewhat of a ‘cultural shift’ towards more acceptance of working with perpetrators – which, participants acknowledged, may take time. Whilst there has reportedly been positive progress in the last 12 months in information-provision and awareness-raising of the project – for example, the MPS now try to provide daily briefings to officers and probation have designated two ‘link officers’ - some stakeholders felt that more could be done to raise awareness through adopting a ‘top-down’ approach; ‘it has got to be from a government level, it has got to be all consuming from every level downwards’. Feedback from practitioners not directly working with perpetrators indicated that they would welcome more official and in-depth data around outcomes, but also case studies and stories to illustrate the work of Drive, types of interventions used, and to highlight learning. Such information may also serve to explain and support the rationale for implementing a perpetrator-focused programme, and address some of the remaining scepticism discussed above. As Croydon prepares for an additional year of delivery, some stakeholders observed that it was a good time to ‘re-launch’ the project to: hold refresher training days; to reengage commitment to Drive and attendance at the DAPP; and, raise the profile of Drive in Croydon amongst agencies and services. There 45 Phillips et al., 2013; AVA, 2018 46 https://gov.wales/sites/default/files/publications/2019-11/violence-against-women-domestic-abuse-and- sexual-violence-public-services-good-practice-guide.pdf. 26 is a clear role for evidence; generating learning, promoting success and demonstrating the effectiveness of Drive, which will be key to sustaining commitment from agencies and staff working in Croydon, as well as raising awareness of the project amongst new partners and in new areas. It is vital that this includes the buy-in and commitment from victim support services. Evidencing and sharing Drive’s successes It was acknowledged by participants that further understanding and sharing the success of Drive in Croydon will be key for the next phase of the programme - to attract and maintain support and buy-in from agencies and staff within Croydon, but also to contribute to the evidence base for perpetrator programmes and the future roll-out of projects such as Drive. At the same time, understanding and measuring the success of Drive was seen as challenging, particularly given the infancy of the project and the cycle of domestic abuse which often occurs over several years. Whilst participants recognised the need to understand and measure effectiveness of the project, defining markers of success for Drive requires further work. Practitioners argued that, given the individualised nature of the intervention work measuring change requires a more nuanced approach than simple recording of generic outcome data. Furthermore, although a reduction in domestic violent behaviour is key, for some participants, success needs to be viewed as ‘all-encompassing’, including evidence of perpetrators leading better, healthier lives, and victims and CYP not only feeling safer, but actually being safer. A review of the outcomes and markers of success should be undertaken to ensure the project continues to learn and build evidence to support further sustainment and rollout. Strengthening partnerships and moving towards local ownership Whilst the Drive Central team have continued to provide a key support function to Croydon, during its second year of delivery, there is evidence of movement towards more local ownership of the model, with Croydon reportedly leading and formulating solutions to problems and implementing local processes. An important aspect of Drive in Croydon is that it is embedded within the well-established local domestic abuse landscape rather than sitting ‘alongside’ it. This has reportedly resulted in increased opportunities to work collaboratively and find innovative ways to support and disrupt perpetrators, further aided by agencies having developed a greater understanding of each other’s roles and the levers they can bring – which may, in part, explain the increase in multi-agency interventions delivered in year two. As one participant summarised: 'A year ago we didn't understand the aims and motives of everyone involved, but now we are in a better place and we've built those relationships’. Indeed, this was seen as particularly evident in Croydon’s work with the MPS in year two. Whilst during early implementation staffing issues led to inconsistencies in the role of the Drive Fellow (typically a police role), during year two stakeholders and practitioners were extremely positive about the cooperative partnership which has developed with the police, presenting opportunities to closely align with existing police operations and activity: 'that's definitely a characteristic of Croydon - they have that complete acceptance by the police that they need to work alongside Drive on these cases’. Despite improvements in partnership working, concerns were expressed by stakeholders and participants that some partner agencies do not appear as invested in the project as others. Practitioners reflected that some agencies, particularly mental health and substance misuse, are becoming increasingly disconnected with the project as workload pressures seemingly result in a de-prioritisation of Drive - a concern given the prevalence of substance misuse and mental health associated needs within the Drive cohort (see Section 1). Drive practitioners also spoke of their frustrations when liaising with services and agencies 27 suffering a high turnover of staff - for whom Drive is not their primary role - which has resulted in cases passed between staff members. If agencies are unable to provide the level of support to the project that is required, the deliberative process of collaboratively developing plans and sharing information may suffer. Delivering Drive: Ways of working Capacity and resourcing At the time of the fieldwork, the Drive Case Manager team was fully staffed – an improvement on year one - which has seemingly made it easier to manage caseloads and brought stability to the team. It was clear from participant feedback, that the Drive workload is demanding, requiring considerable staff time to support information sharing, prepare and attend meetings, and gather intelligence on cases to aid decision making. In year two, an increase in the number of cases accepted onto Drive alongside the continuous flow of Drive cases assigned at the fortnightly DAPP, has sustained the heavy caseloads reported in year one and resulted in caseloads higher than the recommended 25 cases. Furthermore, time spent travelling and arranging the logistical aspects of working with perpetrators, reportedly continues to impact on the management of other elements of the Case Manager role in year two - opportunities to learn from the contingencies implemented during COVID-19 should be explored. Despite these issues, the natural fluctuation of their work means that their roles reportedly remain manageable: ‘We have been at capacity, or even over capacity…But I think it has been manageable, there has been a nice structure in place to manage those cases. Because there is such a fluctuation with the work, there is a freedom in that. But if that wasn’t the case (no fluctuation), I think it would be a very different scenario’. Participating IDVAs felt that their involvement in Drive adds an additional burden to their already busy work schedules and that additional resourcing to deliver the programme is required. Staff feedback indicates that Drive is a resource intense project requiring the right staff with time to do their job. This was illustrated in the role of the Drive Fellow – typically a police role in Drive – where the wider responsibilities of the Drive Fellow beyond Drive and domestic abuse, reduce their capacity to work on Drive. For some practitioners - for whom Drive forms only part of their role - the project creates more tasks, adding pressure to an already demanding workload, reportedly resulting in a reduction in attendance at The Domestic Abuse Perpetrator Panel (DAPP). Training Gaps in training highlighted during year one broadly appear to have been addressed in year two, with participants reporting a general increase in communication, events and information sessions. Opportunities to learn from other Drive sites and role specific network events and workshops have enabled sharing of best practice and learning. In addition, training in areas such as: stalking and harassment; gangs; and, general case manager training such as ACEs, has been positively received, although some felt that it has been too generic focusing on theories of offending as opposed to more in-depth training around behavioural change work which they felt would be beneficial to their role. Furthermore, following a lack of clarity around the ‘disrupt’ element reported in year one, information sessions delivered in the last 12 months have reportedly been helpful, although it was evident from staff feedback that more specific training, awareness raising, and documented guidance in relation to the ‘disrupt’ element is still required: ‘There have been some information sessions, and I do think I’ve got a bit more an idea of about some of the kinds of disrupt stuff. But there was this case the other day at the DAPP which was allocated as ‘disrupt’ and I did have to ask what it means. I know something is meant to be 28 done with it, but I’m not sure what’. Drive practitioners and service managers also indicated that in-depth training on internal IT processes would be beneficial. Information- sharing Multi-agency partnership working continues to be at the heart of Drive. Whilst not a novel concept, the multi-agency approach to tackling domestic abuse in Croydon depicts a change from the more common, unidirectional referral mechanisms which may have limited the opportunity for informed discussions and collaborative action plans typical of Drive. Participants viewed this as a key strength of the model: ‘The information sharing; you have the probation officer, the social worker, the case worker, the IDVA are all having a more holistic view of what’s going on in the family home. So, it’s not a one-sided view but a more accurate whole picture of what’s going on. And that allows for a clearer, more objective plan of how to tackle it rather than other agencies working on it in silo. And from that point of view that’s what’s been the most effective part of Drive’. According to participants, information sharing continues to pose challenges in Croydon and some participants – service managers and case managers – expressed concern that this had worsened during the previous 12 months, resulting in some partners being ‘absent from the picture’. Indeed, case managers reported spending a lot of time ‘chasing’ partners for information, and some had resorted to no longer making information requests to partners outside of the DAPP. Participants speculated that the introduction of GDPR may have impacted on services’ willingness to share information, with some staff seeming to become ‘more protective of their clients’. Participating IDVAs expressed discomfort about sharing information provided by the victim about the perpetrator because of the agreement they have with victims about the use of this information, which falls outside of their control once shared with Case Managers. Gaps in staff awareness and understanding of the data sharing agreements were evident, with some practitioners describing themselves as ‘left in the dark’ about key Drive processes such as sharing intelligence and consent. Given ongoing challenges in this area and the potential impact on delivery, thought should be given to how these issues could be addressed; reviewing local governance structures; prioritising the review, implementation and training of information sharing guidance and processes; and, consulting an information sharing expert to support this work should be considered. The DAPP The Domestic Abuse Perpetrator Panel (DAPP) continues to be viewed as a unique and key component of Drive. Participants praised the newly ‘streamlined’ process that enables members to see which pieces of information and factors will be the focus for each case. As in previous observations of the panel, DAPPs move at a fast pace with 17 cases heard over two hours. Partners appeared more confident when discussing their cases compared to observations earlier in the project, perhaps indicative of Drive being more embedded post-implementation. As well as considering new cases, the panel reportedly serves as a useful forum for reviewing cases that are proving difficult to engage – 10 were reviewed at the panel observed - enabling open discussions between partners about available levers and approaches. As one participant reflected: ’I think some of the review cases have been handy when you can get the partnership information, because sometimes you need that meeting to ask ‘what am I doing with this?’ It’s a good way to get more information or ask people about another way in’. The importance of partners around the table was demonstrated during an observation of a panel – the ‘live’ nature of information sharing and decision making at DAPPs highlights the importance of attendance by all key agencies. It is therefore a concern that participants reported a recent ‘drop off’ in attendance. 29 Possible factors identified by participants as affecting attendance levels, included: the time commitment involved; a high staff turnover resulting in a lack of awareness of meetings and their importance; for some agencies their role may be more limited affecting their commitment to attend, for example reportedly fewer cases are reviewed at the DAPP involving mental health or substance misuse services. Some practitioners and stakeholders observed that the DAPP, unlike the MARAC, has not yet been embedded as part of practitioners’ job roles, as one participant surmised: “People just aren’t turning up (to meetings) as much as they used to. I think with MARAC, it’s kind of drummed into them that they need to be there, But I think with Drive we haven’t really done that”. Participants shared several ways in which attendance at the DAPP could be improved, including: • Holding the DAPP monthly (rather than fortnightly); • Altering the location of the DAPP – it is currently held at the Family Justice Centre in Croydon, which may exacerbate attendance – to a more central location; • Exploring options to join the meeting remotely (e.g. via Skype). Due to the COVID-19 pandemic, DAPP meetings in April have taken place online. Croydon should utilise any learning from these meetings to inform future decisions about the DAPP arrangements; • Raising awareness through Drive roadshows and information sharing events to highlight the importance of the DAPP and the value of partner agencies’ input and attendance and encouraging senior manager within each service to emphasise the importance of attendance; • Partner agencies should receive the list of cases which will be discussed at the DAPP with more notice (rather than a day or two beforehand) so that they have enough time to undertake research and to get a better idea of how they may be of interest to their specific service. Working with perpetrators and victim-survivors Eligibility for Drive There is currently no set of specific selection criteria for Drive as such, and whilst there is some guidance more could be done to set out and standardise who is eligible (or not) for Drive (see Appendix C). According to participants, determining an individual’s suitability for Drive is broadly guided by the level of risk and available levers and includes factors such as: safe levers for contact to ensure victim safety (i.e., if there is no existing intervention/professional contact, that would be deemed unsafe); whether police/probation have indicated elevated risk to professionals; and the extent to which Case Managers are safe to conduct Behaviour change work. Initial misunderstandings that Drive should only work with perpetrators who are ready, willing and able to engage have reportedly since been clarified and there is both understanding and appreciation of Drive’s aims to capture the more complex cases, as summarised by one participant: 'we didn’t want to see cases that were complex, not engaging or not safe to contact in the first instance, just flying under the radar, because Drive was actually set up to focus on that group. If you stop doing that, then those cases just continue to get away with it because they are too hard to work with'. There is evidence that those cases brought to the DAPP are the ‘right’ type with the panel nearly always reaching consensus about the suitability of cases for Drive. One participant commented: ‘We normally agree about what to do with the cases… I know Drive is for high risk cases, but if we think they pose a risk to the case workers, we will reject them. The case managers are not police officers at the end of the day, they don’t have powers to deal with it’. Despite the broad consensus, one stakeholder indicated that more work may be needed to fully develop the eligibility criteria for Drive and potentially expand the cohort to ensure the most appropriate perpetrators are targeted. Thought should be given to further 30 developing the eligibility and selection criteria for Drive in Croydon, to provide clarity for both new and existing staff and agencies, and to ensure and maintain project integrity. Complex cases and complex interventions Panel observations highlighted the complexity of cases with complex needs as well as some presenting long, intergenerational histories of abuse and control. Decisions on whether to close, accept or review cases were not always clear cut, with 9 of the 17 cases heard at the panel brought back for review requiring additional information from partners, further clarity, or due to new information about a perpetrator coming to light, e.g., a breach of a restraining order. Work undertaken on cases requires a ‘flexible’ and ‘responsive’ approach as new information and intelligence shared by partners creates opportunities ‘to close down their [the perpetrators’] space for action’. Fieldwork undertaken to understand the ‘typology’ of cases further highlighted the transient nature of cases; changing risk levels, disengagement and re-engagement, and clients moving between custody and the community are seemingly common in Drive’s caseload47. Although it has been possible to depict several case categories (see Appendix G), it is important to emphasise the complex interdependencies present within Drive’s cohort. Drive’s complex cohort is further illustrated through the ‘perpetrator footprint’ work which, using a dip sample of 25 cases, examined those agencies a perpetrator was known to prior to their involvement in the Drive programme. Children’s Social Care (CSC) and the police, were the most common agencies that perpetrators had previous contact with; 14 perpetrators were known to the CSC and 12 were known to the police. Almost half (n=12/25) of service users were already known to two or more services, prior to their involvement with the Drive programme (see Appendix H). Whilst participants raised the possibility of a link between gangs and domestic violence, the number of gang-related Drive cases in Croydon are extremely low. Indeed, a cross-check of the Croydon Drive cohort with the MPS gangs matrix found a match in only one case. Some participants reflected such low prevalence could be in part due to what they considered ‘strict’ criteria adopted by the matrix. Nevertheless, current evidence indicates no noteworthy crossover between gangs and Drive service users. Further work to develop understanding of Drive’s cohort in Croydon – and high-harm perpetrators more broadly – should be considered; a thorough understanding of the histories and needs of perpetrators, as well as factors in relation to motivation, willingness and ability to engage with an intervention will help to further learning around the role of interventions for perpetrators of domestic abuse. Drive case managers reportedly require ‘a certain approach…a certain skillset which is hard to define’ in order to undertake complex work with a challenging offender cohort. Practitioners described how a considerable amount of preparation - often taking months - is focused on building trust and rapport with a perpetrator before any Behaviour change work can be undertaken. Despite reportedly struggling initially with this aspect of the role and the often-lengthy timescales associated with cases (with some kept open for more than 12 months), in year two, practitioners reflected positively about the work they have undertaken with offenders: ‘we were having difficulty with the lack of time spent on behavioural change work at the beginning…It’s been a process, it’s almost like you have to accept that the first six months are going to be like that for most cases. There are very few cases where you can just dive in and do the work’. This is consistent with learning from wider offender management programmes48. A service user described the relationship with their Case Manager, which was key to their engagement with the programme: ‘I don’t really get on with my family, and most of my mates I don’t really go and 47 Based on one in-depth interview with a Drive practitioner. 48 Sapouna et al., 2015 31 speak to them about those sorts of things. I would say, [CM’s] a very approachable person, [they] makes you feel comfortable to…you know like, some people I feel like judge you, whereas I talk to [CM] about things and like I don’t feel that [they] judge me’. Investing time in building rapport and developing trust, appeared to lead to a sense of accountability for this service user, as they described: ‘they’ve [CM] basically made the first step for me and then that’s made me feel that then I need to take that second step now’. Once behaviour change interventions can be implemented, it was clear from practitioner feedback that there is no standardized approach; Drive delivers a tailored programme of work dependent upon the case, i.e., the environment, the circumstances and the individual needs. Indeed, behaviour change work is highly individual, often involving incremental, small changes, and can occur in many ways, as described by one participant: ‘Behaviour change can mean lots of different things. Talking about your childhood can be a behaviour change – because before a lot of these clients couldn’t talk; they’ve built up an attachment with a professional or any other person to start talking about that stuff’. A service user who engaged with the programme described how as well as helping them to explore the role of relationships in their life and ‘understand why I’ve done things in the past and felt the way I’ve felt’, their Case Manager provided support on ‘coping with stress’ to direct them away from alcohol use, online educational courses, and banking; ‘trying to push me into the right direction’. It is vital to the success of the project that Drive utilises its armoury collectively; ‘change, support and disrupt working together, not separately’. Following gaps raised during year one in understanding of the ‘disrupt’ element of the model, information sessions have been welcomed by staff. Improvements have been noted, with more ‘creative combinations’ involving both support and disrupt, seen in Croydon in year two. Some participants appeared to have a good working knowledge of implementing ‘disrupt’ beyond arrest, describing specific examples of working with partners, such as housing, mental health, probation and social services to disrupt offenders by: undertaking safety planning; putting alerts on addresses; monitoring anti-social behaviour patterns and checking for any other criminal activity. However, whilst there is evidence of improvement, it appears this aspect of the model continues to pose challenges for staff. Uncertainty was expressed by some practitioners around the actions and outcomes that these types of cases typically involve, with a common assumption that ‘disrupt’ equals arrest. Whilst police arrest is the ‘traditional’ route to disrupt, it can be difficult given the nature of domestic abuse offending predominantly ‘behind closed doors’. It was evident that greater understanding and clarity in this area would be beneficial. There may be value for Drive to explore the wider learning from a range of Offender Management programmes delivered over recent years that similarly seek to strike this balance between enforcement and support (i.e., Prolific and Other Priority Offenders Programme; Integrated Offender Management; or the pulling levels approaches)49. 49https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/97811/P PO-5-years-on.pdf. https://www.gov.uk/guidance/integrated-offender-management-iom. Braga et al., 2019 32 Supporting victims Although Drive is a perpetrator targeted programme, victims are also a key focus of the work undertaken. Indeed, during observations at the panel, victim-survivor and child safety were at the heart of all decisions, discussed as part of every case. As noted in earlier observations, child protection - both as a lever and a core part of risk assessment – appeared to guide a lot of cases discussed. The role of the IDVA service in supporting victims was considered an essential part of delivering Drive; ‘if you have not got a robust IDVA service, you are going to struggle on Drive’. The joint working between Case Managers and IDVAs should, according to staff feedback, allow for effective communication and ‘triangulation’ of cases, enabling practitioners to consider the viewpoint of both the victim and perpetrator when making decisions. Positive aspects of the work between Case Managers and IDVAs include their co- location, both useful and important for understanding each other’s job roles and the cases they are involved in, as well as a ‘strong’ working relationship built on ‘mutual respect’, open communication, and trust. As one practitioner described: ‘I think that it is also really helpful that the case managers are based here. So, if you’ve had a conversation with your client you can quickly go and say look, this is what she’s saying, and we can share information very quickly’. IDVAs and Case Managers called for more Case Study A: John* John is a serial perpetrator of domestic abuse who has previously served time in custody for violent behaviours against female partners. He is in a relationship and has children with his current partner, as well as children with his previous partner. John has a history of substance misuse and anger management issues and is well- known to Police. Initially John was difficult to find and engage, but eventually, after several months and numerous attempts, he was engaged via his Community Rehabilitation Community worker and agreed to additional support from Drive. During the initial meeting with his case manager, John presented as angry, asking for support with this, and disclosed high levels of cannabis use. Over time, the Drive case manager developed an effective working relationship with him and, as a result, he was able to openly discuss things and reflect on both recent and non- recent issues he was facing. He acknowledged that he was abusive to intimate partners and wanted to work on changing this. Work undertaken by the case manager uncovered significant childhood traumas and adverse experiences, including: parental mental health; parental substance misuse and domestic violence; young carer responsibilities; and, social workers heavily involved throughout. Throughout the direct, behaviour change work with John, he has shown a willingness to engage, listen and reflect on how his past experiences may have impacted his own. John is attempting to use new strategies to manage anger and conflict. He may need additional therapeutic interventions to support his mental health needs. Alongside direct work with John, Drive has facilitated the sharing of relevant risk information with multiple agencies including Probation, Police and the IDVA service, and John’s current partner has been allocated a social worker following a referral made to the Multi Agency Safeguarding Hub (MASH). Although John remains high-risk, a robust multi-agency response remains in place to manage this. 33 opportunities for inter-team engagement, workshops and partnership meetings to further strengthen and develop their partnership. Despite positive elements of joint working, participants also highlighted some of the challenges of supporting victims as part of a perpetrator focused programme. IDVAs reported that victim feedback about the project is very mixed and many of their clients continue to be sceptical of Drive given its apparent support for the perpetrator, sometimes resulting in them disengaging altogether. There was also concern from participating IDVAs that some victims who remain engaged with Drive do so because they are hopeful of a reconciliation with the perpetrator, or that perpetrators may sometimes use the fact they are on the programme to manipulate the victim into rekindling their relationship. Whilst raised as a risk, it was stressed by participants that this was not unique to Drive, but rather an added complication when working with perpetrators. Moreover, it is likely - according to practitioners - that Case Managers will be working with perpetrators long after a victim has finished being supported by an IDVA, which IDVAs reported feeling ‘awkward’ and ‘uncomfortable’ about when that is the only reason they will continue to contact a victim. As one practitioner observed: ‘It’s kind of saying ‘well we’ve worked with you for a few weeks, and I know that you’ve still got a lot going on, but we’re going to close the case. But just to let you know, I am going to also check in with you once a month, just because we are working with the guy’. Further work to understand the experiences and impact on victim-survivors, when perpetrators engage with perpetrator programmes such as Drive, should be undertaken. Case Study B: Sam* Sam is a serial perpetrator of domestic abuse. He has convictions for domestic abuse, as well as violence and other forms of offending and has served time in prison. Sam presented with pro-criminal attitudes and strong and rigid belief systems, which appear to have influenced his aggressive behaviour in relationships (not only intimate). His negative social circle appeared to further support his abusive behaviour. Sam was approached by the Case Manager whilst he was remanded in custody and was open and willing to engage with Drive. Work undertaken by the Case Manager with Sam highlighted dysfunctional relationships and a pattern of abuse in intimate relationships. Sam reported an unstable childhood with numerous foster care placements before he was 18. He engaged in behaviour change work around his identity and relationships and was open to exploring his behaviours with the Case Manager. Further direct work with Sam is needed; he continues to present as hostile towards the victim and further work to increase his ability to manage his risk to the victim is required. Alongside behaviour change work directly with Sam, Disrupt work has also taken place. Information sharing between the Prison Public Protection Unit, police and Case Manager regarding concerns raised in relation to risks that Sam was presenting to an individual in the community, as well as details of a new partner that Sam is intending to stay with on release, have allowed relevant agencies to take action. As a result of information sharing, Drive have been able to liaise with local IDVA services in order to request a referral to MARAC and a Clare’s Law for his new partner, prior to Sam’s release from prison. Discussions continue between prison, police, probation and the IDVA service to identify new risk concerns linked to the victim or any other individuals. 34 Looking ahead – The future of Drive The future of Drive in Croydon Uncertainty was expressed by some practitioners about Drive’s future beyond the current funding (to March 2021); the assumption that traditional approaches to tackling domestic abuse in Croydon would resume was considered a possibility: “We wouldn’t be running a project like that just ourselves because we need the partners to do it, so we would not do that work anymore, basically.’ An uncertain future for Drive in Croydon has a potential impact on its delivery. During an observation at a panel, the future of the project was raised in relation to whether cases should be taken on. Furthermore, one stakeholder suggested that the recent decline in attendance at the DAPP may be in part due to its uncertain future, highlighting the importance of timely future planning and communication with delivery staff. Stakeholders highlighted several elements which should form the focus of the next phase of project delivery in Croydon to support sustainability: • Further embed the model and ways of working in Croydon – stakeholders anticipated delivering a continuation of Drive in the coming year, with little change to delivery itself or the scale of delivery. • Continue to learn and build evidence – the next phase of delivery was seen to offer a chance to develop understanding of how the project works, broaden the outcomes that Croydon currently monitors, and understand its impact on wider agency working. Outcomes should reflect the nuances of the intervention and its effects on perpetrators, as well as adult and child victims. Identifying which agencies hold data to support the development of outcome data will be a key next step. As the project develops it may be possible to further refine the typology of Drive cases to deepen understanding of how levers work, what combinations work and in which cases. ‘Real stories’ in the form of case studies were also considered vital in cementing people’s understanding of the approach and bringing to life the work of the project. Evidence will be key to harnessing support from operational staff - particularly those not working with perpetrators - as well as maintaining support from senior leaders and commissioners. The programme provides a real opportunity to learn from sites further along in their journey which will be key in sustaining support for the longer-term nature of Drive work; thought should be given as to how this can be best facilitated. • Raising awareness and fostering new links – multi-agency partnership working was seen as a cornerstone of Drive. Participants identified opportunities to forge new links including Health, the Croydon Violence Reduction Unit and London Violence Reduction Unit. As well as fostering links with new agencies, raising awareness of the project was seen as ongoing given staff turnover and organisational restructures. • Take stock and re-engage – following 18 months of delivery in Croydon and confirmation of further funding, stakeholders saw this time as an opportunity to re-engage agencies and staff and ‘re-launch’ the project through workforce development, training refreshers, as well as learning and networking events. It was felt that reviewing attendance levels and the logistical aspect of the DAPP should be a focus for Croydon in the immediate future. • Planning for the future of Drive in Croydon - some stakeholders observed that not enough time was built in to planning for the sustainability of Drive in Croydon. The additional 12 months of funding 35 was seen as an opportunity to begin making such plans and to explore alternative sources of funding for the project. Expanding and Replicating Drive When considering how any potential future roll-out of Drive should occur, stakeholders agreed that a phased approach would be preferable, with pan-London expansion considered after the model has been successfully adopted in other boroughs. Some participants suggested that decisions should be guided by the level of need - prioritising those areas with the greatest need - and considering the local landscape and existing provisions in any planning. The Croydon pilot has generated a set of building blocks to inform and support future expansion and replication of the project (see Figure six): • Buy-in: at every level – commissioner, senior leaders, and operational staff – is vital to the success of the project and the acceptance of work focusing on perpetrators. Sustained support for the project will ensure that sufficient time and capacity are allocated and bring stability and consistency to project delivery. Consideration should be given to longer-term funding cycles to provide more certainty for staff delivering the project. Participants highlighted the importance of setting out clear expectations, outlining the aims of the project and fully briefing key partners and agencies. This will ensure that the correct representatives from each partner agency are bought in to the project and in attendance at the DAPP. As one participant observed: ‘If you really want to make the project sort of work, it is helpful to have individuals who may have had some experience. For example, I think that [they] would have been the better person to attend this meeting. Not because I don’t want to attend’. Whilst securing buy-in is key during implementation, sustaining support and commitment is an ongoing part of the project and will ensure that key functions, such as the DAPP, operate optimally. Evidence will form a key part of attracting and maintaining buy-in and may help to counter initial concerns and scepticism about the project. It is vital that this includes the buy-in and commitment from victim support services. • Evidence & Learning: is crucial to the implementation and delivery of the Drive project, both in sites already delivering Drive as well as any potential new sites about to embark upon project set-up. Evidence has a role in attracting initial buy-in and support for the project and its approach, as well as maintaining ongoing commitment to the project, both operationally and strategically. Participants - both strategic partners and practitioners – highlighted the importance of evidence and learning and its role in sustaining interest in Drive, particularly amongst those who may have concerns about the project’s focus and those who do not work directly with perpetrators. Whilst outcome data is important, participants noted the importance of case studies and narratives which bring the work to life and help others to understand the model and its aims; particularly important in lieu of longer-term outcome data given the project’s relative infancy. There is an opportunity for wider learning between Drive sites through networking and shared learning events and other mechanisms, enabling staff and practitioners to learn from their counterparts. The phased roll-out of the programme benefits new sites enabling them to understand and observe the intended outcomes, as well as anticipate and plan for potential challenges. Generating learning about Drive – how it works, why it works and in what ways - is important for the project itself, as well as its 36 broader contribution to the evidence base for perpetrator programmes. Future rollout could offer new opportunities in this space. • A model adapted to the local landscape: although key principles of the Drive model are consistent across sites, the model affords and encourages flexibility to suit its context and meet local need. Understanding existing local processes, communities, structures, business and third-party opportunities, as well as existing networks, are key to ensuring that the model works within the wider system. Undertaking a mapping exercise will help to identify the local strengths and weaknesses and is a key initial step in planning how to embed Drive. Croydon has demonstrated the benefits of a strong police partnership as well as robust IDVA provision; ‘if you have only got three IDVAs in a borough delivering an IDVA service, they are going to struggle to manage Drive’. The Drive model and its staff are one part of the puzzle; connecting with local agencies and partners, and capitalising on existing activity, is key to fully leveraging the benefits of the project. • A multi-agency approach: at its core, Drive relies on strong partnerships and multi-agency approaches to tackling problems. Whilst targeting perpetrators is the lens through which the project is delivered, work to support adult and child victims is vital to its success, as evidenced in this report and by findings from the UoB. Ensuring agencies and partners have a shared understanding of the model and its aims - as well understanding each other’s contribution to the work – leads to creative and effective interventions. Information sharing and consent processes to facilitate effective partnership working; clear processes and protocols in place and agreed by partners prior to delivery will help to mitigate issues during delivery. Regular review of such processes should ensure optimal working, as well as regular training for new and existing staff. • Skilled and dedicated operational staff: a workforce with the expertise to deliver the work. Staff must fully understand Drive’s novel approach and how to use levers to change, support and disrupt offenders. Training was identified as an important aspect of project set-up and participants reflected that more training ahead of launching the project would have been beneficial. Training on the Criminal Justice System and probation, Drive processes and protocols, as well as operationally focused training in the use of disruption, should form part of the training offer to new staff. As well as defining the roles and expectations of agencies supporting the project, it was noted that a clearly defined job description depicting the responsibilities of the Case Manager, would ensure that staff have realistic expectations of the role and serve to increase staff retention. One participant reflected: ‘I think one thing might be with recruitment. In the early days we had a bit of a turn over with staff, people who had gone for the job and realized it wasn’t quite what they were hoping for’. 37 Conclusion The Drive project in Croydon has continued to deliver a project which addresses the behaviour of high- harm domestic abuse perpetrators to improve outcomes for adult and child victims, and perpetrators. Year two has seen a substantial increase in project activity with a higher number of cases progressing through the service, and a greater number of interventions used. Over the duration of the project in Croydon, Drive has worked with a total of 170 perpetrators, 203 associated victim-survivors and 309 associated CYP. There is much to be taken from process learning from the two-year evaluation of the Drive replication pilot in Croydon. Overall, the model continues to be viewed positively by practitioners and stakeholders in Croydon and was praised for filling a gap in support for high-harm perpetrators. However, it was evident that some practitioners - particularly those working to support victims - remain sceptical or working with Domestic abuse perpetrators. IDVAs in particular felt that Drive exposed an ‘imbalance’ between perpetrator and victim support provision - although there was acknowledgment of the need to work more directly with entrenched, persistent offenders. However, there is growing awareness that working with Figure six: Building blocks for embedding Drive 38 perpetrators is core to delivery50 – given that they are the source of the problem - and tackling offending behaviours can avoid future harm, preventing both future victimisation and new victims. Findings from this evaluation suggest that work is needed to ensure all staff are aligned to the benefits of working with offenders – if not addressed such working tensions will continue. Securing and maintaining buy-in then, is a key enabler of the project, and continuing to build evidence and learning will be vital for Drive in this respect, as well as other perpetrator programmes. Following 18 months of delivery in Croydon, a review of the project outcomes and markers of success should be considered to ensure continued learning for Croydon and any potential wider rollout. Multi-agency working continues to be a key strength of the project and continuing improvements in year two were reported by participants– particularly between Drive and the police. Participants reported improved understanding of both the model itself, and the role of each organisation in its execution, with a sense that Drive is becoming more embedded and established within the local landscape. Going forward, thought should be given to further developing the eligibility and selection criteria for Drive in Croydon, to provide clarity for both new and existing staff and agencies, and, to ensure and maintain project integrity. The Drive model offers a number of interventions available to support, change, and disrupt perpetrator behaviour. Behaviour change work is one tool made available by the model, which does not anticipate that all perpetrators will engage or work directly with Case Managers – indeed, disrupt and risk management activities are the primary interventions in some cases. Staff feedback indicates that behaviour change work is complex, individualised, and difficult given the complex, entrenched and high-risk nature of the client group. Year two has seen some progress in the level of behaviour change support undertaken with perpetrators in Croydon, indicative of the longer-term nature of such work. In addition, Case Managers spoke positively about the work they have undertaken with Drive’s cohort and the small changes which signify meaningful accomplishments. Year two has also seen a notable increase in interventions using both ‘disrupt’ and ‘support’ to target perpetrators, which correlates with the increased understanding of the model and creativity reported by participants - though work to understand such marked growth in activity should be undertaken to support learning. It was evident that the ‘disrupt’ tool continues to cause some uncertainty for staff delivering Drive and further awareness raising, training, and documented guidance, would be beneficial. The two-year pilot has presented some challenges. Agency and practitioner level buy-in was highlighted by participants as vital to project success with varying engagement levels reportedly resulting in inconsistent attendance at the bi-weekly DAPP. The work of Drive adds considerably to already busy workloads and stretched services which poses challenges for partnership working. Given the importance of the DAPP – assembling partners to make informed, intelligence-led, ‘live’ decisions about cases - who is at the table (and who is missing) is key to its success. Issues around information sharing between agencies have seemingly persisted in year two, potentially affecting activity levels outside of DAPPs. Given ongoing challenges in this area and the potential impact on delivery, thought should be given to how these issues could be addressed; reviewing local governance structures; prioritising the review, implementation and training of information sharing guidance and processes; and, consulting an information sharing expert to support this work should be considered. These issues highlight the importance of having clear operational processes in place at the project outset. 50 https://gov.wales/sites/default/files/publications/2019-11/violence-against-women-domestic-abuse-and- sexual-violence-public-services-good-practice-guide.pdf. 39 As discussions regarding the future of Drive in London progress, learning from the replication pilot in Croydon - as well as other Drive sites - should be incorporated into comprehensive and early project planning. Since completion of the initial evaluation period for the Croydon site, the outbreak of COVID- 19 and the subsequent lockdown have likely changed the wider landscape within which programmes such as Drive will be commissioned and delivered. Activities are under-way and evidence is beginning to emerge to better understand the implications with regards to demand for domestic abuse services, offence volumes and severity of offending, as well as budgets and funding availability51. These will need to be considered as part of any sustainability considerations and future planning. 51 For example: Ivandic and Kirchmaier (2020) Domestic Abuse in times of quarantine: https://cep.lse.ac.uk/pubs/download/cp578.pdf; Justice Analytical Services (2020) COVID-19: The experience of victims - a rapid international review, available on request; Currently ongoing research by Dr Katrin Hohl (City University) and Dr Kelly Johnson (Durham University): Responding to the Covid-19 domestic abuse crisis: developing a rapid police evidence base. 40 References Ali, P., Ariss, S., Powell, K., Crowder, M. and Cooper, R. (2017) Doncaster Domestic Abuse Perpetrator Programme: Evaluation Report, 2017 Sheffield: The University of Sheffield Akoensi, T. D., Koehler, J. A., Lösel, F. and Humphreys, D. K. (2013) ‘Domestic Violence Perpetrator Programs in Europe, Part II: A Systematic Review of the State of Evidence’ International Journal of Offender Therapy and Comparative Criminology 57: 1206-1225 AVA (Against Violence and Abuse) (2018) “Capturing the Victim Voice”: Report of a Consultation commissioned by the Drive Project London: AVA Bates, E., Graham-Kevan, N., Bolam, L. and Thornton, A. (2017) ‘A Review of Domestic Violence Perpetrator Programs in the United Kingdom’ Partner Abuse 8 (1): 3-46 Blatch, C., O’Sullivan, K., Delaney, J. J., van Doorn, G. and Sweller, T. (2016) ‘Evaluation of an Australian Domestic Abuse Program for Offending Males’ Journal of Aggression, Conflict and Peace Research 8: 4- 20 Bloomfield, S. and Dixon, L. (2015) An Outcome Evaluation of the Integrated Domestic Abuse Programme (IDAP) and Community Domestic Violence Programme (CDVP) London: National Offender Management Service Boots, D. P., Wareham, J., Bartula, A. and Canas, R. (2015) ‘A Comparison of the Batterer Intervention and Prevention Program with Alternative Court Dispositions on 12-Month Recidivism’ Violence Against Women 22: 1134-1157 Braga, A., Weisburd, D. and Turchan, B. (2019) ‘Focused deterrence strategies effects on crime: A systematic review’ Campbell systematic reviews 15 (3) Bullock, K., Sarre, S., Tarling, R. and Wilkinson, M. (2010) The Delivery of Domestic Abuse Programmes: An Implementation Study of the Delivery of Domestic Abuse Programmes in Probation Areas and Her Majesty’s Prison Service Ministry of Justice Research Series 15/10 London: Ministry of Justice Cox, S.M. and Rivolta, P.M. (2019). Evaluative Outcomes of Connecticut’s Batterer Intervention for High Risk Offenders. Journal of Aggression, Maltreatment & Trauma, pp.1-19 Dawson, P. and Stanko, B. (2013) ‘Implementation, Implementation, Implementation: Insights from Offender Management Evaluations’ Policing: A Journal of Policy and Practice 7 (3): 289-298 Dobash, R. P., Dobash, R. E., Cavanagh, K. and Lewis, R. (1999) ‘A Research Evaluation of British Programs for Violent Men’ Journal of Social Policy 28, 205-233 Dunford, F. W. (2000) ‘The San Diego Navy Experiment: An Assessment of Interventions for Men who Assault their Wives’ Journal of Consulting and Clinical Psychology 68: 468-476 Feder, L. and Dugan, L. (2002) ‘A Test of the Efficacy of Court-Mandated Counselling for Domestic Violence Offenders: The Broward Experiment’ Justice Quarterly 19: 343-375 41 Gordon, J. A. and Moriarty, L. J. (2003) ‘The Effects of Domestic Violence Batterer Treatment on Domestic Violence Recidivism’ Criminal Justice and Behaviour 30: 118-134 Haggard, U., Freij, I., Danielsson, M., Wenander, D. and Langstrom, N. (2017) ‘Effectiveness of the IDAP Treatment Program for Male Perpetrators of Intimate Partner Violence: A Controlled Study of Criminal Recidivism’ Journal of Interpersonal Violence 32 (7): 1027-1043 Hasisi, B., Shoham, E., Weisburd, D., Haviv, N. and Zelig, A. (2016) ‘The “Care Package”, Prison Domestic Violence Programs and Recidivism: A Quasi-Experimental Study Journal of Experimental Criminology 12 (4): 563-586 Hester, M., Eisenstadt, N., Jones, C., Magnus, L., Morgan, K. and Bates, L. (2019) Evaluation of Year 2 of the Drive Project – A Pilot to Address High Risk Perpetrators of Domestic Abuse Bristol: Centre for Gender and Violence Research, University of Bristol Hester, M., Eisenstadt, N., Ortega-Avila, A., Morgan, K., Walker, S.J. & Bell, J. (2019) Evaluation of the Drive Project – A Three-year Pilot to Address High-risk, High-harm Perpetrators of Domestic Abuse: Centre for Gender & Violence Research, University of Bristol HM Government (January 2019) Transforming the Response to Domestic Abuse: Consultation Response and Draft Bill London: HM Government https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/ 772202/CCS1218158068-Web_Accessible.pdf HMIC (July 2014) Everyone’s Business: Improving the Police Response to Domestic Abuse London: HMIC. Home Office and Ministry of Justice (2009) Prolific and other Priority offender programme, Five years on: Maximising the impact. London: Home Office and Ministry of Justice https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_d ata/file/97811/PPO-5-years-on.pdf. (accessed 06/10/2020) Jenkins, J. A. and Menton, C. (2003) ‘The Relationship Between Incarcerated Batterers’ Cognitive Characteristics and the Effectiveness of Behavioural Treatment’ The Peer-Reviewed Journal of the American Correctional Association 28: 1-27 Kelly, L. and Westmarland, N. (2015) Domestic Violence Perpetrator Programmes: Steps Towards Change. Project Mirabal Final Report London and Durham: London Metropolitan University and Durham University Labriola, M., Rempel, M. and Davis, R. C. (2008) ‘Do Batterer Programs Reduce Recidivism? Results from a Randomised Trial in the Bronx’ Justice Quarterly 25: 252-282 Lilley-Walker, S., Hester, M. and Turner, W. (2018) ‘Evaluation of European Domestic Violence Perpetrator Programmes: Toward a Model for Designing and Reporting Evaluations Related to Perpetrator Treatment Interventions’ International Journal of Offender Therapy and Comparative Criminology 62 (4): 868-884 McGinn, T., Taylor, B., McColgan, M. and Lagdon, S. (2016) ‘Survivor Perspectives on IPV Perpetrator Interventions: A Systematic Narrative Review’ Trauma, Violence and Abuse 17 (3): 239-255 42 Mennicke, A. M., Tripodi, S. J., Veeh, C. A., Wilke, D. J. and Kennedy, S. C. (2015) ‘Assessing Attitude and Reincarceration Outcomes Associated With In-Prison Domestic Violence Treatment Program Completion’ Journal of Offender Rehabilitation 54: 465-485 MOPAC (2019) Drive Project: An Interim Evaluation report from the Croydon site Office for National Statistics (2018) Domestic Abuse in England and Wales: Year Ending March 2018 Newport: Office for National Statistics https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/domesticabuseine nglandandwales/yearendingmarch2018#main-points (accessed 19/02/19) Office for National Statistics (2018) Annual Population Survey: Ethnic Groups by borough (2018) Newport: Office for National Statistics https://data.london.gov.uk/dataset/ethnic-groups-borough (accessed 04/09/20) Oliver, R., Alexander, B., Roe, S. and Wlasny, M. (2019) The economic and social costs of domestic abuse: Home Office Research Report [https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/ 772180/horr107.pdf] Pence, E. and Paymar, M. (1993) Education Groups for Men Who Batter: The Duluth Model New York: Springer Publishing Company Phillips, R., Kelly, L. and Westmarland, L. (2013) Domestic Violence Perpetrator Programmes: An Historical Overview Durham: Durham University, London: London Metropolitan University Price, P., Ragagopalan, V. and Donaghy, P. (2009) Domestic Violence Intervention Programme (DVIP): Report and Evaluation of the East London Domestic Violence Service, January 2007 – September 2008 London: DVIP Rempel, M., Labriola, M. and Davis, R. C. (2008) ‘Does Judicial Monitoring Deter Domestic Violence Recidivism? Results of a Quasi-experimental Comparison in the Bronx’ Violence Against Women 14: 185- 207 Richards, T.N., Jennings, W.G. and Murphy, C. (2019a) Risk and protective factors for batterer intervention treatment program attrition: how completers are distinct from dropouts and no-shows. Journal of interpersonal violence, p.0886260519834096 Sapouna, M., Bisset., C., Conlong, A.M. and Matthews, B. (2015) What Works to Reduce Reoffending: A Summary of the Evidence: Justice Analytical Services Scottish Government report [https://www.gov.scot/binaries/content/documents/govscot/publications/research-and- analysis/2015/05/works-reduce-reoffending-summary-evidence/documents/works-reduce- reoffending-summary-evidence/works-reduce-reoffending-summary- evidence/govscot%3Adocument/00476574.pdf] Stanley, N. (2011) Strength to Change: Report of the Evaluation of a New Initiative for Perpetrators of Domestic Violence Hull: NHS/PCT. Taylor, B. G., Davis, C. R. and Maxwell, C. D. (2001) ‘The Effects of a Group Batterer Treatment Program: A Randomised Experiment in Brooklyn’ Justice Quarterly 18: 171-201 43 Vigurs, C., Schucan-Bird, K., Quy, K. and Gough, D. (undated) The Impact of Domestic Violence Perpetrator Programmes on Victim and Criminal Justice Outcomes: A Systematic Review of Reviews of Research Evidence What Works: Crime Reduction Systematic Review Series No. 5 London: College of Policing Welsh Government (2019) Good Practice Guidance for Non-specialist Welsh Public Services on Working with Adult Perpetrators. Wales: Welsh Government https://gov.wales/sites/default/files/publications/2019-11/violence-against-women-domestic-abuse- and-sexual-violence-public-services-good-practice-guide.pdf (accessed 06/10/2020) 44 Appendices Appendix A: Drive Model Source: SafeLives 45 Appendix B: Drive Case Manager training summary Nationally accredited Responding to High-Harm Perpetrator – Case Manager Training (Level 3 Certificate in Domestic Abuse - Prevention and Early Intervention) Inclusive material in 3 blocks of 4 days • Psychology of DA Perpetrators • Typologies of abuse • Trauma and Trauma informed practice • Advocacy Skills • Motivational Interviewing • Working with denial, minimisation & blame • Risk management and intervention planning ----- • Sexual violence • FGM. HBV and Forced Marriage • Mental health • Impact on Children • Male Victims • Assessing counter allegations • Criminal Law • Assertive engagement with clients with complex needs • Housing • LGBTQIA* • Responding to crises situations • Domestic Violence Disclosure Scheme ---- • Specialist 4-day block on holding 1-2-1 intensive interventions with high-harm, serial and complex need perpetrators In addition, the Case Manager can access CPD on: • Stalking • Cultural awareness/responsivity training • DVPO’s and DVPN’s • Disrupt tactics • Risk Management of Prison leavers • Formulation training (with forensic psychologist) • Adverse Childhood Experiences • Child Protection/Safeguarding • Substance misuse and domestic abuse 46 Appendix C: Drive Referral Guidance Source: Drive presentation to MOPAC Violence Against Women and Girls Board, 4 December 2018 47 Appendix D: Croydon Drive interim evaluation: Learning from year one of the Drive Project in Croydon The interim report reflected on initial learning generated in year one of the Drive Project in Croydon. The report presented the ‘activity’ of the project through performance data and reflected on process learning generated throughout fieldwork conducted. Key Findings: • During the performance period (1st July 2018 to 8th May 2019), 73 cases were accepted and allocated to the Drive Service, slightly below the target for that period (n=100). There were 89 victim-survivors and 125 children and young people associated with these cases. • At the time of reporting, 69 cases were open. Drive Case Managers had made direct contact in just under half (n=32) of open cases, with twelve perpetrators fully engaging in the service (most remaining cases were undergoing additional background research, in the process of contact, or on hold). Four cases had closed, with lack of service user engagement the most common reason. • A total of 105 interventions had been made in cases across the lifetime of the Drive Service to date, including support, disruption, or a combination of the two. These were delivered jointly by the Drive Case Manager and a range of agencies, predominantly the police, Children’s Social Services, and Independent Domestic Violence Advisors (IDVAs). Eleven open cases had received behavioural change support - focused interventions to explicitly address perpetrators’ attitudes and behaviours. • A multi-agency partnership approach is at the ‘heart’ of Drive, and the DAPP was welcomed by practitioners and stakeholders as a forum to assess risk and agree approaches to tackle abuse. • However, challenges were recognised around initial stakeholder engagement, identifying levers to tackle perpetrator behaviour (particularly disrupt action), delivering consistent, ongoing training and communication, and establishing data sharing agreements. Furthermore, some stakeholders struggled to resource Drive, particularly time-consuming background research on cases, compounded by broader organisational changes. • Engaging with perpetrators was a challenging part of the Drive Case Manager role, particularly knowing individuals’ whereabouts and identifying ‘levers’ to encourage participation. Co- location with IDVAs was welcomed, with this collaborative approach viewed as promising in terms of making a positive difference for victim-survivors and perpetrators. 48 Appendix E: Fieldwork conducted to date Method Participants Date Evidence and Insight Team, MOPAC Quarterly performance data returns Provided by Croydon Drive team May 2019 – March 2020 Semi-structure face-to-face stakeholder interviews x 5 Central and local (Croydon) Drive stakeholders involved in delivery of Drive November 2019 – February 2020 Observations at DAPP x 1 DAPP members November 2019 Service user interview x 1 Service user July 2020 Opinion Research Services Depth telephone interviews x 4 DAPP members January 2020 – February 2020 Focus groups x 2 Drive CMs; IDVAs February 2020 49 Appendix F: Drive case manager reported outcomes: the Severity of Abuse Grid (SOAG) Source: Drive Performance Management Croydon dashboard, 9th May 2019 – 31st March 2020 Figure 2: Abuse Profile: Intake vs. Exit (All closed cases) (number of victim-survivors associated with cases, n=83) Figure 3: Abuse Profile: Intake vs. Intake Exit Intake Exit Intake Exit Intake Exit Physical Abuse Sexual Abuse H&S JCC Unknown 20% 7% 54% 11% 34% 12% 25% 11% None 12% 77% 24% 83% 6% 69% 6% 67% Standard 5% 2% 2% 1% 6% 8% 8% 5% Moderate 11% 1% 4% 1% 16% 4% 11% 4% High 52% 12% 16% 4% 39% 7% 49% 13% 0% 20% 40% 60% 80% 100% % of closed cases Type and Severity of Abuse Closed Cases High Moderate Standard None Unknown % reduction Physical abuse Sexual abuse H&S JCC High 77% 77% 81% 73% Moderate 89% 67% 77% 67% Standard 50% 50% -40% 43% 50 Exit (Closed Direct contact cases) (number of victim-survivors associated with cases, n=40) % reduction Physical abuse Sexual abuse H&S JCC High 95% 100% 85% 89% Moderate 100% 100% 100% 100% Standard 100% 100% -100% 33% Intake Exit Intake Exit Intake Exit Intake Exit Physical Abuse Sexual Abuse H&S JCC Unknown 25% 5% 50% 8% 43% 5% 28% 8% None 13% 93% 23% 93% 8% 85% 8% 83% Standard 3% 0% 5% 0% 3% 5% 8% 5% Moderate 8% 0% 3% 0% 15% 0% 10% 0% High 53% 3% 20% 0% 33% 5% 48% 5% 0% 20% 40% 60% 80% 100% % of closed contact cases Type and Severity of Abuse Closed Contact Cases High Moderate Standard None Unknown 51 Figure 4: Abuse Profile: Intake vs. Exit (Closed non-direct contact cases) (number of victim-survivors associated with cases, n=43) % reduction Physical abuse Sexual abuse H&S JCC High 59% 40% 79% 59% Moderate 83% 50% 57% 40% Standard 33% - -25% 50% Intake Exit Intake Exit Intake Exit Intake Exit Physical Abuse Sexual Abuse H&S JCC Unknown 16% 9% 58% 14% 26% 19% 23% 14% None 12% 63% 26% 74% 5% 53% 5% 53% Standard 7% 5% 0% 2% 9% 12% 9% 5% Moderate 14% 2% 5% 2% 16% 7% 12% 7% High 51% 21% 12% 7% 44% 9% 51% 21% 0% 20% 40% 60% 80% 100% % of closed non-contact cases Type and Severity of abuse Closed Non-Contact Cases High Moderate Standard None Unknown 52 Appendix G: Initial typology of Drive cases (based on one Case Manager interview) 53 Appendix H: Perpetrator ‘Footprint’ (based on a dip sample of twenty-five cases) Using a dip sample of 25 Drive service users, the below infographic demonstrates the ‘perpetrator footprint’ – agencies that the service user was previously known to prior to their involvement in the Drive programme. Data indicating whether the service users were classified as ‘engaging’ or ‘not engaging’ was available for mental health and substance misuse services only. Notably, perpetrators were most commonly known to child social care, police and substance misuse services. Almost half (n=12/25) of service users were already known to two or more services, prior to their involvement with the Drive programme. This snapshot illustrates the complex nature of the ‘typical’ Drive service user and indicates that service users may have multiple needs beyond their domestic violence offending. •!! ,a. MAYOR OF LONDON OFFICE FOR POLICING London’s use of GPS Tags for Domestic Abuse Final process and performance summary September 2022 Tim Read, Barry Charleton & Laura Duckworth. London’s use of GPS Tags for Domestic Abuse (DA) • There is significant political interest in the use of electronic monitoring to enhance the management of people on probation. Legislation over the last 20 years has enabled it to be used in what the Probation Inspectorate describe as ‘increasingly inventive and intrusive ways’. There is, however, ‘an overall lack of evidence about the longer-term 1 effectiveness of such interventions’. • The Mayor’s Office for Policing And Crime (MOPAC) has been at the forefront in the use of electronic monitoring, piloting the Alcohol Abstinence Monitoring Requirement (AAMR) tags in 2014. Since February 2019, MOPAC has been piloting the use of the GPS tags for individuals released from prison on licence who have committed knife- related offences2 , which was extended in March 2021 to include offences of domestic abuse (DA) across London. • The aims of the domestic abuse GPS pilot are to: o challenge the thinking and behaviour of offenders subject to GPS through the use of location data in supervision; o use GPS data as an integral part of risk management plans; o use GPS data to improve the management of the risk posed to past, current and future partners and other known adults and children; o act as a deterrent to domestic abuse perpetrators; o increase the effectiveness of enforcement action and managing compliance with licence conditions, particularly exclusion zones; and o develop best practice learning in the use of GPS with this group. • The service model for DA is the same as introduced for knife crime tags. Fitting takes place usually at prisons on the morning of release or else at probation offices or approved premises. • Additional data can be supplied by the monitoring provider (Buddi) to the probation practitioner (PP), such as real time key location notifications and summary reports, with heatmaps and top addresses visited, etc. The police are also provided with a list of current tag wearers to support joint risk management with probation and to enable data requests to be considered in appropriate circumstances. • The pilot also involves a crime mapping element - where the movements of those tagged are matched against the location of reported serious crimes and details of matches are provided to the police for further analysis. Tagged individuals are most likely to be White, males, in their 30’s, who had committed violence • To date (mid-October 2022) 261 DA GPS tags have been fitted. • 98% (n255) of those tagged were male, 2% (n6) female. 1HM Inspectorate of Probation (2022) ‘The use of electronic monitoring as a tool for the Probation Service in reducing reoffending and managing risk’. 2 Interim knife crime tag evaluation report available on MOPAC website. Final impact evaluation due in 2023. 1 a mean age of 36. Where ere Black, 10% were Mixed sentenced to custody, returning to , showed that the DA tag sample is equally White and slightly less likely group. individuals were: OF LONDON OFFICE FOR • Recipients of DA tags were aged between 19 to 70, with ethnicity was known (n=244), 49% were White, 33% w White / Black, 7% Asian and 1% ‘Other’ ethnicity. • Data for April 2022 for a similar cohort of those London and meeting the same risk threshold likely to be male, more likely (6 percentage points) to be to be Black (3 percentage points) than the comparison • The most common main sentencing offence for tagged o Violent offences: 54% (n142); o Breaches of victim protection orders: 24% (n62); and o Stalking and harassment offences: 7% (n19). Around half of the tagged individuals successfully completed their order • Of the 190 individuals who had finished on the tag3 - 51% (n97) of the cohort had successfully completed their order while 49% (n93) had been recalled. • The reasons for recall were: o General non-compliance – reasons unrelated to GPS: 35 (38% of those recalled); o Breach of other conditions identified using GPS: 23 (25% of those recalled); o New incident detected through crime mapping or evidenced through EAR: 13 (14% of those recalled); o Breach of GPS monitored exclusion zone: 10 (11% of those recalled); o Multiple reasons with GPS as secondary reason: 10 (11% of those recalled); and o Primarily for GPS related non-compliance: 1 (1% of those recalled). • The proportion of those completing is noticeably lower than the comparative figure of MOPAC’s knife crime tag (which is typically completed in just under 2/3rds of cases), but similar to that reported in E&I’s 2018 Interim report on GPS tagging for an IOM cohort (56% compliance rate). Probation practitioners believe the tags help them and partners manage the offender • PPs outlined the primary benefits of using GPS DA tags, the most popular were: o to establish if the tag wearer had been in an exclusion zone(s); o the ability to monitor the risk to the victim; o to monitor and share information about breaches of victim protection orders (Restraining Orders, Non-Molestation Orders), and o the ability to challenge the tag wearer in supervision. • PP’s preferred email alerts from Buddi about individual incidents, followed by looking at the weekly reports. Less popular were contacting Buddi via email or phone for specific data (such as heat maps, data snapshots etc); logging onto the Eagle system4 to look at GPS data directly; and/or speaking the MOPAC GPS Project team directly. • PPs contacted other agencies/bodies in relation to the GPS DA tags, most frequently the police, but also MARAC, victim liaison officers, social services, independent domestic violence advisors and (less frequently) MAPPA and housing. • Contact with the agencies tended either to be for either of the following: o Information purposes - informing that an individual was on a tag, providing licence conditions/exclusion zone details, victim safety information, breach 3 71 were still ‘live’. 4 The Buddi portal that allows practitioners direct access to GPS data and alerts. 2 that an individual had not This could include updates about a breach/entry into offender management/gangs zones for cases when a wearer had not been informed person. Having a defined the new address whereas be monitored without the specific location. although there were concerns seconded probation officer to OF LONDON OFFICE FOR information, arrested or recalled, or information breached an exclusion zone. o operational purposes - particularly with the police. of location to facilitate arrest, sharing information exclusion zones; provide integrated intelligence5. • Practitioner feedback highlighted the benefits of interest victim or person at risk had been relocated and the tag about the new address to enhance protection for the exclusion zone on the licence would alert the tag wearer to having a non-enforceable interest zone allowed the area to tag wearer being necessarily informed of this There are suggested improvements for the service model • General feedback was the service model worked well, about the model’s resilience (essentially relying on one manage the process). • Most PPs were happy with the system, but areas for improvement were: o maps could be clearer as some staff were unfamiliar with the locations; o more awareness of, and support to use, the Eagle platform to access to live map data; o improvements to the portal, which was seen as ‘confusing and sometimes clunky’. It was highlighted data was not always provided over the course of a day, so difficult to gain a big picture where individuals go. • Communication issues with fitting tags for DA remain consistent with prior issues found within the knife crime tagging evaluation. Some tag wearers reported that they had not been told by probation that they were going to receiving a tag prior to release, and as a result were unhappy, something confirmed by the tag supplier. • There were issues with the accuracy and completeness of lists provided to police detailing who is currently on tag (e.g. missing PNC reference number). The tagging supplier also identified there was scope to reduce the number of short notice referrals from probation to ensure resources can be planned and deployed as required. In 44% of cases time between referral and tag fitting was 5 days or less, in 19% 2 days or less. • When asked, PPs preferred the MOPAC GPS tag to that introduced by the MoJ, largely on the basis that there was better, quicker communication from Buddi. Perceptions of the DA tag are generally positive across staff and offenders 6 Police & probation stakeholders & probation practitioners • Feedback from PPs about the DA tag was overwhelmingly positive – it was felt tags play a role in more robust risk management; public protection; providing reassurance to professionals and victims and was more suited for DA (compared to knife crime). • Senior police and probation stakeholders were also supportive – commenting it was a useful tool in monitoring high risk DA perpetrators. 5 PP feedback was obtained from an online survey that was completed by 17 respondents. 6 PP feedback was obtained from the online survey above. Tag wearer data is taken from an online survey completed when the tag was removed. 30 tag wearers responded, data on perceptions of the tag’s impact was available for 24. This small sample precludes firm findings. 3 expanded to cover other as well as Indeterminate tagged (n=24) have completed the This limits our confidence in the they were generally positive it improved their to others (i.e., housing, with friends and family). served as a reminder they were or because they were less OF LONDON OFFICE FOR • Stakeholders and most PPs would like to see the use of tags offence types (notably stalking); to include Life Sentences; Sentences for Public protection; and to additional prisons. Tag wearers • To date, only a relatively small number of those online survey undertaken when the tag is removed. findings. However, of those to answer – offending behaviour, but that it made no difference employment situation, mental health, or relationship • They attributed the offending result to the fact that it being monitored and of the consequences of reoffending likely to mix with people they might get into trouble with. The ‘Crime mapping’ aspect for DA has significant limitations • The crime mapping for the DA tags is the same process used for knife crime tags with one exception – knife offences have a SPOC at the BCU level, who receives significant match details and those currently on the tag, for DA these details are sent to the predatory offender unit (POU) email inbox. • Both models have the same limitations - high levels of attrition from the crimes uploaded to the number of hits generated; and few or no arrests from the hits provided to the BCUs. The process involves a lot of administration and resource for little apparent reward. • It is currently unclear if the lack of significant matches and consequent prosecutions is due to low/reduced levels of offending by tag wearers. Impact analysis has not been undertaken for this group, but it will be possible when the number of individuals who complete the tag increase. • There is little certainty within the central MPS governance structure for DA about what is happening in various BCUs with the data provided (the hits, and details of those on the DA tag). Stakeholder and practitioner feedback on the future and sustainability of the model • Senior stakeholders (and practitioners) were keen for the DA element to continue, and to be expanded with an associated concern that without an appropriate replacement there would be a gap in service provision. • However, there was also concern about the duplication of electronic monitoring systems, and the confusion that might arise from different systems; providers; platforms; and criteria for inclusion on a GPS tag. Ideally a ‘one system’ approach is required (for London and nationally). • Stakeholders (particularly the police) emphasised this should continue to be a probation-led system to which the police respond. • Service sustainability in terms of design, funding and governance should be addressed. • For this review it has not been possible to ascertain whether self-reported opinions on the effect of the tag on offending behaviour has transferred through to police recorded contact or offending. This aspect will be looked at within E&I’s final Knife Crime tag evaluation (due in 2023). 4 1 A Research Deep Dive into Domestic Abuse in London Paul Dawson, Barry Charleton & Lynne Conroy MOPAC Evidence and Insight June 2022 2 A Research Deep Dive into Domestic Abuse in London Executive summary About the Deep Dive into Domestic Abuse Domestic Abuse is a key issue for the Mayor’s Office for Policing And Crime. The deep dive used an innovative and in- depth approach to examining Domestic Abuse within London. A total of 277 cases were randomly selected from 2019 and systematically coded against 147 variables resulting in a novel dataset from which analytics can then proceed. Key Findings: Victim characteristics Key Findings: Suspect characteristics Key Findings: Offence circumstances Key Findings: Procedural characteristics Key Findings: Case outcome and predictive factors Conclusion The Research deep dive provides valuable insights into the nature of Domestic Abuse within London. As the MPS continues to record an increasing volume of cases, it will become more and more important to improve the response and it is hoped that the results herein can aid such discussions.  Most suspects were male (78%, n216) with a smaller proportion female (22%, n60). The average age was 36 years of age. 51% of victims were White, 26% Black and 19% Asian.  44% (n123) of suspects presented with at least one additional ‘need’, with 16% (n44) presenting more than one need. The most frequent need for suspects was that of mental health issues (27%, n76); substance abuse (25%, n70) and suicidal thoughts (22%, n62).  Three quarters of cases ended with the victim withdrawing support (73%, n201). In half of these cases it was noted the victim reported only to remove the immediate threat.  There was an average of 15 days between initial report and victim withdrawal. Exactly half of victims (50%, n101) withdrew on the same day as reporting.  Overall, 12% of cases resulted in a CPS charge – however, this figure represents a majority of cases which were submitted to them (68%).  A small number of cases progressed to the CPS without victim support (6% of the overall sample but over a third of cases referred to the CPS).  Three quarters of cases (75%, n206) involved an (ex) intimate partner as suspect with one quarter (25%, n71) involving another relation (i.e., parent/sibling/child).  77% of victims were female (n214) and 23% (n63) male. The average age of victims was 36 years of age. 56% of victims were White, 22% Black and 19% Asian.  Most cases (72%, n199) involved some measure of repeat victimisation.  19% of victims presented with at least one additional ‘need’, the most frequent being a mental health issue (12%).  Nearly 70% of cases occurred within the victim’s home.  A third of cases (35%) were noted as having children residing at the location.  Overall, physical (59%, n163) and verbal (46%, n127) abuse were the most frequent behaviours and a third of cases (31%, n85) involved both.  Analytics were only able to explore those variables influencing ‘victim withdrawal’.  The strongest predictor increasing victim withdrawal was the variable ‘police were called to remove the threat’ (these cases were 9x times more likely to withdraw).  The strongest variable in decreasing victim withdrawal was where a ‘dispute over child access’ was noted. These victims were 14x times less likely to withdraw. 3 Background The Mayor’s Police and Crime Plan (PCP) 2022-25 has a specific focus upon reducing and preventing violence and making London a city in which women and girls are safer and feel safer. This continues a rich vein of related work such as the appointment in 2017 of Claire Waxman as London’s first independent Victims’ Commissioner (taking on a key role in making sure victims’ voices are heard and their experiences of services are improved); the publication of the Mayor’s Violence Against Women and Girls Strategy in 2018 and a refreshed VAWG strategy published in 2022. This direction has been supported by a wide array of research analytics conducted by MOPAC – be it a relevant data dashboard to monitor the numbers1; two indepth research reviews exploring Rape2; the evaluation of the Drive programme3; a two year evaluation of the London Survivors Gateway4, in-depth monitoring of victim satisfaction through surveying (the MOPAC User Satisfaction Survey and a new Victims Voice survey); as well as an upcoming evaluation of the MOPAC emergency DA accommodation provision. Specifically focussing upon Domestic Abuse (DA), MOPAC analytics has previously sought to explore the underlying prevalence, something that proved to be a challenge. Indeed, a wide range of research shows the crime is under-reported by victims and under-recorded by the police. In 2014 Her Majesty’s Inspectorate of Constabulary (HMIC)5 highlighted that one in five of all reported crimes (over 800,000) go unrecorded each year, particularly in cases of violence and sexual offending6. In examining the topic, the MOPAC report, Beneath the numbers7 set out a ‘data conundrum’ – that is, as measured by Police Recorded data, DA has seen a substantial increase since 2014 (the MPS increased by 21% in the 12 months ending in March 2015 compared to the same period 2014), whereas prevalence measured by the Crime Survey for England And Wales (CSEW)8 has shown little change (comparing 14/15 or 17/18). Overall, the weight of evidence in Beneath the Numbers pointed towards the increases observed by police being primarily attributed to better compliance with National Crime Recording Standards rather than improved victim reporting. Using 2017 Mid-Year population estimates with self-reported information from a London- level CSEW sample, taken across multiple years (March 2016 – March 2018) the Beneath the numbers report estimated approximately 246,700 adults aged 16 to 59 years who live in London would have experienced any form of domestic abuse in the previous year. This would equate to a prevalence rate of approximately 4 in 100 adults, with women more likely to experience domestic abuse than men (5.9% compared with 2.9%) resulting in an estimated ~164,000 women and ~82,000 men. The landscape of crime and victimisation saw clear changes subsequent to the COVID-19 pandemic and subsequent lockdowns. Within the wider documented crime changes - which saw stark increases within ASB, drug offences and Hate Crime as well as notable reductions in overall Total Notifiable Offences, especially acquisitive crimes9 - Domestic Abuse emerged as an area of concern. 4 Chart 1 shows the volume of recorded domestic abuse flagged offences, month by month for the MPS area between April 2014 to March 2022. Interrogating these data shows the steady increase in police recorded DA with historically consistent peaks, typically within the summer months, as well as the peak in August 2020 where unprecedented levels were reached. This peak follows on from the first National lockdown and could be related to the easing of such restrictions and an increased opportunity to seek help/report. Ivandic, Kirchmair & Lindon (202010) conducted analysis related to DA and the impact of Lockdown on a range of factors including calls for service. They reported that subsequent to lockdown, domestic calls in London increased by 400 calls on average per week compared to the year previous; noting an increase in intimate cases, a reduction in ex-partner cases, and an increase in third party reporting. Outside of the police statistics, there has also been a range of supporting research that has explored the impact that lockdown conditions can have upon DA – on controlling behaviours11; creating conditions that triggered domestic abuse12; magnifying existing abusive behaviour13; as well as affecting the availability of and access to support services14. Anderberg, Rainer and Siuda (202115) explored the magnitude of DA during the COVID-19 pandemic noting the limitation of solely using police data, and instead developed a technique based upon internet search activity which estimated a 40% London increase at peak within DA during lockdown. Wider data from support agencies also suggest an increase subsequent to lockdown – the National Domestic Abuse Helpline (NDAH) reported a sharp increase in calls (+65%) in April and June 2020 (compared with the first three months of the year)16, and the ManKind Initiative helpline also reporting a consistently higher contact rate each month in 2020 than the previous year, with a particularly large increase (+32%) in June 202017. Additionally, data Chart 1: monthly domestic abuse flagged offences for the MPS area between April 2014 and March 2022 (source MPS – Hate Crime or Special Crime Dashboard) 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jul-15 Oct-15 Jan-16 Apr-16 Jul-16 Oct-16 Jan-17 Apr-17 Jul-17 Oct-17 Jan-18 Apr-18 Jul-18 Oct-18 Jan-19 Apr-19 Jul-19 Oct-19 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21 Jan-22 Domestic Abuse flagged offences recorded by the MPS April 2014 and March 2022 (source - MPS Hate Crime or Special Crime Dashboard) Domestic Abuse Domestic Abuse (National Lockdowns) Average Upper Control (+2 SDEV) Lower Control (-2 SDEV) Linear (Domestic Abuse) 5 showed a reduction in the number of referrals to the CPS for a charging decision by 19% in the year to March 202018. In totality, this illustrated a changing landscape within DA which prompted the MOPAC Violence Against Women and Girls Board to request new research. This brief report presents findings from the research, exploring key characteristics, demographics, needs and vulnerabilities of a sample of those who report DA to the police and those who are suspected of perpetrating DA in London; as well as the circumstances of the recorded offences and attrition through the criminal justice system. 6 Methodology The research is inspired by the MOPAC Rape Review (2019)19 adopting an intensive deep dive coding approach to learning. The technique requires the development of a systematic coding framework, which is then completed by accessing individual case files to capture relevant information. The research randomly selected 277 cases from recorded crimes flagged as DA by the Metropolitan Police Service (MPS) in 2019 from their crime reporting system (CRIS). This timeframe allowed for enough time to have passed for the majority of cases to have reached a procedural completion. The coding framework defined 146 variables divided into five categories: • Victim/survivor characteristics; • Suspect characteristics; • Offence circumstances; • Procedural characteristics (the police response and investigation) and • Case Outcomes. This approach enabled the research to generate rich data that is not captured or published elsewhere. The final aspect involved a series of logistic regressions to explore whether specific variables could predict outcomes. All analyses were conducted in SPSS version 24. This method is not without limitation. The research draws from police systems, which of course were not designed as a research tool. These systems can involve inconsistent data capture by officers. For example, we are unable to determine if the absence of a variable was due to a true omission or was not recorded accurately. The approach was also limited to only those who came to police attention (i.e., we know DA is under-reported) so caution is advised in any wider generalisations. Finally, it was not possible to obtain wider Crown Prosecution Service (CPS) data on the cases to allow consideration of the full journey. Nevertheless, the data recorded provides an innovative contextual deep dive, documenting information not available elsewhere to generate new insights. 7 Results As outlined, the research explored a random selection of 277 cases from the calendar year of 2019. Results begin with an overview of the recorded DA victim/survivors, before presenting the suspects, aspects of the offence, procedural factors and ending with an exploration of the predictive factors of victim withdrawal. Who reports Domestic abuse in London? The majority (77%) of victims/survivors in the sample were female (n214) with 23% (n63) being male. The average age was 36 with a range between 2 – 78 years (there were 4 victims below the age of 16). Over half (56%, n139) of victims were white, 22% (n55) Black and 19% (n46) were Asian20. There were a very small number of same sex partnerships (4%) in the data. Three quarters of cases (73%, n203) were reported by the victim themselves, while a further 11% (n30) and 8% (n22) were reported by another household or family member, or a neighbour respectively. Just 4% (n12) were reported by social support or medical services. Most cases (72%, n199) involved some measure of repeat victimisation21. Repeat victimisation was high for all, but females experienced this slightly higher than males (74%, n158 vs 65%, n41). The coding also sought to capture additional needs identified through the investigation. Results demonstrate that 19% (n52) of victims/survivors presented with at least one additional need (e.g., having mental or physical health needs; expressing suicidal thoughts; being under 18 years old; having substance abuse/dependency; and/or having insecure immigration status), and 4% (n11) presented with more than one need. The most frequent need for victim/survivors was related to mental health (12%, n34). The coding identified that DA support services were offered to the victim in 57% (n159) cases. There were some differences between recorded offers across ethnicities (i.e., 61%, n85 of white victims; 49%, n55 of black victims and 65%, n 44 of Asian victims) but none of these were statistically significant. While support services were offered in 159 cases, they were declined in almost half (48%, n76) of these instances. There were a range of differences coded across males and females – overall, male victim/survivors were significantly more likely to have a recorded history of abusing a partner or family member themselves (24% for male versus 11% for female). Female victims/survivors were significantly22 more likely to be offered support services (61%, n131 vs 44%, n28), but this difference was largely driven by the thematic split across the ‘intimate’ and ‘other relation’ categories – with only 12% of male victims in the ‘other relation’ group offered support compared to 51% of females. A significantly23 greater proportion of female intimate partner cases were also submitted to the CPS than male cases (21% vs 5%). 8 Domestic Abuse Suspects The majority of identified suspects were male (78%, n216) with a smaller proportion female (22%, n60). The average age of a suspect was 36 years old with a range between 15 – 80. Two suspects were aged under 16. White suspects accounted for 51% of cases, 26% were Black and 19% were Asian24. As before, coding was able to document identified needs. Results show that almost half of suspects (44%, n123) were identified as presenting at least one additional need, with 16% (n44) presenting more than one need. The most frequent need for suspects was that of mental health issues (27%, n76); substance abuse (25%, n70) and suicidal thoughts (22%, n62). Significantly more suspects (44%) than victims (19%) were identified as presenting a need. This is thought to be related to how the information was captured – namely the risk assessment tools (i.e., the DASH questionnaire25) used by the police focuses more on the management of suspect risk, rather than victim vulnerability. To illustrate, in the DASH there are more questions concerning suspect needs rather than victim. This is an issue we will return to in the discussion. The intimate versus other relation dichotomy Three quarters of cases (75%, n206) involved an intimate relation. Breaking this down, over half (54%, n111) were ex-intimate partner with just under half being a current partner (46%, n95) as suspect. The remaining quarter involving a non-intimate other relation such as sibling, parent, or child (25%, n71). See table 1. This intimate versus other relation dichotomy is not routinely presented in the statistics relating to DA26 27, although the CSEW does make the distinction (with a consistent level of prevalence – 73% of the 2.3 million adults aged 16-74 years who according to the CSEW experienced DA in the year ending March 2020, by a partner or ex- partner, compared to 35% by another family member28). Table 1: intimate and other relation demographics 9 Compared to the intimate cases, the other relation cases were primarily incidents where the suspect was the (grown-up) child29 of the victim/survivor (51%, n71); 32% (n23) were between siblings; 14% (n10) the suspect was a parent of the victim/survivor; and 3% (n2) another relation e.g., an Uncle. The other relation group also presented a differing thematic profile – relatively speaking a larger proportion of male victim/survivors (i.e., 37%, n26 compared to 18%, n37 for intimate partners), who tended to be older (41 years compared to 35 years for intimate partners)30; and a larger number of female suspects (32%, n23 compared to 18%, n37), who were younger (33 years compared to 37 years for intimate partners).31 Offence circumstances This section presents some of the wider circumstances behind or during the offence itself, such as where the offence occurred, the different types of abuse, whether children were involved in some way and so on. Findings indicate most coded DA cases occurred within the home (69%, n192). In over a third of cases the victim was reported to have children (under 16) (34%, n95) and similarly, in a third of cases, children were noted as residing at the location of the abuse itself (31% n86). Children residing within the residence was found to be statistically higher within the intimate cases (37%, n76) compared to other relations (to 14%, n10), but regardless - this raises clear safeguarding concerns. Cross allegations were made by the suspect in a fifth (21%, n57) of cases. Just over a third of female suspects made a cross allegation (23/60, 38%) where male suspects were less likely to make one (34/216, 16%). On a similar note, coding indicated that the victim had previously been a suspect in a DA report in 14% (n39) of cases. In a small proportion of instances (4%, n10) both of these applied, and one or the other applied in nearly a third (31%, n86) of cases. These proportions are similar for both intimate partner and other relation offences, except for the 10 cases where both cross allegations were made, and the victim had been a previous suspect – all of these were intimate partner cases. In terms of the offence itself, physical (59%, n163) and verbal (46%, n127) abuse were the most frequently coded. Nearly a third of cases (31%, n85) involved both physical and verbal abuse. Half of cases (49%, n135) involved two or more different types (e.g., physical, verbal, controlling behaviour, stalking/harassment, sexual assault, threats to kill). While coercive control could be identified as being part of the reported offence in 12% (n34) of all cases, the victim responded ‘yes’ to the DASH question in 43% (n118) of cases. Again, there are differences seen across intimate and other relation cases, with (ex)- intimate partners significantly more likely to involve controlling behaviour32 (49%, n100 compared to 25%, n1833); both individuals (victim and suspect) recorded by the police as suspects for domestic abuse cases at some point (17%, n34 compared to 7%, n534); and have previous instances of harassment (30%, n61 compared to 13% n935). A more detailed breakdown of types of abuse across female/male, intimate partner/other relation splits is shown in Table 2 below. With one exception (the reported offence involving 10 stalking/harassment against female intimate partner victims and female other relation victims) there were no statistically significant differences either between the sexes within each relation group, or between the same sexes between each relation group. Finally, in this section - of those cases referred to the CPS (n50) more than half had a classification on CRIS for violence (n29). A quarter (n12) for a harassment, with the remaining selection made up of various offences including Burglary, Criminal Damage, Public Order and Theft and Threats to Kill. Over half (n29) involved more than one form of abuse. Children resided at the location of the abuse in 15 cases; and the victim sustained an injury in 22 cases, this being deemed moderate or severe in 3 cases. Procedural characteristics This section presents key findings relating to the procedural aspects of the case – covering aspects such as the overall attrition, police investigation and wider timings. In terms of the police investigation, coding revealed the police spoke to suspects under caution in almost two thirds of cases (60%, n167) – which was either by arrest (55%, n153) and/or arranged interview (8%, n21) or both. Body worn video was documented within 55% (n153) of all cases; and an evidential statement was arranged in 59% (n163) of cases and obtained within 35% (n97). In terms of overall case attrition through the criminal justice process, less than 1% of cases were ‘no-crimed’ by the police; 12% resulted in Police ‘No further Action’; 73% resulted in the victim withdrawing support and 18% were referred to the CPS. See Figure 1 for the attrition flowchart. Table 2: Types of abuse in reported offence by sex and relation 11 As outlined, by far the largest grouping regarding attrition was the victim withdrawing support. Examining those victim/survivors in more detail, almost half (47%, n95; or 34% of total cases) were recorded by the police as ‘the victim reported to have the immediate threat removed’. Whether this is accurate recording by officers and reflects a victim/survivors wishes or is somehow related to how the officers are presenting the criminal justice journey is something we will return to later. A significantly36 greater proportion of victims of Black ethnicity (53%, n29) were logged as ‘report to have the immediate threat removed’, than that of either White or Asian victims (32%, n45 and 24%, n13 respectively). The proportion of those reporting to remove the threat that were (ex)partners (73%, n69), female (75%, n71) and co-habiting with the suspect (40%, n 38) were similar in each case to the overall figures. There were no significant differences between victims of different ethnic groups in terms of stage of case progression/attrition, a breakdown of which is shown below in table 3. With regards to victim withdrawal, 70% (n150) of female victims withdrew compared to 81% (n51) of male victims; and 70% (n144) of victims in intimate partner cases withdrew compared to 80% (n57) of victims in other relation cases. These differences were not statistically significant. Within our sample, the CPS authorised a charge in 34 cases (12% of all those recorded, and 68% of those submitted to them). This is consistent with the national charging rate for DA cases, as ONS indicates this to be around 73% of the cases submitted. Figure 1: Case progression for DA allegations Allegations made to the police (n277) Crime recorded & investigation begins (99%, n275) Case referred to the Crown Prosecution Service (18%, n50, 18% of total cases) Charge authorised by the CPS (68%, n34, 12% of total cases) No Crime is recorded (1%, n2) Victim recorded as Withdrawing Support from the investigation (73%, n201, 72% of total cases) Victim states they only reported the offence to have the threat removed (47%, n95, 34% of total cases) Police No Further Action (NFA) (12%, n34, 12% of total cases) 12 A small number of cases were identified as being referred to the CPS ‘without victim/survivor support’ (18 cases or 6% of the entire sample – but which is 36% of all 50 cases referred to the CPS). This figure is far higher than observed in the Rape Review (a figure of 0.4%). Of the cases submitted in this manner, 7 of the 18 cases were successfully charged. Obviously, these are very small numbers, something that precludes the identification of clear patterns, but when looking in more detail, most of these cases included an element of tangible evidence – for example, reference to body worn video (in 13); an evidential statement (in 9) and referencing other technology evidence (in 5). This suggests such investigative aspects may be key when pursuing a prosecution without the support of the victim/survivor. The coding was also able to document the time cases took to move through the system. DA cases appeared to reach an outcome faster than observed within the previous Rape Review research. To illustrate: - There was an average of 15 days between initial reporting and withdrawal. However, this was skewed by a small number of cases with lengthy timescales (a max of 186 days). Looking in more detail, exactly half of victims (50%, n101) withdrew on the same day the report was made to the police; this figure rises to 67% withdrawing within the first week since report. Only 5% of cases withdrew after 100 days. - There was an average of 40 days between initial reporting and CPS submission, but again this is skewed as almost half of cases (48%, n24) were submitted to the CPS within the first week. - There was an average of 24 days between CPS submission and Charge. The CPS completed 56% (n19) of those they charged on the same day they received the case. - Comparing the above statistics to the MOPAC Rape Review: 38% victims withdrew within 7 days and 13% CPS submissions were within a week of the report. Table 3: Stages of case progression / attrition by ethnicity 13 Case Outcome and predictive factors The final section seeks to replicate previous analytics conducted within the MOPAC Rape review, to establish whether specific variables were able to predict case outcomes. Given most cases were recorded as ‘victim withdrawing their support from the investigation’ (73%, n=201) it was not possible to explore modelling on any wider outcomes (i.e., NFA, CPS charge) due to a low sample size. Regression analysis was used to explore which variables have the strongest association with victim withdrawal. Results can be observed within Table 4. In terms of predicting what makes victim withdrawal more likely, the strongest predictor (9x more likely to withdraw) was where a case was coded as police were called just to ‘remove the threat’. Other predictors of withdrawal were when the offence included a physical assault (4x more likely), noting the severity of abuse was increasing (3x more likely) and if the police reported some reconciliation or return to cohabitation (3x more likely). These variables illustrate the complex and contradictory nature of the situation, but also one in which there appears a continued risk. When predicting aspects that made withdrawal of support for the investigation less likely - a dispute over child access was the strongest predictor (14x times less likely). Other predictors were the police making an arrest (4x less likely) and a history of threats to kill (3x less likely). Table 4 – Factors significantly associated with victim withdrawal 14 Discussion Domestic Abuse is a key issue of interest for the Mayor’s Office for Policing And Crime. Attempting to understand the prevalence of DA has proved to be a challenge. Research indicates the crime is underreported, and while police recorded data has suggested an increase over recent years (likely influenced by HMICFRS inspections on crime data integrity) wider National Statistics (i.e., CSEW) have not shown comparable increases. On a wider note, this itself neatly demonstrates the need to incorporate multiple data across agencies and charities and wider to generate a holistic and accurate understanding (i.e., public perceptions, crime data, helpline data, internet data). Of course, the landscape changed subsequent to the response to the CV-19 pandemic – which fostered a range of research noting the potential impact specifically upon DA. Inspired by previous Rape Review research, MOPAC Evidence and Insight were tasked with conducting similar analytics in relation to DA, focussing upon the demographics of victim and suspect, as well as needs, procedural aspects and case outcomes. This deep dive coding used an innovative and in-depth approach to examining DA within London. A total of 277 cases were randomly selected from 2019 and coded against 147 variables resulting in an original dataset enabling a rich vein of analytics. In terms of the findings, much of what we have identified was already established – the description of victim and suspect characteristics or the findings on repeat victimisation are widely documented in the literature. Likewise, the ethnicity breakdown from our sample shows results broadly consistent with the London population in terms of White & Asian groups, with the Black group over-represented (both in terms of victims & suspects). However, this profile is also broadly comparable to overall DA statistics for London where in 2020/21, 73% victim/survivors were female and 27% male; 55% of victim/survivors and 64% of perpetrators were aged between 25 and 44; 46% of victim/survivors were White and 36% were from an ethnic minority group; 88% of perpetrators were male and 11% were female; 53% of perpetrators were White and 46% were from other ethnic groups37. However, there are findings that are more worthy of discussion. We turn to these now. The intimate versus other relation dichotomy Most published research and statistics considering DA tends to present the group as a singular whole. Results of the deep dive highlighted the diversity of the cases – for example a quarter of cases involved ‘non-intimate other relations’, where the suspect was a wider family member, parent, sibling or (grown) child. As outlined, these cases presented a somewhat different victim and suspect profile and were less likely to be offered support services from police officers. In terms of prevalence, similar proportions of ‘non-intimate other’ DA cases have been reported elsewhere (i.e., in the CSEW) but far less is known when considering the deeper context, background or nature of such cases. For example, what different dynamics would present themselves between a parent, sibling or (grown) child presenting as DA suspect. Further work to explore this group in more detail within a larger sample would clearly be beneficial. Ultimately, the question should focus on how best to ensure adequate and 15 specialist provision to this group and whether existing staff working in the area would benefit from additional learning and development to improve the support offer. The effective capture of victim need…? A notable finding obtained from the coding was that fewer needs were identified for the victims as compared to the suspects. Indeed, only 19% of victims had an identified need. This could point to limitations in police data capture or coding, but in most likelihood, this is attributable to the nature of the police response to risk reduction and the investigation itself. To illustrate, within the DASH there are far more suspect focussed questions as opposed to victim focussed (i.e., questions explore whether the suspect - not the victim - has threatened suicide or has substance abuse/dependency). This raises the wider issue of effective risk assessment. In 2009, National Policing leads endorsed the DASH as a model to improve the policing response to DA; and the majority of police forces in England and Wales use this tool. Turner et al (202138) stressed the value of the DASH approach especially in encouraging positive change towards the policing of DA but also noted the variable quality of data capture. Similar findings on the inconsistent implementation of the DASH, officer usage, and the accuracy of the tool to identify the risk of harm have also been reported (see Robinson et al 201639 or Wire & Myhill, 201840). This is clearly a sensitive area, but it is worth exploring how to better capture the potential needs of victims (within a timely and broader procedurally just and non-judgemental approach) not only to better understand risk or the impact(s) of the offence but to enable better signposting to support services. Addressing suspect needs Almost half of suspects (44%) presented with at least one identified need. The most frequently identified were mental health (27%); suicidal thoughts (22%) and substance abuse (25%), with 16% presenting more than one need. This reveals a considerable amount of presumably unmet need. The MOPAC evaluation of the London Croydon Drive scheme41 - a scheme working with high harm DA perpetrators - also showed clear areas of need for this group (i.e., 61% of preparators had multiple needs) and broadly similar results were found in the national Drive evaluation conducted by the University of Bristol42. From a rehabilitative perspective, there is a wealth of research that indicates that addressing offender need is a core aspect of future behaviour change (McGuire, 201343). In this way, there may be learning from the offender management landscape (i.e., such as Integrated Offender Management (IOM), Multi-Agency Public Protection Arrangements (MAPPA) or the historic Prolific and Other Priority Offenders Programme (PPO)) and their approaches which seek to pull together partners; delivering a local response and balancing ‘carrot and stick’ in the management of offenders. Likewise, the ‘pulling levers approach’ (Braga, 200844) and its focus upon an evidence base; key offenders; a varied repertoire of legal actions towards the offender, as well as balancing other social and wider community communication may provide value (Morgan et al, 202045). On this point, the Domestic Abuse Perpetrator Panel (DAAP) within the DRIVE programme would appear to be the nearest template to the above, pulling in a variety of support and disruption interventions within a regular meeting of partners to discuss cases. The MOPAC 16 evaluation of Croydon Drive outlined that the DAAP was viewed as a unique and key component of the overall service. However, it also outlined partner attendance to the DAAP was an ongoing issue during the programme and that whilst most agencies and practitioners were supportive of working with offenders, there remained some scepticism of the benefits of such perpetrator work. In this way, securing and maintaining continued staff buy-in to the rehabilitation and management of DA perpetrators would be beneficial in any future endeavours. Understanding victim withdrawal The majority (73%, n=201) of DA cases were closed by the police under the category of ‘victim withdrawing their support from the investigation’, this is higher than seen in the Rape Review research (65%). The majority of these withdrawals also occurred quickly – with exactly half of victim/survivors (50%, n101) withdrawing on the same day the report was made to the police; rising to 67% within the first week since report. Half of those that withdrew were coded as wanting the immediate threat removed. This highlights multiple issues. Immediately, it illustrates the narrow window of opportunity afforded to police, community responses or other agencies when seeking to engage or promote follow-up. Being able to respond in such a timely fashion would further likely require prompt mobilisation, rapid data sharing as well as multi-agency support amongst other aspects. The limited available timeframe would also likely necessitate a different communication style to best engage. Work to assess the current practicality of such a timely and coordinated response would be worthwhile so to drive improvements. The second aspect of reporting to remove the threat questions as to whether this is an accurate reflection of victim/survivor ‘wants’ and how they are using the police for threat removal; if it is linked to the volatile nature of DA offending itself; or if it is somehow related to how the officers are presenting the case to the criminal justice system. Knowing what ‘success’ looks like for a victim/survivor is key, as a justice conclusion may not be the goal. In this manner, our previous Rape review research highlighted issues such as not wanting to relive the stress, a desire to move on and the act of reporting in itself being enough as reasons for withdrawal. Similar issues have also been reported within DA attrition as well as wider factors around wanting longer term protection or an understanding of the cyclical nature of the offence (see also Barrow-Gint, 201646 or Hester, 200547). Understanding this in more detail would be clearly beneficial, as well as consideration and usage of evidence-led prosecution, something that occurred in 6% of the overall sample but over a third of cases referred to the CPS. The analytics on the predictors of victim withdrawal also illustrates the complexity of the challenge herein. Other than ‘called to remove the threat’, the aspects that made withdrawal more likely were the presence of physical abuse (4x), noting the severity of abuse was increasing (3x), but also an apparent reconciliation or return to cohabitation (3x). This mix of these variables again clearly illustrates the complex and somewhat contradictory nature of the offence, whilst also demonstrating the continued risk in the situation. The variables that predicted a reduced likelihood of victim withdrawal were more straightforward – that of disputes over child access (14x), police making an arrest (4x) and a 17 history of threats to kill (3x). The disputes over child access overall was the strongest overall variable in predicting withdrawal (or the lack of in this case). This could be interpreted in several ways – for example, that where children are involved the victim is more determined to seek an outcome and remain engaged. Similarly, there could be different levels of support or investigation when there are such disputes. The Police making an arrest could well be showing the victim of the ability of the system to protect them thus encouraging engagement, and the final aspect potentially indicates a threshold where victims/survivors may feel they want to proceed with a prosecution. Sitting comfortably underneath the concept of victim withdrawal is the issue of public confidence in the system. There is a wealth of research exploring how confidence in the police can facilitate prosocial and positive engagement. For example, as measured by the MOPAC Public Attitude Survey, those individuals holding positive views of the police are more likely to state they would help the police, provide information to police and even comply with police orders (for example, 71% of those with high levels of confidence report to be ‘very likely’ to provide information to police versus only 58% of those with low levels of confidence48). Within victim withdrawal, overall there was little difference in that approximately three quarters of each ethnicity withdrew. However, Black victims were significantly more likely to be logged as reporting to remove the threat. Research demonstrates that Black Londoners have lower public perceptions in the police – be it confidence, trust, issues of fair treatment or many other aspects49. Furthermore, given the overrepresentation of Black victims and suspects in the sample, this further suggests work to redress such inequality (such as the Mayoral Action Plan50) may also be able to yield wider benefits such as improved willingness to engage with the system if successful. Ultimately, regardless of ethnicity, victims need to have confidence in the system not only to initially reach out but to subsequently engage and maintain throughout. While there is a strong evidence base relating to confidence in the police and the role of fair treatment, effective engagement and police effectiveness (Jackson et al, 201251) less is known about the larger drivers of confidence towards the Criminal Justice System. This should be addressed. New research There are a variety of areas where new research would add value and develop the evidence base further. Many of these have already been raised such as understanding the needs of both victims and suspects. Some of the gender differences are worthy of further exploration – such as understanding cross allegations, the roles of children in attrition, or why males were less likely to receive support or their cases to be submitted to CPS. This itself leads to more nuance in exploring the ‘other relation’ cases and support provision. Likewise, better end-to-end research tracking cases through the entire journey, incorporating CPS information would be advantageous and elucidate a more complete understanding of the process. However, by far the clearest need for future research would be aimed at exploring the issue of victim withdrawal. This could explore facets involving victim decision making, victim confidence of the system, exploring what success looks like for different victims and/or police communication styles and outward support with victims. Across all of the 18 above, there would also be benefits in looking beyond criminal justice research/data, into wider domains such as children’s social work, health, hospitals and family courts (and so on) which would all contribute to the far richer understanding of the issue. Summary The research deep dive into Domestic Abuse provides valuable insights into the nature of the offence within London. As the MPS continues to record an increasing volume of cases, it will become more and more important to improve the response and it is hoped that the results herein can aid such discussions. 19 References and endnotes 1 MOPAC Domestic Abuse and Sexual Violence Dashboard. Domestic and Sexual Violence Dashboard | London City Hall 2 MOPAC Evidence and Insight (2019). The London Rape Review: A review of cases from 2016. https://www.london.gov.uk/sites/default/files/london_rape_review_final_report_31.7.19.pdf MOPAC Evidence and Insight team (2021). The London Rape Review 2021. An examination of cases from 2017 to 2019 with a focus on victim technology. Microsoft Word - London_Rape Review_2021 Final draft_feedbackAddressed.docx 3 MOPAC Evidence and Insight (2020). The Croydon Drive Project: A 2-year Evaluation: Final Report https://www.london.gov.uk/sites/default/files/croydon_drive_final_evaluation_report.pdf. 4 MOPAC Evidence and Insight (2020). The London Survivors Gateway pilot: a 2-year evaluation. https://www.london.gov.uk/sites/default/files/london_survivors_gateway_final_evaluation_report.pdf 5 Now Her Majesty’s Inspectorate of Constabulary and Fire and Rescue Services (HMICFRS). 6 Crime-recording: making the victim count’ © HMIC 2014 ISBN: 978-1-78246-660- 7 Beneath the Numbers: An exploration of the increases of recorded Domestic Abuse and Sexual Offences MOPAC Evidence and Insight https://www.london.gov.uk/sites/default/files/201902_beneath_the_numbers_- _an_exploration_of_the_increases_of_recorded_domestic_abuse_and_sexual_offences_v1.pdf 8 Domestic abuse https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/crimeinenglandandwales/yearending march2021. 9 See MOPAC quarterly Performance Update report for an overview https://www.london.gov.uk/sites/default/files/mopac_q1_2021-22_monitoring_report_1.pdf. 10 Ivandic, R., Kirchmair, T., & Lindon, B. (2020). Changing patterns of domestic abuse during Covid-19 lockdown. Centre for Economic Performance, no.1729. November 20. dp1729.pdf (lse.ac.uk) 11 GLA City Intelligence. (2020) COVID-19: Summary of external research. London, UK: Greater London Authority. 12 Fraser, E. (2020) Impact of COVID-19 Pandemic on Violence against Women and Girls, VAWG Helpdesk Research Report No. 284. London, UK: VAWG Helpdesk. 13 Williamson, E., Lombard, N., & Brooks-Hay, O. (2020). Domestic violence and abuse, coronavirus, and the media narrative. Journal of Gender-Based Violence, 4(2), 289-294. 14 House of Commons Home Affairs Committee (April 2020) Home Office preparedness for Covid-19 (Coronavirus): domestic abuse and risks of harm within the home. 15 Anderberg, D, H Rainer, and F Siuda (2021), “Quantifying domestic violence in times of crisis: An internet search activity-based measure for the COVID-19 pandemic”, Journal of the Royal Statistical Society: Series A, forthcoming. 16https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/domesticabuseinenglandandwaleso verview/november2020 17 The ManKind Initiative operates a confidential helpline available for all men across the UK suffering from domestic abuse. Source: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/domesticabuseduringthecoronavirusC ovid19pandemicenglandandwales/november2020#domestic-abuse-during-the-coronavirus-Covid-19-pandemic-data 18https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/domesticabuseinenglandandwaleso verview/november2020 19 MOPAC Evidence and Insight (2019). The London Rape Review: A review of cases from 2016 https://www.london.gov.uk/sites/default/files/london_rape_review_final_report_31.7.19.pdf 20 Ethnic categorisation was based upon police IC code as recorded on the CRIS reports. White refers to combined figures for IC1 (white, northern European, n118) and IC2 (white, southern European, n21). Asian refers to the combined figures for IC4 (Asian, n44) and IC5 (East Asian, n2). In addition, there were n8 IC6 (Arabic) victims (3%). 20 21 Coding of repeat victimisation was based upon victim account/response to DASH questions as well as five year intel checks – therefore could be at any time and/or with a different suspect. 22 p0.05 23 p0.05 24 As detailed above (ref footnote 19), ethnic categories were based upon police IC code and were grouped. White suspects comprise n104 IC1 and n14 IC2. Asian suspects comprise n44 IC4 and n1 IC5. In addition, there were n8 IC6 (Arabic) suspects. 25 The Domestic Abuse Stalking and Honour-based violence (DASH 2009) questions are a risk identification, assessment and management model used by police services in the UK. A bank of questions is asked when attending what appears to be any domestic incident to identify, assess and manage ongoing risk. 26 MOPAC Domestic Abuse and Sexual Violence Dashboard. Domestic and Sexual Violence Dashboard | London City Hall. 27 Metropolitan Police Crime Dashboards. Crime data dashboard | Metropolitan Police. 28 Crime Survey England and Wales. Domestic abuse in England and Wales overview: November 2020 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/domesticabuseinengland andwalesoverview/november2020. 29 Where the suspect was a child of the victim, be aware that these were mostly adult children (average at 29). The full age range was 15 – 50, but of the 36 cases only 8 were below the age of 20. 30 Almost half (46%, n12) of male other relation victims were abused by their child, while in 39% (n10) cases it was a sibling. 31 Victim and suspect age ranges for other relation cases were 2-72 years and 15-64 ears respectively. The victim withdrew from the police investigation in 80% (n57) other relation cases, while 18% (n13) were submitted to the CPS, 7 cases (10%) being charged. 32 Based on response to relevant DASH question. 33 p0.01. 34 Not statistically significant with Yates’ correction. 35 p0.01. 36 p0.01. 37 Metropolitan Police domestic abuse recorded offences. https://www.london.gov.uk/sites/default/files/london_domestic_abuse_safe_accommodation_needs_assess ment_-_november_2021.ppsx 38 Turner, E., Brown, G., & Medina, J. (2021). Predicting domestic abuse (fairly) and police risk assessment. CrimRxiv. https://doi.org/10.21428/cb6ab371.3f0ab9af. 39 Robinson, A. L., Myhill, A., Wire, J., Roberts, J., & Tilley, N. (2016). Risk-led policing of domestic abuse and the DASH risk model. What Works: Crime Reduction Research. Cardiff & London: Cardiff University, College of Policing and UCL Department of Security and Crime Science. 40 Wire, J., & Myhill, A. (2018). Piloting a new approach to domestic abuse frontline risk assessment. Evaluation Report for the College of Policing [online] https://whatworks. college. police. uk/Research/Documents/DA_risk_assessment_pilot. pdf [accessed 4th March 2019]. 41 Geoghegan-Fittal, S., Keeble., C & Wunsch, D. (2020) The Croydon Drive Project: A 2 year evaluation. MOPAC. croydon_drive_final_evaluation_report.pdf (london.gov.uk) 42 Hester, M., Eisenstadt, N., Ortega-Avila, A., Morgan, K., Walker, S.J. & Bell, J. (2019) Evaluation of the Drive Project – A Three-year Pilot to Address High-risk, High-harm Perpetrators of Domestic Abuse: Centre for Gender & Violence Research, University of Bristol Drive-Evaluation-Report-Final.pdf (driveproject.org.uk) 43 McGuire, J. (2013). 'What works' to reduce re-offending: 18 years on. 44 Braga, A. A. (2008). Pulling levers focused deterrence strategies and the prevention of gun homicide. Journal of criminal justice, 36(4), 332-343. 21 45 Morgan, A., Boxall, H., Dowling, C., & Brown, R. (2020). Policing repeat domestic violence: Would focused deterrence work in Australia?. Trends and Issues in Crime and Criminal Justice, (593), 1-20. Policing repeat domestic violence: Would focused deterrence work in Australia? (aic.gov.au). 46 Barrow-Grint, K. (2016). Attrition rates in domestic abuse: time for a change? an application of temporal sequencing theory. Policing: A Journal of Policy and Practice, 10(3), 250-263. 47 Hester, M. (2006). Making it through the criminal justice system: Attrition and domestic violence. Social policy and society, 5(1), 79-90. 48 MOPAC Public Attitude Survey R12 to Q3 19-20. 49 Action Plan Dashboard | London City Hall 50 Mayor's Action Plan focuses on disproportionality of police powers | London City Hall 51 Jackson, J., Bradford, B., Stanko, B., & Hohl, K. (2012). Just authority?: Trust in the police in England and Wales. Willan. Developing an evidence based police degree-holder entry programme Final Report Home Office Police Innovation Fund 2016-18 Edited by Mike Hough & Betsy Stanko Contributing authors Winifred Agnew-Pauley Jyoti Belur Jennifer Brown David Gamblin Gillian Hunter Almuth McDowall Brendan McGinley Tiggey May Lisa Tompson DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME MOPAC and its academic partners would like to thank the Home Office for providing a grant for this work from its Police Innovation Fund. We would like to thank the staff of the Police Now programme for offering themselves as a test bed for research. We are also grateful to the second and third Police Now cohorts for participating in the research. Research can make unwelcome demands on those who agree to submit themselves to the research process, and we appreciate their forbearance. We would also like to thank all those staff from police forces, social work departments and teaching institutions who made valuable contributions of their knowledge and time to this project. For their thorough and illuminating peer reviewing, we would like to thank Chief Constable Andy Rhodes and Professor Ken Pease. Thanks are also due to the College of Policing, which became an increasingly important partner and stakeholder in this project as it evolved. The College has provided us with a great deal of encouragement and practical help throughout the life of the project. We would like to thank Betsy Stanko for the vision to initiate this project, and for providing strategic advice throughout. Finally, we should thank the MOPAC project director, Robin Merrett, and the project coordinator, Lizzie Peters, who both did exceptional jobs at keeping a complex and ever-evolving project on schedule. Lizzie Peters was also involved in aspects of the research design and in fieldwork interviews. Winifred Agnew-Pauley Jyoti Belur Jennifer Brown David Gamblin Mike Hough Gillian Hunter Almuth McDowall Brendan McGinley Tiggey May Betsy Stanko Lisa Tompson ACKNOWLEDGEMENTS 3 CONTENTS Acknowledgements 2 Contents 3 Foreword by the Mayor’s Office for Policing and Crime 4 Foreword by the College of Policing 5 Summary 6 Introduction 12 Chapter 2: The evidence on conversion courses 26 Chapter 3: The systematic review on effective police training 34 Chapter 4: Organisational readiness for the Degree Holder Entry Programme 46 Chapter 5: Motivations and appreciation of evidence based policing amongst Police Now participants 58 Chapter 6: View from the shop-floor: perspectives on the Police Now programme 76 Chapter 7: Insight from the research 92 References 100 Appendix A: Stakeholder interviews about graduate conversion courses 104 The safety of Londoners is the key priority for the Mayor. It is one of the essential foundations of London’s success as a global city. Our Police and Crime Plan provides a strong pathway for progress, and sets out MOPAC’s role in ensuring the Plan is delivered. However, I recognise that if we are to move forward effectively, we need to place the officers, staff and volunteers within the Met at the core of our plans. People are central to all we do; we need to recruit the right people, and properly support them to learn and develop. As part of the Plan, MOPAC and its partners aim to transform police learning in London, ensuring that we deliver a service to the public that is fit for current and future challenges. Having police officers with the right skills, knowledge and behaviours is absolutely fundamental to effective policing. MOPAC created this project to develop the evidence base for future training, learning and development of police officers. One of the key pathways is to transform the police service into a learning organisation that fully embraces evidence based practice, and this is consistent with the National Police Chiefs Council’s Vision 2025. As part of this the College of Policing has set out the Policing Educational Qualifications Framework (PEQF). The project findings presented here are one element in this process of change. MOPAC led this collaborative project, working with the Metropolitan Police Service and other forces nationally, the College of Policing, two universities and Police Now. The ambition was to provide support in developing the Degree Holder Entry Programme, one of the three PEQF initial entry routes to policing - whereby on joining the police service graduates embark on a programme to convert their degree into a policing qualification. These research findings have already informed the development of this programme, which police forces need to implement by 2020. The project was made possible through a grant from the Home Office Police Innovation Fund. The fund is intended to drive innovation in policing through collaboration, and this report is ample proof that the project has been successful in achieving this. SOPHIE LINDEN Deputy Mayor Policing and Crime, Mayor’s Office for Policing and Crime FOREWORDS DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Modern policing is changing. Crime is taking on new forms, generating a new set of challenges to which the police service must respond. More than ever, those working in policing are required to engage in critical thinking, deal with complex policing situations and environments, work with a high degree of autonomy, make decisions independently, communicate and negotiate successfully, engage in professional reflection, and contribute in visible ways to supporting and protecting the public. The role of police constable remains the bedrock of effective policing, but increasingly it requires knowledge and skills to be applied across a range of different situations, as well as behaviours consistent with efficient, appropriate and supportive policing. The College of Policing is working with the police service to introduce three new degree-based entry routes into policing at the rank of police constable. We are determined that programmes of education within policing will foster the highest standards of educational and professional development, enhancing knowledge, skills and competence within the service. The way in which we equip those working in this key role must reflect its wide- ranging responsibilities and support them to succeed. In this regard, policing has much to gain from partnership working with those in the education sector who are directly engaged in policing-related research, in ways that can improve police learning and development and, ultimately, practice. We welcome the publication of the research findings of this MOPAC project as a significant contribution to the research base on the education of new recruits to the police service. Although the report concentrates upon graduate entry into policing, many of the research findings have a broader relevance to policing education. Emphasised throughout the report is the importance of full integration of theoretical and practical elements within policing education. As the professional body setting the national education standards for the police, we are committed to ensuring that the new degree-based entry programmes for the Police Constable (PC) role are fundamentally based upon demonstration of effective, evidence-based professional practice in the workplace. Delivery of a new framework of practice-based education for the police constable will play an important part in the service’s delivery of what the Policing Vision 2025 describes as ‘a more sophisticated response to the challenges we face now and in the future’. The MOPAC report will help the service to frame that response, supporting us to deliver a better policing service to the public. MIKE CUNNINGHAM Chief Executive, College of Policing 5 The College of Policing launched a consultation in 2016 on plans to require all police recruits to hold a policing qualification at degree level. The plans were agreed in 2017 as part of the College’s PEQF with the expectation that the degree requirement would take effect from 2020. The PEQF provided for three entry routes for police constables: • the Police Constable Degree Apprenticeship (PCDA), introduced from April 2018, involving academic study in parallel with operational work • the pre-join degree in professional policing, introduced from 2020, where prospective recruits acquire their degree in policing in advance of recruitment • the Degree Holder Entry Programme (DHEP), introduced from 2020 – the focus of this report. The aim of the project reported here, funded by the Home Office Innovation Programme, was to contribute to an evidence-informed dialogue about the nature, quality and purpose of police education. More specifically the project examines questions that relate to the development of the DHEP: • How do graduate police recruits learn about the standards for policing and the public expectations of a police officer? • How do graduate police recruits apply their critical skills (gleaned from their university study) to their decision- making as a police office? What is the role of evidence here? • How do graduate police officers learn the essential skills/craft of policing? Do they have different learning styles from non-graduate recruits? What is the role of evidence here? • What is the theory of change for requiring graduate entry – that is, what are the key elements of policing that are expected to result from graduate training, and how? The project involved several separate elements, carried out by researchers at Birkbeck, University of London and University College London. The team conducted a review of the UK research literature on analogous conversion courses provided for graduates entering law, teaching, and the social work profession, and a further review of the international literature about effective police training. Team members interviewed staff in police force Learning and Development (L&D) Departments to assess readiness to introduce the DHEP from 2020. They also interviewed recruits undergoing a graduate-entry programme designed and delivered by Police Now (PN), to track experiences of the pre- existing graduate programme that is closest to DHEP. SUMMARY DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME The learning from analogous conversion courses The single most important aspect of a successful professional conversion course is its ability to integrate theoretical learning and the learning of practical skills. This can be achieved through a variety of teaching tools, resources and opportunities to put theoretical learning into practice under the supervision of a capable mentor or tutor as well as providing the space for reflection on the experience and how practice can be improved. It is also important to be clear about the assumptions being made about graduate recruits’ critical thinking abilities, relevant skills and preconceptions about policing. This would sensitise curriculum developers and those involved in delivering education to understand the needs of graduates from the outset. However, it is equally important to understand that different individuals have different learning styles – and adapting suitable teaching methods for adults is needed for the delivery of both theoretical learning and practical skills. Another critical aspect of curriculum design and delivery is to put in place appropriate forms of assessment. These should serve as tools not only to test but also to enable learning. Assessment needs should be properly quality-assured and validated by experts. Finally, and fundamentally, outcomes which learning should achieve both in terms of practical skills (getting the job done) and practice skills (ensuring that behaviour promotes the proper goals of the organisation) must be specified. Focusing on practical day-to-day competence whilst losing sight of what makes for good policing is counter- productive. Research on effective police learning A thorough review of the international English-language research on police training identified 33 high-quality studies of police recruit training. Some very consistent messages emerge. The first is that teaching and learning methods have to be tailored to adult learning styles, and where recruits are degree holders, teaching methods appropriate to graduates are required. The principles of andragogy (adult learning) and of problem-based approaches need to be followed if deeper learning is to remain with recruits beyond the classroom and to survive the more corrosive elements of the police occupational culture. The theoretical and practical parts of programmes must be properly integrated if the theoretical elements are not to be jettisoned as soon as recruits arrive in the field and become inculcated with the occupational culture. This is essential but challenging to achieve. Related to this point, several studies stressed the importance of organisational buy-in to the aims of the learning programmes, and in particular in ensuring that teachers and mentors in the field are appropriately qualified and provide proper support for recruits when they acquire skills. Findings on programme aims and outcomes from the studied police force in-house learning and development approaches were disappointingly thin. 7 Aims were often poorly articulated and we have been unable to say much of significance about outcomes. Clearly the PEQF provides an important opportunity, that should be firmly grasped, to evaluate the outcomes of the three new training routes into the police in England and Wales. Any such evaluations would need to state fully and carefully, the aims of the graduate entry programme in terms of the qualities and skills that recruits should acquire. Police forces’ readiness to implement the DHEP and other parts of the PEQF There needs to be a significant shift in organisational culture to recognise and accept the value of graduate training for recruit officers. The workforce needs to accept the case for the PEQF. For this to happen, police leaders and the College of Policing need to articulate the rationale – and the benefits for the workforce – of professionalisation and graduate level education. Forces’ senior management teams need to support, and have the full support of, their L&D staff in this process. So too do forces need to think holistically about the changes to learning that require additional support. As this report makes clear, mentors are critical to the success of transferring new knowledge into practice. How mentors learn their role, and are supported in doing so in situ, must be considered as a fundamental part of the graduate entry programme. The daunting logistical challenges, such as getting contractual arrangements in place with Higher Education Institutions (HEIs), upskilling in-house learning staff and mentors and bringing the changes in on budget may distract L&D units from finding solutions to the difficult intellectual and pedagogic/andragogic challenges in designing and implementing a conversion course that equips recruits with complementary practical and intellectual skills. Tight budgets increase the risks of making false economies. The cheapest and quickest routes to learning how to be a police officer and bringing recruits to the level of independent patrol status may not be the best long-term investment for a police service. Without more careful thought about how all recruit officers achieve the competence of independent patrol status (regardless of which entry route recruits choose into policing in England and Wales), changes introduced may undermine the aspirations of professionalisation. Given the complexities of setting up partnerships between forces and universities and sharing the design and delivery of learning, there is a need for an independent body – with expertise in both the theoretical and practical aspects of police learning – to oversee these arrangements and to quality-appraise and assess the various training programmes on a national level. The College of Policing could take on this role, or an independent assessor could be found or set up. Finally, there is little evidence to date about the impact that making policing a graduate profession has on people’s experience of their police. The PEQF promises to achieve a radical transformation of police learning, and of the nature of policing in England and Wales. Now is the time to put in place rigorous evaluation plans to track the outcomes achieved by the PEQF’s three entry routes, so that recruit learning can be closely monitored and refined, as implementation proceeds. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Motivation and appreciation of evidence based policing practice amongst Police Now participants This part of the project involved psychometric data collected from PN participants. They emerged as a highly motivated and able group of graduates, with a clear commitment to ‘making a difference’. Measured on their emotional intelligence, they scored higher on empathy, emotional control, conscientiousness and structure than a benchmark group of comparable professionals, but lower in decisiveness and influence. Other key points relate to the way recruits learn about evidence-based policing practice. • Repeated exposure to examples of evidence based practice for the PN participants reaped rewards by increasing knowledge of basic principles and increased the likelihood of a more focused approach to addressing real-life policing problems. • Enabling PN participants to frame stronger research designs together with a greater appreciation of some basic tests of statistical significance, or appropriate qualitative methodological approaches, may reap earlier rewards for adopting evidence based practice. • There was little evidence to suggest that when PN participants undertook systematic problem solving approaches, these ideas were taken up elsewhere within force or the learning exploited for trialling elsewhere. • Further research is needed to show how graduate recruits learn about policing – as a craft as well as through better exposure to evidence based practice – and how better to transfer this knowledge to the whole of the workplace. In-depth interviews with Police Now participants Our interviews with PN participants and their police line managers about their experiences of this programme highlight some important learning points for developing the DHEP • While our interviewees could largely cope with learning over a condensed period of time and were happy to complete some of these elements independently, there was a general call for more discussion and reflection alongside lecture-style learning. This might also help reinforce the connection between the more theoretical or ‘soft-skills’ aspects of the training curriculum and policing practice. Common concerns about a lack of practical knowledge or skill on completion of learning the curriculum components could be addressed through better coordination in the immersion and mentoring period and clearer plans for how graduates’ practice will be developed in force and through top-up learning and reflection opportunities. • More generally, issues between in- force learning and development staff, HEIs (when these come on board) and those front-line staff who provided entry recruits with support and management – show a potential risk for the DHEP. PN participants are expected to apply and promote evidence-based practice. 9 However, interviewees met with various obstacles, including the lack of support for Evidence Based Practice (EBP) in participants’ local forces, and this would need to be addressed in readiness for the DHEP as well as the other two entry routes. • The PN impact presentations could be viewed as one opportunity for reflective practice, and these are promising practice for PN. The 100- day assessment process used by PN has several objectives including: to enhance participants’ problem-solving skills and to inspire reflective innovation and creative problem solving. There is clear scope for thinking further about how these presentations test and improve the police knowledge base, not only for PN, but across the developments in recruit education. Evidence based practice: researching police learning This project was fortunate to have the cooperation of Police Now (PN), whose participants, force based mentors and staff agreed to be interviewed and observed. The research on police learning across the globe is thin, and the establishment of the PEQF provides the opportunity to learn more about how police learn. This report is not an evaluation of Police Now. Police Now’s participants who took part in this research did so with expressed research consent and we thank these participants for taking the time to contribute to making police education better. The issues identified here are issues relevant to all the forces across England and Wales who will be adopting the DHEP. Fundamental issues and principles arising from this project’s research A number of key issues and fundamental principles need to be addressed for successful implementation of the DHEP: •• There is a need to achieve clarity on the model of police professionalism to be adopted •• The model of professionalism steers the respective underpinning principles for education and continuous professional development and should draw on learning gleaned from other occupations’ experience •• There is a need to refine on an ongoing basis the operationalisation of evidence based policing practice, integrating the recognition and inclusion of craft-based practice •• There remains a suspicion of the added value of graduate officers, and for the sake of those entering the police service through the DHEP, there is a need to address continuously the service’s receptivity to graduate officers so that they experience the welcome they deserve in the job •• There is a need to support stronger and more collaborative engagement with HEIs to ensure best fusion of theory and practice as well as coherent standards and delivery across forces •• It would be useful to establish a framework for and undertake an iterative longitudinal study of the DHEP that becomes a basis for continuous improvement in the education of graduates who join the police service. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Recommendations of this report for the College of Policing in developing the DHEP There are key findings that the College of Policing may wish to consider in the iterative development of the DHEP: •• To articulate clearly the desired outcomes of graduate entry education with explicit criteria for becoming a ‘good police officer’ and for achieving ‘good policing’ •• To articulate the aspired model of professionalism that underpins the PEQF by drawing on learning from other professions •• To review the published curriculum for the PCDA and concentrate on fusing theory and practice in the different modules including explicit consideration of applying EBP principles to practice •• To provide specific guidance for forces on working collaboratively with HEIs – especially in creating new ways of co-producing learning and integrating HEIs’ contribution to this new curriculum •• To give more consideration to the candidate (learner) perspective in policing education from their initial selection; including the consideration of personal motivation and other characteristics, and through ongoing support and monitoring •• To highlight the receptivity of the Police Service as a whole to the creation of a graduate profession of policing as an ongoing issue •• To support the undertaking of an on-going major evaluation of the DHEP, alongside its evaluation of the PCDA •• To encourage a suitable funder to allocate monies for a major study of the introduction of the graduate entry requirement. 1 1 The context of policing in England and Wales is continually and rapidly evolving, as is the profession itself. There are 43 police forces in England and Wales. The professional body developing the standards for the skills and knowledge necessary to prevent crime, protect the public and secure public trust is the College of Policing (COP), established in 2012. The College sets standards for the police service on training, development, skills and qualifications, and provides support to help the service implement these standards.1 A significant development has been the introduction, by the College, of the PEQF. Whilst this is ultimately designed to cover a wide range of the service’s educational needs, a key element of the PEQF, which has been first for implementation, is the requirement that all new recruits should acquire a degree level policing qualification either before recruitment, upon entry or in the first years of service. The project summarised in this report2 supported the COP in the development of one of the three entry routes, the DHEP, which is envisaged as a conversion course akin to those for graduate entrants to law, teaching or social work. The project addressed four sets of questions: 1. How do graduate police recruits learn about the standards for policing and the public expectations of a police officer? Is there relevant learning from academic research on other graduate entry professional conversion programmes? 2. How do graduate police recruits apply their critical skills (gleaned from their university study) to their decision- making as a police officer? What is the role of evidence based policing here? 3. How do graduate police officers learn essential skills/craft of policing? Do they have different learning styles from non-graduate recruits? What is the role of evidence here? 4. What is the Theory of Change for requiring police entrants to have degrees – that is, what are the key elements of policing that should be instituted by effective police training and education, and what are the key steps in making this happen? INTRODUCTION DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME 1. www.college.police.uk/aboutus 2. Fuller – and fully referenced – reports are available relating to each chapter. Contact details for the relevant academics are provided in each chapter. 3. http://www.college.police.uk/What-we-do/Learning/Curriculum/Initial-learning/pages/initial-learning.aspx 4. http://www.college.police.uk/What-we-do/Learning/Policing-Education-Qualifications-Framework/Pages/Policing-Education- Qualifications-Framework.aspx The Policing Education Qualifications Framework (PEQF) For much of the post-war period the training of police recruits was in Regional Training Centres, such as those at Ashford, Bruche, and Hendon, with a focus on learning the policing ’craft’ and the law as it related to policing. The regional framework was reviewed in 2003 and abandoned in 2006, with training provided thereafter through a patchwork of local provision, known as the Initial Police Learning and Development Programme (IPLDP). The IPLDP is pitched at Level 3 of the National Qualifications framework, equivalent to A-levels, and is delivered by police training centres, supported by mentoring arrangements in police divisions.3 Former Chief Constable Peter Neyroud’s (2010) review of police leadership and training reset the direction of travel of police learning on a path towards professionalisation. He argued that policing should move from being a service that ‘acts professionally’ to becoming a ‘professional service’. He proposed (amongst many other recommendations) a Level 4 (i.e. sub- graduate) Point of Entry Qualification along with a range of additional provisions to provide for better qualified managers and leaders. The College of Policing’s draft PEQF, published in 2015, following the Neyroud Review, went further in its recommendations. Firm proposals, published in 20164 after consultation, stated that a point of entry qualification would be required, at Level 6, commensurate with a university degree. The PEQF provided for three entry routes for police constables: •• the PCDA (introduced from April 2018), involving three years of study in parallel with operational work •• the pre-join degree in professional policing, introduced from 2020, where prospective recruits acquire their degree in policing in advance of recruitment •• the DHEP (introduced from 2020) – the focus of this report. The PEQF sets out a consistent, national approach for the training of police recruits. This forms part of a programme of police practitioner accreditation across the ranks, which is ultimately intended to achieve a level of professionalisation in policing that has only really existed in the past as an aspiration (cf Holdaway, 2017). Key features in this professionalisation process include: •• the establishment of a professional membership body •• the specification of an organised body of knowledge which is central to learning the skills of the occupation •• making a professional qualification, usually at degree level, a precondition for practice •• developing a code of ethics that governs professional practice. 1 3 Rationales for professionalisation It is important to ask not only whether the current push towards police professionalisation exhibits the recognised traits or criteria that constitute professional status, but why such policies are in place. There have been calls for the greater professionalisation of policing from the 1960s and 1970s onwards, notably by Sir Robert Mark (1977). Sometimes these were clearly intended to address conduct that was clearly unprofessional, especially corruption. Other calls were for improvements in quality – to ensure that the workforce simply do the best possible job. And of course, underlying both these impulses for professionalisation was a hope that policing could thereby acquire more professional status. Beyond these justifications is a more significant one, that professionalisation provides the key to an alternative form of institutional regulation. Historically the structures of professional bodies emerged in the 19th century as a specific mode of regulation for occupations requiring skill, judgment and integrity, involving geographically dispersed and autonomous, generalist practitioners, where incompetence carries high social costs.5 This mode of regulation is distinctively different from others, such as military hierarchy, Taylorian scientific management systems and Weberian administrative bureaucracies. These all place more emphasis on authority structures, and the latter two also involve well-defined divisions of labour. The burden of knowledge in these more hierarchical forms of accountability fall upon senior commanders or managers, with the requirements of obedience falling upon less skilled front-line staff. Different professions place different weight on the various elements of professionalisation. Professional knowledge carries particular weight in the case of the traditional “learned professions” of medicine and law. The same is true for those ‘helping professions’ whose practitioners are geographically dispersed, and often unbureaucratised, such as psychotherapists and physiotherapists. It is their claim to have a specific body of knowledge that legitimates their authority to practice. Amongst the functions of their various professional bodies, ensuring that practitioners remain ‘up to speed’ with new developments in the knowledge base (continuous professional development) is important – unsurprising, given that the body of knowledge of these professions is constantly developing. Another important function is the enforcement by the relevant professional body of their code of ethics. For other professions, the requirement that practitioners have at entry a degree-level qualification is better understood as a strategy for attracting people into the workforce with particular intellectual abilities: analytic skills, the ability to think independently and critically, and the acquisition of a theoretical understanding of the problems that they address, essential for problem solving (Tilley and Laycock, 2017). DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME 5. There is, of course, an alternative and more critical account of professions as self-serving organisations designed to exclude competition and to maintain the benefits of their members. If professionalisation emerged as a form of regulating uncoordinated, independent practitioners in the 19th Century, many of the occupations now recognised as professions have been bureaucratised. Most teachers, academics, and social workers, and many medical professionals, work in state bureaucracies. Some professionals in the private sector now work in very large – often national or multi-national – organisations: accountants, architects, and engineers, for example. In these cases, the bureaucracy provides a further layer of regulation, specifying functions and best practice, and enforcing standards. These two dimensions – the salience of the professional body of knowledge and the extent of bureaucratisation – form a matrix in which the various professions can be located, as in Figure 1. There is obviously room for argument about precisely where each profession should be located. We leave it to readers to make their own judgments about where in the matrix they would locate teachers, for example, or lawyers, GPs, surgeons, engineers, architects and other professionals. However, we hope that it is clear that different occupations belong in different quadrants. What is less clear is where the police would sit in the various prescriptions for professionalisation. It is obvious where the police were located on the matrix in the past: in the top of the top-left quadrant. There was no requirement for any entry qualification, no conception that formal knowledge, rather than skill and craft, should guide front-line practice, and no expectation that front-line staff should operate fully autonomously. Of course front-line staff have always operated with considerable discretion, and one reading of the hierarchical nature of policing is that it is precisely a system for regulating discretion. It remains unclear – at least to us – precisely where the current professionalisation process will relocate policing on the matrix. It is clearly intended to transform policing into a knowledge-based profession to some considerable extent. It can also be seen as part of a process for establishing what Holdaway (2017: 591) has called “a hybrid, fragmented form of regulatory governance”, which will certainly distance the police service from the Home Office, and arguably loosen the grip of police force hierarchies on their workforces. The medical model of professionalisation, if that is to be followed, would certainly confer more autonomy on front-line staff – akin to GPs, who make significant treatment decisions by reference largely to the medical knowledge base. On the other hand, the sort of professionalisation envisaged by Peter Neyroud’s review – with additional education and qualifications for senior staff – suggests that familiarity with the professional knowledge base is mostly required in the middle and top of the police management hierarchy. 1 5 DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME FIGURE 1: STYLES OF PROFESSIONALISATION BUREAUCRATISED EMPLOYMENT INDEPENDENT PRACTITIONER STATUS QUALIFICATION NEEDED ONLY AT ENTRY KNOWLEDGE BASE DETERMINES EVERYDAY PRACTICE 1 7 The authors of this report encourage the College of Policing to articulate more clearly the style of professionalisation it envisages, on these dimensions of hierarchical accountability and professional knowledge. There may be a case for retaining some constructive ambiguity about this, because it remains unclear how far down the route to evidence-based policy and practice the police service can realistically travel, and over what time period. Clearly it can make greater use of evidence, but defining the boundaries of policing’s professional body of knowledge – and filling the gaps within this boundary – is a significant undertaking that is likely to take decades. And the degree to which policing can be transformed into an evidence-driven institution in turn determines whether loosening of the hierarchical structure of police accountability is possible or desirable. On the other hand, designing the shape of the entry level qualification, and the routes by which it may be acquired, does require some clarity, both about the relationship of the knowledge base to policy and practice and about changes envisaged in the hierarchical structure. Our findings can speak to some – but not all – of these issues. Figure 2 below sets out what we take to be the theory of change, and the logic model that is implicit in the College of Policing’s current proposals for the DHEP. FIGURE 2: THE LOGIC MODEL AND THEORY OF CHANGE UNDERPINNING THE GRADUATE ENTRY PROGRAMME Graduate entry: Academic degree harnessed and accelerates learning / training in becoming a Police Officer in England and Wales Uplift approach to recruit training and integrate HEI content to assure functional equivalence of three entry routes into policing Focus on police constable basic field skills together with an uplift in delivery level of learning DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Entry programmes cross-fertilise field and knowledge; iterative approach to the balance of field and learning; integrate police cultures within three entry routes; recruits use information and evidence to improve constable ways of working Achieving independent patrol status; creating independent decision makers who can draw on best knowledge at beginning of career; foster high professional standards What does success look like? Police officers are trained in EBP techniques and methodologies; knowledge is applied to everyday policing in a everyday context; uplift in police officers using EBP approaches; Police training materials reflect evidence base; Recruits embed an approach to policing that enables improvement in managing local problems; Gaps in knowledge base of transmitting policing practice are continually identified and proactively addressed How will we know we have achieved success? 1 9 DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Researching the value of a degree for police work Clearly the remit for this research took as given that degree entry would confer considerable benefits on policing, and we were not asked to review the research evidence in support of this. Nevertheless we should say something about the supporting evidence. First, we should stress that there are different versions of the argument for graduate entry. One is that a degree qualification – regardless of subject – serves as a guarantee that the holder will have certain qualities, such as analytic skills and independence of mind. A more narrow claim, and one implicit in the PEQF, is that providing recruits with relevant skills and knowledge to degree level will help them do a better job. Secondly, research evidence in support of either proposition is unlikely to be clinching. Some differences are likely to be subtle, with the desired effects accruing in the long-term; and in the short-term, degree-holders have had to find their place in a workforce whose culture has not been positive towards higher education (Lee and Punch 2004; Silvester, 1989, Hallenberg and Cockcroft, 2017). There is research, mainly from the United States and much of it dated, on the value of in-service acquisition of degrees. Paterson (2011), reviewing the 1970s American research, found graduate officers were less authoritarian, less cynical, and behaved more ethically and professionally than non-graduates. This body of research concluded that it was not so much a particular course but the overall university experience that made the difference. Paterson’s review of research in the 1990s found that graduate officers had improved knowledge of the criminal justice system, made better sense of managerial issues and were less likely to use coercive force. On the other hand, in a separate review, Paoline, Terrill and Rossler, (2014) found few differences between graduate officers and others in terms of attitudes and professionalism, job satisfaction and stress, a conclusion supported by their own more contemporary study. They put the lack of difference down to the overriding effects of socialisation into the police occupational culture, and graduate officers’ frustration at being assigned to low level policing tasks. The safest conclusion to draw from the available research is that the outcomes of decisions about professionalisation and graduate entry are likely to be highly context-specific, and in particular that desired effects can be undermined by staff resistance – but that there are real potential benefits to be grasped by requiring a degree-level qualification at the point of entry into the police service. Different professional knowledges: what counts as the policing evidence base? The natural assumption to be made about the nature of the policing evidence base is that it only addresses ‘what works’ in reducing crime – by analogy to the National Institute for Health and Care Excellence’s (NICE) evidence base6 for medical and social care. NICE’s website sets out – with a high level of granularity – the best treatments for medical and social care problems. 2 1 In reality research evidence has charted the wider range of police functions, from tackling crime to public protection, maintaining public order, and dealing with emergencies that need an immediate response. There is a growing but very incomplete body of knowledge derived from experimental research about what works best in discharging all these functions. Overlaying this work is a more reflective body of academic research that offers various ‘theories of policing’ that characterises how public order and the rule of law is best maintained.7 Abutting this work are studies of the police as an occupation, charting the policing culture(s) and exploring how best to secure organisational commitment and compliance (and all this leaves to one side the more scientific body of knowledge associated with issues such as forensics). The key point here is that the evidence base can be conceptualised as sitting on a continuum. At one end of the continuum, it can be thought of as reliable what works knowledge,8 formulated at a high degree of granularity, that provides evidence-based solutions to the vast majority of problems that a patrol officer (or neighbourhood officer or detective) will encounter. At the other end of the continuum the core elements of the evidence base can be thought of as theories or perspectives that provide police officers with principles against which they can test different solutions to the problems they have to tackle. This sort of knowledge will generally be more provisional and contestable than the highly granular knowledge that one might expect to find in a well-developed ‘what works’ warehouse, but with a much wider range of applicability. Our own view is that police too often want and expect the policing evidence-base to yield unambiguous solutions to specific problems; whilst what they need is a set of well-thought-through principles from which they can derive the solutions they seek. The latter is severely lacking in police learning. Whichever characterisation of research evidence is judged more appropriate for policing, both are in tension with the assumptions firmly embedded in current learning approaches for today’s police recruits. These assumptions are that policing is a craft, and that training police recruits in ‘what police do, when they do it, and how they do it’ is best done by former or serving police officers who hand on the knowledge in a practical apprenticeship. In this report we are making a distinction between ‘training’ and learning. PEQF, we suggest, requires learning strategies and learning outcomes. The PEQF gives HEIs a central role as a co-producer, together with police forces, of learning delivery and assessment of competence. The contribution that HEIs will make to the creation of innovative 6. https://www.nice.org.uk 7. Much of the sociology of policing from the 1960s and 1970s remains relevant, e.g., Banton’s work on police and community, Packer’s models of crime control and due process policing, and Bittner’s work on the functions of the police. More recent work such as procedural justice theory, is also of central importance. 8. No ‘what works’ evidence is likely to lie beyond challenge. Better solutions come along, in time. And in any case, research evidence generally needs to be blended with practitioner judgement before it is fully usable. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME approaches to useful principles and learning frameworks is not a simple ‘add and stir’ formula. Bringing academic approaches to learning in this new PEQF approach to police recruit induction to the profession will present challenges to HEIs and current police training centres who are more attuned to traditional models held by each. Both sets of learning professionals are not yet ready to translate operational craft or scholarly knowledge into grounded, blended teaching and assessment. As this research strongly suggests, for the PEQF to succeed, HEIs and police must collaborate on a novel hybrid approach to learning – one that grounds scholarly ‘what works’ knowledge into principles which guide new ways of thinking, decision making and practice. So at this critical juncture in the journey of PEQF, intelligent strategies are needed for blending into the curriculum the craft skills and professional knowledge of experienced officers with the more formal research evidence that is generated within universities and other research centres. This takes us to one of the golden threads – the introduction of EBP – that must be welded into the PEQF curriculums for police officer recruit entry. The College of Policing offers a definition of evidence based policing as a way of working that ‘creates, reviews and uses the best available evidence to inform and challenge polices, practices and decisions.’9 This approach recognises both knowledge that is generated through collaboration with academics and other partners. It also recognises that evidence – such as professional consensus and peer review – may be regarded as the ‘best available’, if this is gathered and documented in a careful and transparent way. Evidence based policing practice (EBP) is a required part of the curriculum for all entry routes to become a police constable. At the risk of oversimplification, the curriculum for the policing entry qualification, regardless of which route is taken, will need to synthesise, in the right quantities, three rather different forms of evidence: a. The practical craft knowledge acquired through experience and tested against professional judgment, held within the organisational culture and transmitted through the organisation by formal and informal mentoring and training; b. Where the research already exists, the experimental ‘what works’ evidence that addresses questions framed at a very granular level about effectiveness and cost-effectiveness of different interventions; c. Research that offers theoretical understanding, or perspective, on the police role, including theories of crime, victimisation and offending, that will guide officers in their problem-solving roles. Under the PEQF, these three conceptions of evidence imply different approaches to teaching and learning, and how learning and development units approach the preparation of modules and materials covering ‘basic policing’. The transmission of craft skills is basically a process of demonstration of skills, supplemented by learning by rote of processes that have to be performed. The ‘what works’ model involves the transmission of formal knowledge at a higher level of generality and complexity 2 3 but does not privilege analytic ability and conceptual grasp of the nature of policing in the way that the third model does. Policing training, learning and development has traditionally worked with model (a), although police management now has quite a long history of engagement with research both in the hard and social sciences. The EBP movement has given much more priority to the second type of evidence – experimental evaluative research – and has tended to assume that this is the model applied in the archetypal profession of medicine. Critiques of the experimental approach are numerous and sustainable: not enough Randomised Control Trial (RCT) knowledge as yet; limited sensitivity to context or mechanisms; and insufficient recognition of problems of reflexivity/adaptation. The third sort of evidence fits in to a somewhat different model of evidence based practice, described well by Tilley and Laycock (2017) as the ‘engineering model’ of evidence based trial and error, which they see as highly applicable to policing. Finding a test-bed for new knowledge: Police Now Police Now (PN) is not a graduate conversion programme, but there are similarities between it and the proposed DHEP. Like the DHEP, PN trains graduates for a career in policing, using a different approach to traditional recruit training, and placing a greater emphasis on promoting an evidence-based approach to policing. PN was designed, using the COP’s recruit training syllabus, to train outstanding recruits over a two-year period. It exists to ‘transform communities, reduce crime and increase the public’s confidence in policing’, and its aim is to recruit and develop ‘an outstanding and diverse group of individuals to be police officers and leaders, working on the front line and contributing to wider society’.10 The programme consists of an eight- week Summer Academy (SA) ‘designed and delivered by outstanding, high performing frontline police officers... it is underpinned by a comprehensive pre-learn course so that participants have the legal and procedural knowledge they need to hit the ground running as well as a period of one-to-one mentoring afterwards’. For the remainder of the two-year training period, most PN recruits work in a neighbourhood policing team. Police Now describes the Academy as ‘innovative, dynamic, inspiring and challenging’. Police Now initially worked only with the Metropolitan Police Service (MPS), expanding to work with 7 forces in 2016, 19 forces in 2017, and 17 forces throughout England and Wales in 2018.11 The project has much to thank Police Now for in its willingness to be a test-bed for the ideas and challenges discussed in this report. None of these findings are meant to evaluate Police Now as a programme. The curriculum for inducting recruits into policing is set by the College of Policing, adapted and applied by 43 police forces across the UK. 9. https://www.whatworks.college.police.uk 10. https://www.policenow.org.uk/wp-content/uploads/2016/09/Police-Now-Programme-Booklet.pdf 11. https://www.policenow.org.uk/the-programme/about-the-programme/#forces-we-work-with DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Police Now relies on the IPLDP curriculum in its current approach. The project team had frequent and continuous conversations with various PN staff, and the findings of the studies were fed back routinely to PN over the course of the project. The project reported here explores what the challenges are not only for Police Now, but for all the 43 forces across the country when they will transition to the PEQF curriculum. The project This project’s origins predated the College’s announcement of the PEQF. MOPAC, with UCL and Birkbeck as academic partners and Police Now as a key stakeholder, secured a grant from the Home Office Police Innovation Fund. Its aims were to: “Enhance the effectiveness and professionalism of police officers by developing and testing a syllabus to embed evidenced ‘what works’ knowledge on tactics and styles of policing into police recruit training.” Police Now agreed that its graduate entry programme could be used as a test-bed for research to enable the development of its syllabus to include a sharper focus on ‘what works’ evidence. The College had started its consultation over the PEQF at about the time that the project formally began in Spring 2016, and it was clear at that point that the project would be particularly relevant to PEQF’s graduate entry route. The College of Policing was a formal project partner; and progressively over the first few months, the project was reshaped to provide whatever support it could for the development of the DHEP, which is, in essence, a two-year graduate conversion course that will provide existing graduates with a Level 6 Graduate Diploma in Professional Policing Practice. Police Now remained a central stakeholder for the project, as the graduate participants on its three cohorts have experiences and aspirations which are likely to be closely comparable to those enrolling in any graduate conversion course in the future. Detailed reports are available on the various elements of the project’s full work programme from the MOPAC website. 2 5 The key elements were: 1. A rapid evidence assessment of what can be gleaned from other graduate conversion courses for the PEQF graduate entry route, supplemented by interviews with experts. 2. An international systematic review of what is known about effective police training methods. 3. Completion by recruits of: a ‘tracker survey’, a psychometric instrument completed at three points in time during the course; a tool assessing recruits’ emotional intelligence; and an ‘evidence based policing toolkit’ assessing understanding of, and commitment to EBP. 4. In-depth qualitative interviews with Police Now participants in cohorts 2 and 3 of the programme, complementing the quantitative material collected at (3), tracking progress on the course, assessing reactions to it, and levels of commitment and satisfaction. 5. Interviews with Learning and Development (L&D) Unit staff across the country to capture their understanding of the DHEP and their readiness to implement the significant changes to learning approaches required by the PEQF. The shape of this report This report is structured as follows. Chapter 2 present the results of the rapid evidence assessment on graduate conversion courses in other fields. Chapter 3 summarises the results of the systematic review about effective police training. Chapter 4 deals with findings about forces readiness to implement PEQF. Chapter 5 uses three quantitative monitoring tools to describe the experience of Police Now graduates. Chapter 6 provides qualitative findings on the experiences of Police Now graduates. A final chapter offers some concluding thoughts and draws out the lessons of these findings for the DHEP. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Chapter 2 The evidence on conversion courses Jyoti Belur, Winifred Agnew-Pauley and Lisa Tompson 2 7 This chapter presents the results of a rapid evidence assessment (REA) carried out by UCL, examining effective practice in graduate conversion courses for other occupations. The objective was to inform the development of this route in the PEQF. Fifty-one studies were identified and summarised. The main themes emerging from the review concerned: •• learning styles •• translating theory into practice •• teaching methods •• assessment •• reflective practice. Inclusion in the REA was based on three criteria: that the study should relate to a graduate conversion programme for a professional qualification for entry into an occupational role; should have some description of programme content, pedagogy, duration, modes of delivery, outcomes, trainee experience or theoretical content; and should be in English. Over half of the REA studies were from the United Kingdom (n = 30), mostly from England (n = 14). Research was predominately within the teaching discipline (n = 40) with other studies relating to either law (n = 6) or social work (n = 5). Many of the conversion courses ranged between 36 weeks and one year (n = 29). Three courses were between 18 months or two years. Eighteen courses either did not provide any information or had a variable or unclear course length. Cross-cutting sub-themes were: the use of information technology (IT) or virtual learning environments (VLE); collaborative learning. The REA highlighted a number of more detailed questions worth examining in more detail, and e-interviews were conducted with seven people selected for their knowledge of conversion courses in education, law and social work. The results are presented at Appendix A. A full report of the findings in this chapter is available from j.belur@ucl.ac.uk Course structure Graduate conversion courses in teaching, such as the UK Postgraduate Certificate in Education (PGCE), tend to structure their practical training alongside the educational aspects of the course, with periods of university teaching interspersed with periods of practical training. This training is most often in the form of ‘placements’ made possible through partnerships between schools and higher education institutions. Teaching placements range from around 4 weeks to 24 weeks. The most common arrangement within the reviewed 51 studies specified that the teaching placement would comprise two thirds of the course (for example, 24 weeks of a 36 week course). The remaining studies examined shorter placements (for example, 12 weeks) or had variable arrangements. Graduate conversion courses for law tended to structure their courses as stand-alone training preceding practical training. Hannibal and Pope (2005: 237) state that the Legal Practice Course (LPC) is “an intensive nine-month full-time programme of study… which prepares students for their vocational engagement as a trainee solicitor”. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Therefore, the acquisition of professional skills and practical experience comes at a later stage in the process and is not included on the graduate conversion course for law. Most of the social work studies relating to graduate conversion programmes gave no details on course structure, but according to the ‘Evaluation of Social Work Degree Qualification in England’ Team (2008: 95) the requirements for Social Work Training set the minimum standard for students as being at least 200 days in academic learning and 200 days in practical training, at either undergraduate or postgraduate level. All these courses involve partnerships between training institutions and HEIs. Six of the studies made specific reference to partnership arrangements. There are various partnership models, with some courses having informal arrangements, or arrangements in the early stage of development, and others having formalised partnerships. One study (Carter 2015) found that partnership arrangements that follow the ‘clinical practice’ model are considered to be best practice, whereby students spend sustained periods of time in the same policing setting under the supervision of an experienced mentor. Furlong et al. (1996) differentiate three models of partnerships in teacher education: collaborative, HEI-led and separatist: “In ‘collaborative’ partnerships, teachers and HEI tutors work together in planning and discussing professional issues. The process is characterised by a genuine search for shared understanding and mutual respect. In an ‘HEI-led’ partnership, assessment and the content of the placement and planning is designed by the HEI with, at most, consultation with small groups of teachers. In the ‘separatist’ model, each sector is seen to have its own responsibilities without any attempt at dialogue and with integration being achieved by the students themselves.” Ideally, the REA suggests that graduate conversion courses should have formalised partnership models in place between learning and development and HEIs that are collaborative and based on the ‘clinical practice’ model. Learning styles Graduate conversion courses need to acknowledge that applicants, as adult learners, come with a diverse range of educational and academic experiences and work or practical expertise. This should inform course design and the range of teaching methods used, as it is important to take account of the starting point of trainees. The literature on learning styles stressed three key points: understanding that students learn in different ways and have different learning styles; the importance of acknowledging past learning and experiences; and taking into consideration prior perspectives and concerns. Students of graduate conversion courses not only bring prior experiences, but also concerns and fears surrounding entering the workforce. Students starting initial teacher education courses have different levels of knowledge, understanding and experience about teaching, schools and pupils which may result in different causes of concern for students. It is important for courses, and educators, to provide opportunities for students to identify and address individual concerns so that they might develop as professionals. 2 9 The take away message is that conversion courses are designed for adult learners and must consider individual learning styles of students as well as their prior experience, preconceptions and concerns about the profession. One potential way forward, addressed later in this report, is for course tutors to identify these concerns and preconceptions about the profession early on and address them by focusing on the theory and conceptual knowledge underlying the practice elements of the profession. Translating theory into practice This theme refers to the relationship between the educational, academic or conceptual components of a course and its practical or field training aspects. While these two components are often referred to as separate entities, both are integral and (ideally) interconnected. The practical components of a course are the practical tools introduced in both the fieldwork and classroom settings, whereas the conceptual theoretical components involve the “conceptual tools which facilitate teachers framing and interpretations of practice, but they do not offer specific solutions for practice” (Tang et al. 2016: 55). Within practically orientated professions, such as teaching, social work, or policing, there is a consensus that the practical components of the courses provide students with the essential skills to be a competent professional. Furthermore, studies have shown that students tend to privilege the practical elements of the courses over conceptual or theoretical aspects, as this is where ‘job preparedness’ is established. However, as outlined above, it is the conceptual or theoretical components of training that inform these skills and provide the foundational guidelines that underpin practice. It is therefore essential that courses emphasise the integration of theory and practice, or devise innovative ways to translate what is learnt in the conceptual components of the course to the skills acquired during practical components. Theoretical components of a course stimulate higher order thinking, debate, theory building and discussion. Costello and Aung (2015: 590), discussing social work education in Myanmar, outlined that social work education is “guided by principles of social justice, human rights and ethical practice and their development of critical thinking, reflection, and analysis of power, inequality and disadvantage guides their practice”. This demonstration of how conceptual components guide ethical practices and alternative perspectives is particularly important for graduate conversion courses, as students may come into the course with certain perspectives or understandings, that new knowledge has the ability to challenge. This is referred to as conceptual change, which can be made possible when “students feel dissatisfied with their current conceptions and have access to alternatives they perceive as intelligible, plausible and fruitful” (Wong et al. 2006: 2). Without these stimulating elements of conceptual learning, students are more at risk of succumbing to institutionalised professional culture and losing their enthusiasm: “without a robust theoretical knowledge, student-teachers are ill- equipped to resist institutional constraints on their teaching… and are unable to resist the sometimes anti- intellectual climate of school staffrooms” (Wong et al. 2006: 2). DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME This is not to say that students’ prior experiences are wrong, or need to be disproven, however it is important that prior learning is acknowledged, alternative perspectives are offered, opportunities to reflect are built in, and tools are provided to translate attitudes, beliefs and values into actions. The take away message is that the integration of the practical elements of training and the theoretical elements are achieved in order to ensure that the practical element, as it is considered to be indicative of job worthiness, is not privileged over the theoretical elements, which underpin the ethos of the profession relevant to the context it is practised. Teaching methods Methods of teaching were noted by a majority of the studies. A wide number of teaching methods were discussed, all linked to different pedagogical theories. These included IT/VLE, the use of videos or video-papers,12 academic teaching methods such as exams and written assignments, peer review, lesson planning or lesson study, portfolios, critical incident analysis, oral presentations, role play, blogs and interviews. Overall teaching methods ought to “promote cooperative, active, inquiry-based learning” (De Jong and Chadbourne 2007: 15), and utilise the best available evidence when it comes to innovative and diversified teaching methods, utilising tools such as IT, and framing methods with pedagogical practice. Where possible, it is important to shift focus away from the traditional classroom-based model of teaching, and encourage collaborative learning and critical thinking through engagement between student and teacher as well as among peers. Methods of teaching must have “a strong emphasis on inclusion, active and participative learning, student empowerment and cooperative approaches” (Ofsted 2005). Equally important is a diverse range of teaching and learning activities to stimulate interest and keep students engaged. Optimally courses should be designed around a blended model of face-to-face teaching and traditional lectures and tutorials with online, collaborative components. The main pedagogical practices in the literature were reflective practice, communities of practice, IT/VLE, peer learning and mentoring. These are discussed in more detail below. The main take away message is that teaching methods ought to be innovative, engaging and appropriate for adults; should integrate theory and practice; encourage reflection, and include a diverse range of teaching methods in order to take into account students’ different learning styles. Assessment The evidence indicated that, like choosing appropriate teaching methods, assessments of students should be based on pedagogical principles that incorporate a diverse mix of innovative methods. Aside from the necessary evaluation of student competence, assessment should also be a learning experience, have practical relevance and, as noted, be guided by pedagogical principles. 12. A video paper is when text is accompanied by video within a single electronic document so that the reader is able to “activate different modes of presentation, watching the context in which the text is placed as well as presenting authentic examples to improve the validity of text” (Krumsivik and Smith 2009: 271). 3 1 As outlined above, graduate conversion courses consist of the two essential components of theory and practice, both of which need assessment. When this is focused on the theoretical aspect, it should be closely linked to innovative learning methods; when addressing practical competency, it should be practically relevant. As with effective teaching methods, measures of assessment ought to shift away from a heavy dependence on traditional methods (such as written assessments and exams), and incorporate a diverse and innovative mix of assessments that promote learning and are practically relevant. As Roness and Smith (2010: 180) remark: “Students may get frustrated if they experience heavy demands for written assignments that lack relevance to their practical teaching”. Practices identified as effective for assessment that also encourage learning include: “effective questioning, sharing learning objectives and assessment criteria, providing effective feedback about how to improve self-assessment and peer assessment” (Winterbottom et al. 2008: 194). Studies provided several examples of effective methods of assessment. Portfolios, as both a teaching and assessment method, reoccur as a tool for showing how students integrate theory and practice, reflect critically on their learning, and demonstrate the breadth and depth of their learning. Concept mapping, where students visually represent concepts and how they are linked together, has been shown as effective in measuring not only students’ content knowledge, but also for “promoting and assessing conceptual change” (Reitano and Green 2013). Assessing a student’s practical competency is of equal importance. The most common method is observation of trainee professional performance. The important take home learning from the evidence is that assessments, when used innovatively can both promote learning (assessment for learning) as well as evaluate student learning (Assessment of learning), especially if it encourages critical reflection. Reflective practice Reflective practice emerged as an integral theme across most of the included studies. Reflective practice is fundamental to producing competent professionals. The literature revealed that infusing graduate courses with reflective practice would produce a range of benefits essential to enhancing the learning experience, integrating theory and practice and developing good practice, higher order thinking, and professional development. Reflection can take place through a range of different tasks or observations. At the simplest level “these processes of reflection are often understood to be triggered through completion of various administrative and technocratic mechanisms”, for example planning lessons in teaching (Lamb and Aldous 2016: 100). However, the key to these reflective tasks is to engage with individual reflexive processes and employ a cyclical process of planning, practice, reflection and revision. Reflexivity can be understood as a “set of meaningful processes in which the individual agent reaches a heightened point of awareness regarding their own reflective practice” and their understanding of how experience can shape understanding (Lamb and Aldous 2016: 102). DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Reflective processes should occur continually throughout education and training, in particular before, during and after practical placements, in order to ‘scaffold’ or continually build on professional development. The different stages of reflection are referred to as reflection-on-action and reflection-in- action: “Reflection on-action can be defined as the self-evaluative thinking that teachers engage in after most lessons. Killen (2007: 96) regards reflection-on-action as a deliberate attempt by the teacher to understand past events in order to shape future actions” (Taloe 2012). Furthermore, reflection should occur both individually, and among peers as an outcome of collaborative learning. The concept of reflection has been embedded in all themes that emerged from the qualitative synthesis, and while it is presented here separately, it is also a unifying theme throughout the literature on graduate conversion courses. Allowing time for reflection gives space for both students and trainers to consider how prior experience and learning styles will impact on the understanding of new knowledge. This space to reflect also helps develop the capacity to build on and integrate both old and new knowledge and to translate theory into practice. As put forth by Lamb (2015: 344), “acknowledging uncertainties and complexities in practice and engaging in meaningful deliberation are necessary for generating insights and for interrogating one’s practice in a way that leads to sustained change”. This process of reflection is also an integral element of higher order thinking and critical thinking: ability to compartmentalise knowledge and draw on different forms of knowledge and experience in different contexts is the mark of a competent professional. The message here about reflective practice is critical to elevating training to learning, the central argument throughout this report. The use of IT/Virtual Learning Environment (VLE) VLE is a teaching method that featured in many studies as a tool to facilitate the teaching/learning process and foster reflective practices.13 Some examples of use included: online platforms such as Blackboard or Moodle, electronic discussion boards, using videos or video- papers and blogs. Online platforms create spaces for formal and informal interaction between students and between the students and the tutor. It is also a useful tool for distance learning and can reduce isolation while students are on placement. Falloon (2012: 9) suggests that there is a “need to blend significant offline preparation with synchronous online interaction, so that what goes on in classrooms represents deeper, more meaningful learning, rather than shallow, superficial interaction”. Collaborative learning acts This is an umbrella term covering: mentoring, communities of practice, peer learning and placements. These sub- themes are all examples of active engagement in the learning process between students and their peers or educators. While each of these sub- themes can be discussed as separate concepts, they are all linked by this guiding principle of collaborative learning. Mentoring is a crucial element in the teaching and learning process. The role of mentors is to observe, guide, support, motivate and evaluate recruits. In contrast to supervision – where the teacher observes and oversees learnt 3 3 material – mentors are more actively engaged in the learning process. They can give students evidence of good practice, offer feedback and support, and describe the roles and responsibilities of a competent professional. Peer learning is a further example of effective collaboration which is thought to be key to developing professionals. Peer learning can involve mutual peer dialogue and support, peer observation and peer assessment or review. Peer learning is underpinned by the concept of communities of practice. Implications for police conversion courses The evidence presented here indicates that the following factors need to be considered in creating any graduate conversion course for police entrants under the PEQF: 1. Articulating assumptions that are being made about the students’ critical thinking abilities, relevant skills and profession related preconceptions that the students might bring to the table. This would help the curriculum developers and those involved in delivering the teaching to understand the needs of prospective students and address gaps at the outset. For example, it is essential to understand students’ perception of the police as a force in order to engender a discussion about the purpose and role of the police (exposure to various policing theories) in order to underpin the purpose of the learning programme. 2. Understanding that different individuals have different learning styles and adapting suitable teaching methods for adults are essential for delivering the theoretical and conceptual learning as well as operational skills required in policing. Adapting learning space to learning styles based on adult experiential learning theory has been shown to have some value in higher education. 3. The most important aspect of a successful professional conversion course is its ability to integrate theoretical learning in practice. This can be achieved through the use of various teaching tools, resources and opportunities to apply theoretical learning to practice under the supervision of a capable mentor, as well as being provided the space for reflection on the experience and how practice can be improved. 4. Conceptualising appropriate assessment so that it is a tool to enable learning (assessment for learning) as well as test learning (assessment of learning) and ensuring that the assessment is quality assured and validated by objective experts is another important aspect of curriculum design and delivery. 13. “Communities of practice are groups of people who share a concern or a passion for something they do and learn how to do it better as they interact regularly” http://infed.org/mobi/jean-lave-etienne-wenger-and- communities-of-practice/ DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Chapter 3 The systematic review on effective police training Jyoti Belur, Winifred Agnew-Pauley, Brendan McGinley and Lisa Tompson 3 5 Introduction This chapter presents the findings from a systematic review of the available research evidence on police recruit training. The review, which took a ‘realist’ approach, includes a total of 33 studies conducted in different countries. We searched 11 relevant electronic databases, including grey literature and dissertation databases, and searches of publications by relevant government, research, and professional agencies. We included studies in this review if they: 1. related to an entry level training programme for new police recruits; 2. reported the findings of an empirical research project on police recruit training; 3. covered substantive content on police training; 4. and were in English. We initially identified 109 studies meeting these criteria, covering six broad themes. After consultation, it was decided to focus on the 33 studies that were (a) methodologically rigorous and (b) focused on two themes: academic and field training, and how recruits learn. Fuller details and additional findings from the review are in two detailed reports and can be obtained from j.belur@ucl.ac.uk The studies Twenty of the 33 studies examined programmes in the Unites States and five covered ones from the UK. Other studies were from Australia, Europe, Canada, Saint Lucia, South Asia and South East Asia. The studies mainly covered police training academies, but often working in partnership with a university or a college. Two of these programmes in the US led to the attainment of a degree. Whereas research on recruit training in the UK studies post-employment, the US studies covered both pre-join and post-join models. Most training programmes covered in the studies were delivered in standalone blocks of academic and field training components, for example 20 weeks of academic and skills training in a training academy, followed by 10 weeks field training at a police force. Some were structured with the first block in the academy, second block in field training, and a final block returning to the academy. Five studies described training programmes with interspersed periods of academic and field training. The field training component also appeared to be longer in programmes included in the US studies than in the UK studies. The 33 studies identified a wide range of mechanisms which were regarded as key factors in achieving the desired educational or training outcomes. We shall first present findings on best practice identified in the studies relating to teaching and learning strategies. We then examine issues relating to the integration of theory and practice. Next we summarise what research says about implementation. Finally we report on aims and outcomes. Teaching and learning mechanisms The studies discussed a range of student-centred learning theories, the main ones being andragogy and problem-based learning (PBL). Andragogy is a learning philosophy specifically relating to the teaching of adults, contrasted to pedagogy – the teaching of children (cf. Vodde, 2008). DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Andragogy emphasises active involvement of the learner in the learning process, as opposed to passive listening, and building on existing knowledge and experience. Andragogy also recognises that adults vary in their learning styles. Andragogy was identified as a mechanism in nine of the studies and PBL in five. Problem-based learning follows a similar approach to andragogy in that it is “learner-centred and emphasised experiential learning with passive facilitation and active learning” (Lettic, 2016: 25). Both learning theories centre on active participation in the learning process. Whilst andragogy focuses on building on past experiences and different learning styles, PBL emphasises working through problems in a collaborative way: “it is through learners’ active participation in the learning event, the use of language and interaction with others, that learning occurs” (King Stargel, 2010: 14). The goal or outcome of PBL is to develop the learner’s problem-solving and critical thinking skills. Within the studies examined, both these learning theories were identified as particularly important for police recruit education: “Police education should make greater use of adult learning principles that encourage a learner- centred approach, promoting critical thinking skills within authentic learning contexts related to their policing duties” (Shipton, 2009: 58). Other techniques that were considered to be effective for encouraging critical thinking and problem solving were scenario-based training, group discussions and debates and opportunities for reflection. The first key teaching mechanism that we found related to the use of “practical, scenario-based teaching that includes active participation” (McCay, 2011: 87). Both scenario-based learning and ‘active involvement’ of the learner were identified as a mechanism in six studies. Learning techniques such as “group discussions, simulation exercises, hypothetical case scenarios, PBL activities, the use of the case method, and peer-helping activities, all serve to capitalise on the experiences of the learner, thus enhancing the learning experience” (Vodde, 2008: 94). This is in line with previous research which found students prefer and respond well to practical learning, as opposed to traditional lecture-based teaching. Scenario-based teaching also allows students to be presented with useable examples to draw upon. An example of this was trainers’ narration of stories about actual experiences and how to respond to them which was found to be very helpful by recruits in answering exams. Peer learning through group discussion and/or debate was considered a mechanism in five of studies. This is another example of experiential learning technique during interaction between peers: “Students do not come to know a concept just by having it presented to them. They must construct it through exportation, reasoning, and discussion. Various concepts and skills need to be scaffolded on top of students’ prior experiences and together with each other” (Porodzisz, 2004: 13). 3 7 This quote demonstrates the importance of recruits formulating and constructing knowledge based on prior experiences, in line with andragogy, and in discussion with other peers, adhering to PBL. The learning process must involve debate and discussion in order to challenge recruits’ prior learning and encourage thinking about different perspectives. Lettic (2016: 26) found that “students benefited from the differing opinions, literature resources presented, and disagreements [which led] to better information retention, understanding, integration and application of knowledge”. Nine of the included studies referred to the importance of reflection, as in the process of thinking back on one’s practice or experience and considering why something was done a certain way, what the possible impact of this was, and if it could have been done differently. Reflections challenge students’ pre- conceived knowledge, past experiences or attitudes. The use of ‘reality checks’ can ensure that student views are challenged and placed within the correct context and, trainers frequently check in with students to ensure the correct learning is taking place. Further, collective reflection “exposes subjects to a wider base of experiences and involves mutual resources” (Rantatalo and Karp, 2016: 720). However, the key to all of these mechanisms was considered to be effective facilitation and leadership by academic and field trainers and tutors. The evidence suggests that seamless integration of andragogy, combined with learner specific teaching methods that encourage group discussion and reflection on practice at both the individual and collective levels would be very effective in delivering the curriculum while encouraging critical thinking skills. Integrating theory and practice Integrating theory and practice emerged as an important mechanism in seven studies – even if they did not always articulate precisely how this should be done. The importance of consistency between the academic learning component and field training component through the application of learning principles, the role of the tutor, and reflection and debriefing between the recruit and the tutor were identified as field training mechanisms, which encourage recruits to draw links between their academic learning and practical experience. The design of recruit training programmes has important implications for the integration of theory and practice. Interspersing academic and practical components of the course provides more opportunities for recruits to apply what they have learnt in practice, and to reflect on what they have experienced operationally, to help integrate theory and practice. This can include introducing time spent back in the university while recruits are in the field training phase, or interspersing university learning with community placements. Heslop (2011b: 337) comments: “There were perfectly sound logistical and pedagogical reasons for sending the officers on placements only seven weeks into their training. This was at a time when it was envisioned that they would be open to learning new things as well as seeing things from different perspectives. At the same time, however, this also meant that while their experiences on the community placement had some meaning, they were not situated in their own field of policing”. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Moving away from the ‘traditional academy model’ (echoing the strict, hierarchical, ‘paramilitary’ model), by interspersing academic learning with practical experience, may also help counteract the negative aspects of recruit socialisation and police culture. Placements of recruits in organisations such as schools, hospitals, and charities have also been used to help recruits to better understand diverse communities and the workings of organisations within their policing area. Opportunities to integrate the academic components with practice arise during the field training phase. During this phase, recruits get their first opportunity to apply what they have learnt in the academy, and begin to experience socialisation. Furthermore, this gives tutors and trainers an opportunity to observe recruits’ performance and witness their development. Specific aspects of the field training component were identified as promoting the integration of theory and practice, considered pivotal for the training to be effective. Unless field training complements academic content, recruits may begin to “question the validity of their [academic learning] as a result of their initial exposure to operational policing” (Chan et al., 2003: 122). Successful integration of theory and practice occurs when students are able to apply what they have learned in the academic component to the practice environment, or ‘transfer’ their knowledge. This integration encourages recruits to “see the big picture”, and understand the theory behind their actions: “If the recruits understood the rationale behind their actions they… would be better prepared to learn from their own experiences in the future” (Charles, 2000: 81). Course consistency and coherence, whereby the learning theory and principles applied in the academic components carry through to the field training components, is crucial. As some police agencies have transitioned towards training police recruits in the principles of community policing, the disconnect happens because recruits “receive academy training in community policing, then go into field training programs based on traditional philosophies of policing” (Chappell, 2007: 499). A technique that can be used to facilitate reflection during field training and link theory and practice in a cyclical process is the Model of Experiential Learning Cycle (ELC). This theoretical framework involves a cyclical process of four key components: experience, reflect, interpret, and plan action. When applied in the policing environment, it can encourage recruit officers to critically analyse and reflect on practical experiences (cf Thorneywork 2004). Involving the learner in an active rather than passive way means there is a greater emphasis on applying knowledge instead of memorising it. One example of encouraging this is through linking key topics or ‘threading’ within the curriculum design: “Integrating key topics, or ‘threading’, means that selected important themes will be discussed in relation to each substantive topic or module (i.e., woven throughout the curriculum). For instance, communications lessons (or diversity training or officer safety) can be reinforced in arrest scenarios, in crowd control exercises, and in community relations material. 3 9 Integration helps recruits draw connections among multiple subject areas, which facilitates mastery over the curriculum and prepares recruits for problem-solving challenges when they enter the field” (Chappell and Lanza-Kaduce, 2010: 191). This encourages recruits to construct links between concepts and themes and to see the bigger picture of what is being taught, encouraging the ability to think critically and problem-solve. Another useful field-training technique to encourage integration of theory and practice was de-briefing, whereby the tutor discusses a practical experience with the recruit, encouraging the recruit to first reflect on their practical experience (for example, after an incident) and then giving useful feedback. Having a competent tutor to facilitate this process is vital. Tutors also need to be encouraging, and not overly critical. Encouraging recruits to reflect and improve on their experience is important in accustoming recruits with the cyclical feedback process that is an inherent part of problem-solving. The importance of the tutor’s role during this phase cannot be overstated, as the field training officer/tutor has a “long term impact on the recruit” (Novakowski, 2003: 147). “Tutors are seen to be the bridge linking the training environments where the students have to contend with group dynamics and the realities of policing” (Thorneywork, 2004: 48). Tutor constables need to be supportive, trustworthy and approachable, to maximise opportunities for development, and to be able to give and receive feedback. Tutors’ approaches to their role can either be authoritative or facilitative, the latter producing more effective outcomes. Some examples of authoritative techniques include repetition of tasks, demonstrations, feedback (without reflection) and vicarious learning. Examples of facilitative techniques include hands-on experience, debriefing after incidents, pre-briefing prior to incidents, posing scenarios or role plays, task setting and setting action plans. Thus the evidence highlighted the role of the tutor as being key not just for recruit learning, but also for the development of a professional police officer. Implementation Implementation quality is an important element in any intervention, including police training: “For both successful and unsuccessful initiatives, it is important for the practitioner to know what was done, what was crucial to the intervention and what difficulties might be experienced if it were to be replicated elsewhere” (Johnson, Tilley and Bowers, 2015: 468). This section presents findings on organisational factors, teaching styles within the police environment and trainer and tutor roles. Organisational buy-in Most of the studies which discussed implementation examined the introduction of andragogy or PBL into police training. Charles (2000: viii) argued that implementing a new learning model into police training requires “an organisational transition of epic proportions”, as the transition to an andragogical and learner-oriented model challenges existing police culture and conventional styles of police training. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Several studies identified the ‘paramilitary model’, characterised by hierarchy, obedience to authority, isolation, stress and discipline, as a particular barrier to recruit learning. The influence of police culture was also identified as an impediment to newer, more innovative models of policing. While socialisation was identified as a positive mechanism by some studies, it was also found to have a negative impact in prioritising police culture over academy learning. Recruits are receptive to the environments within which they experience operational police work, and the effect of traditional police socialisation, characterised by obedience to authority, stress and hierarchy, was found by several studies to be contrary to the academic development of the independent, critical thinking and problem-solving recruit. Integrating theory and practice was often not achieved with the training programmes, if operational policing (described as “real police-work”) is privileged over academy work that is seen as “scenarios” and “made up reality” (Chan et al., 2003: 155). Another barrier that can arise from a lack of support from the organisation, or poor implementation, is a lack of adequate resources for recruits. This can include equipment that would aid recruits in their learning, such as access to computers or a library, as well as physical attributes of the learning environment. Facilitation of specific learning activities such as group discussion and debate is aided by the physical environment which must be conducive to such activities. For example, lecture halls make it difficult for classes to have this small group interaction. Poor facilities can contradict any progress that is being made with the introduction of new learning models. Importantly, organisations must avoid the “quantity over quality” policy, whereby students are over recruited, limiting staff numbers and putting pressure on existing resources. Organisational buy-in also covers ensuring that field training is organised and delivered so as to be effective. Field training is critically important in facilitating the integration of theory and practice, as well as exposing recruits to operational police work and socialisation into the culture. The most significant factor in the field training component is a supportive and effective tutor constable or field training officer (FTO). Key to this is continuity, whereby recruits remain with the same tutor for their field training period. This involves having enough tutors available to meet the demand of recruits, and appropriate scheduling. Continual change of tutor and environment is not optimal for recruit development. Field tutors need to be carefully selected, choosing officers who are motivated individuals with adequate expertise, training, and motivation to teach. Six studies indicated how tutors or trainers were selected for their role, with five following some type of formal selection process. Ideally, tutors and trainers should volunteer for the role and have the required skills or show an interest in teaching or alternatively be selected by their superiors on the basis of ability and potential, which may then require further education for the tutors. 4 1 Teaching styles in police settings Essential for the field training phase is management by shift supervisors. Managers must recognise the needs of recruits, and as such be allocated shifts or jobs that are beneficial to their development. The pressure to resist mismanagement lies with the tutor who must “manage the development of their probationers and also deal with the demands of the public, requests for assistance from colleagues on shift and orders from supervisory officers” (Green 2001: 13). It is important that adequate time is allocated for recruits and tutors to engage in reflection or de-briefing, during this crucial phase, despite organisational and operational pressures. Studies reported that consistency of values between academic and field training was essential in order to ensure that learning is correct and retained as intended. This includes avoiding the negative aspects of police culture or recruit socialisation particularly during field training. Significant problems arise when field tutors and trainers fail to acknowledge academy learning, or actively encourage recruits to ‘forget’ what they learnt in the academy as ‘real training’ occurs through operational experience. As one study put it: “For all the rhetoric of ‘professionalism’… the bulk of recruit training was based on the notion that policing is a craft to be learned ‘on the job’... Probationers developed a number of strategies for coping with the challenges of the job and for fitting in at their new environment… This strategy of ‘keep your mouth shut’ or ‘be seen and not heard’ was a demonstration of respect for experience and rank; it was also a survival technique” (Chan et al., 2003: 304). This provides further evidence of the need for field training officers and tutors to be provided with adequate training and support, to ensure that there is cohesion between academic learning and the field training component. Implementation of the field training programme therefore requires support and coordination between all agencies, including the academic institution, the field training provider and all training staff. Andragogical models seemed to demand a different set of skills than traditional teachers, requiring knowledge of methods which put the student at the centre of learning – methods which would often be contrary to the model under which they were trained. Evaluations showed frustration by recruits with instructors attempting to teach using student-centred methods without the proper training or real “buy in” (Lettic, 2016: 95). In general, student-centred models were praised for their effectiveness in teaching recruits and were considered more likely to provide a more satisfying experience. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Trainer/tutor roles The evidence suggested that one of the most important factors for implementation was the need for trained and motivated trainers. Both trainers, during the academy phase, and tutors, during the field training phase, need to be sufficiently trained in order to deliver a new learning model and to understand their role. This is key within andragogy as “the adult education philosophy requires highly skilled and knowledgeable instructors that must be versed in various learning methods, which are often contrary to the quasi-military model under which they were trained” (Charles, 2000: vii). The instructor’s role within both andragogy, and PBL, is notably different to the traditional lecture-based, teacher focused model of police training. The teacher’s role within PBL is to “live in the background… which is a difficult role to assume when the instructor is used to be the focus of the students” (King Stargel, 2010: 140). However, trainers also must develop specific facilitation skills, for example, withholding immediate answers to encourage students to find information themselves, but also being conscious of when to ‘check in’ with students to ensure they are finding and understanding the right answers. And Heslop (2013: 21) stresses the importance of effective partnership between academics and practitioners: “Along with well-qualified academics and tutors, experienced police practitioners play a necessary and large part in all pre-join training programmes including those delivered in academic institutions”. To secure trainer and tutor buy-in, they need to be persuaded about the value of a new approach, and their role within it. Yet, training received by trainers appears to be variable. Some receiving no new training at all; others receiving training varying from eight hours, two days, a week or 140 hours. Training Aims and Training Outcomes The studies offered some information on the outcomes of training programmes, but their stated aims were variable, and outcome were measured in different ways. This limits what we can say. Aims The studies were sometimes unclear about whether they or the programme had defined aims; furthermore, some covered overall aims, and others were more focused, for example on teaching and training methods. Most studies defined training aims only implicitly. Chan et al. (2003: 44) explains the general aim of the Police Recruit Education Programme (PREP) was to produce “reflective practitioners” who were accountable, effective decision makers, and operationally independent. Following their training, officers should have “knowledge of policing, effective communication skills, commitment to ethical standards, respect for individual rights, self-awareness, empathy, and problem solving skills” (Chan et al., 2003: 44), though specific training objectives were not detailed. Similarly, Conti (2011) does not provide aims of training, but rather notes recruits must meet performance requirements for “physical fitness, self-defence, fire-arms, and the like”. He argues the academy experience was when “recruits must perpetually demonstrate that they are worthy of an eventual elevation to the status of police officers” (Conti, 2011: 411), and failure to meet those requirements means the recruit is unfit to become an officer. 4 3 Other studies focused on reporting the development of specific skills as part of general training. For example one study reported that the aim of various parts of the training was to improve communication skills, in the very specific context of report writing, in order to enable recruits to pass a written communication examination. Another assessed a course that aimed to improve tactical and communication skills within the community-policing model. Aims for specific parts of the field training were sometime very general: “producing constables fit for independent patrol status” (Green 2001) or “to reach a level of competence in the workplace to the satisfaction of their tutor constable” (Thorneywork 2004:2). Other studies described field training as “a ‘bridge’ between the academy and police work” (Hundersmarck 2004: 37), or said that field training “bridged the gap between academy training and operational police work through supplemental training and real world experience” (Novakowski 2003). Aims of academies which incorporated a community-oriented approach were said to produce officers capable of developing rapport and building trust with local communities, to have strong presence, and to be decisive and assertive. Where programmes had a narrower focus, their aims were – unsurprisingly – stated more sharply. Charles (2000) examined the implementation of an adult education model designed to help recruits to think creatively and problem-solve. An alternative ‘narrative’ method of teaching, examined by Poradzisz (2004) aimed to teach recruits to deal effectively with people with mental disorders. Outcomes Few of the 33 studies reported on training outcomes, and where they did, some focused on measures of course completion and others on achievement of skills. Chan et al. (2003: 200) state “For most of the cohort, the transition from recruit to constable was successful: less than 10 per cent of the cohort left the police service through resignation or termination”. Green (2001: 112) reported “it can be seen that the constabulary’s minimum standard is being achieved in the majority of cases”. An expected outcome in academies was that the programme would be completed by most recruits and would equip them with the necessary skills and outlook to function as a police officer. A few studies reported on programmes’ success in increasing specific skills or qualities. Not all outcomes were judged to be positive, however. For example, Heslop (2011b: 340) found placements which took training out of the classroom and into the community “had only a limited effect”. Foley (2014: 194) found training produced physically fit recruits, due to two hours of physical education per day at the academy, but recruits felt they actually had “little or no acquisition of skills but knew how to take a direction from authority”. In short, one of the biggest gaps we found in our review was the lack of evaluation of overall training outcomes. The articulation of specific training aims was limited, and measures of outcome were variable and inconsistent. As a result, we have been unable to synthesise training outcomes in any meaningful way. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Conclusions This chapter has reviewed the findings from 33 high-quality studies of police recruit training across the world. Some very consistent messages emerge. The first is that teaching and training methods have to be tailored to adult learning styles – and where recruits are degree holders – teaching methods appropriate to graduates. The principles of andragogy and of problem-based learning need to be followed, for example. A second very consistent message is that the theoretical and practical parts of programmes must be properly integrated if the theoretical elements are not to be jettisoned as soon as recruits arrive in the field and become inculcated with the occupational culture. This is essential but challenging to achieve. Related to this point, several studies stressed the importance of organisational buy-in to the aims of the training programmes, and in particular in ensuring that trainers and mentors in the field provide proper support for recruits while they acquire practical skills. Findings on the studies’ programme aims and outcomes were disappointingly thin. Aims were often poorly articulated and we have been unable to say anything of significance about outcomes. Clearly the PEQF provides an important opportunity, that should be firmly grasped, to evaluate the outcomes of the three new entry routes into the police service in England and Wales. Any such evaluations would need to state fully and carefully, the aims of the training, in terms of the qualities and skills that recruits should acquire. This probably requires some sort of meaningful statement of what ‘good policing’ looks like, and what a ‘good cop’ looks like. And outcomes need to be measured on a range of dimensions including: programme completion rates; measures of recruits’ acquisition of conceptual and practical skills; in-force retention rates; and disciplinary records. This would amount to a major research programme, that has not been attempted to date by any county – but its findings could be of great significance. 4 5 DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Chapter 4 Organisational readiness for the Degree Holder Entry Programme Jyoti Belur with Winifred Agnew-Pauley and Brendan McGinley 4 7 Introduction This chapter presents findings on the organisational readiness of police forces in England and Wales to implement the new Degree Holder Entry Programme (DHEP). We interviewed Learning and Development (L&D) Managers, HR managers or senior managers with responsibility for training. We carried out 17 interviews, sometimes with multiple interviewees, covering 17 police forces and one civilian police organisation in England and Wales. Some interviewees represented two or more forces since they shared training and learning resources. Interviews were conducted either at force headquarters or at the university. The sample included a mix of police officers and civilian staff; almost all regions in England were represented, but no Welsh forces took part. We used a semi-structured interview schedule covering the main areas of research interest. Interviews were taped with the interviewees’ permission and later transcribed. We conducted framework analysis NVIVO software (Ritchie and Spencer 2002). Full details of the research can be obtained from j.belur@ucl.ac.uk. Our findings are presented in four sections. First, we summarise current recruit training in the 17 forces. The second section describes the challenges foreseen by L&D managers in setting up DHEP partnerships with HEIs. The third section examines other concerns about implementing DHEP and the PEQF more generally. A final findings section focusses on the arguments for and against whether recruits should learn as omni-competent or specialist officers. We then offer brief conclusions. Current training arrangements The current recruit training system is the Initial Police Learning and Development Programme (IPLDP), defined by the College of Policing. Though intended to be a standardised curriculum, there is a great deal of variability between forces in duration, delivery and assessment methods. Table 1 summarises how training was organised in the 17 forces (and the civilian security force). The proportion of graduate recruits ranged from 25 to 50 percent in the forces, consistent with the College of Policing’s estimate of 38 percent. Thus, most L&D managers expected that the DHEP would be a major feeder route into policing and thought that it was an attractive option for the police, mainly because of the shorter learning time involved as compared to the apprenticeship route. Training and L&D structures One force had outsourced its learning and development provision to an outside provider. Another group of forces had formed a consortium that contracted with the local university to provide academic inputs. However, most forces provided training in-house either independently or in partnership. Some forces provided all learning under one roof; others had set up separate units responsible for initial recruit training, specialist training (such as firearms, driving, public order etc.) and training for trainers. Senior L&D managers were a mix of police officers and civilian staff. Some forces had very well- resourced L&D departments, but others were struggling with a skeleton staff. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME TABLE 1: CURRENT TRAINING STRUCTURE AND ACCREDITATION FORCE CLASSROOM TRAINING FIELD TRAINING IPS STATUS IN ACCREDITATION ACCREDITING BODY % OF GRADUATE RECRUITS Force A 103 days 10 weeks 8 to 9 months Level 3 Diploma in Policing Skills for Justice 30 Forces B & C 18 weeks - 6 months to 1 year Level 4 Diploma in Policing City & Guilds 35 Force D 15 weeks 15 weeks ~ 30 weeks No formal qualification City & Guilds 23 Forces E, F, G, H 17 weeks 10-12 weeks ~ 30 weeks Level 4 Diploma in Policing Skills for Justice 45 Force I 16 weeks 10 weeks 26 weeks No formal qualification No external body 50 Forces J & K 18 weeks 10 weeks 28 weeks Level 3 Diploma in Policing Skills for Justice 33 Civilian Security Force 17 weeks 320 hours No set period Level 3 Diploma in Policing Skills for Justice Unsure Forces L & M 15 weeks 10 weeks ~ 25 weeks Level 3 Diploma in Policing Skills for Justice 50 Force N 11 weeks 17 weeks ~ 28 weeks Level 3 Diploma in Policing Unsure ~ 45% Force O 15 weeks 15 weeks 30 weeks Level 3 Diploma in Policing Pearsons Unsure Forces P & Q 18 weeks 12 weeks 30 weeks Level 3 Diploma in Policing City & Guilds Unsure 4 9 Role of tutors during field training Officer trainees spend an average of at least 10 weeks in the field with the tutor constable, who is then responsible for assessing them in various areas of competence before signing them off as fit for independent patrol status (IPS). In some forces the role of tutor and supervising officer was clearly demarcated but in others the terms tutor, coach assessor and mentor were used interchangeably. Three forces had external assessors other than tutor constables. Most forces aspired to one- to-one tutoring, but did not always achieve it. Some forces clearly valued tutors in recruit induction, but not everyone shared that view, and some (smaller forces) gave the role low priority. There is little continuous development for trainers and tutors, unless they have personally chosen to qualify themselves to degree or postgraduate level. The difficulty in upskilling currently employed trainers was a theme that will be picked up in the challenges identified by interviewees in implementing the PEQF, and especially the DHEP. Challenges in setting up partnerships with HEIs Despite some initial resistance, L&D staff had accepted as inevitable HEI involvement in recruit police learning post-PEQF. But staff varied in their thinking about the extent to which universities would be involved in the design and delivery. Some did not know how to go about negotiating relationships with HEIs or what they should be negotiating about; others had already established long-standing partnerships delivering foundation or policing degrees prior to the PEQF. The issues raised by interviewees can be divided in to logistical and intellectual (or pedagogic) challenges. Logistical challenges Interviewees expressed some concern about delivering recruit learning in close partnership with HEIs, especially those with limited experience of working with universities. Chapter 2 showed that effective professional education involves integration of theory and practice – implying that recruit learning should be designed and delivered by academics and practitioners together. The partnership requirements need to be incorporated in the initial stages of procurement. While most interviewees were aware that their choices were constrained by the availability of capable and willing local HEIs, only a few realised that the procurement process would impose further constraints. There were four specific areas of concern: i. Bespoke single force arrangements vs. consortium style approach Interviewees – especially those with limited procurement experience - were unsure whether it was better to have bespoke arrangements with individual universities, or to pool resources with a consortium of forces. On balance most interviewees favoured collaborative arrangements with regional learning and development hubs, serviced either by a single university or an academic consortium. A variant was that learning and development provision would come under regional management, but tenders would involve separate lots for each force’s needs. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME ii. One or more HEI providers for the three routes? Another issue of concern was whether it was best to have one provider (single HEI or consortium) or a mix of providers for the three PEQF entry routes. Opinions were divided, with some wanting the simplicity of a single provider and others wanting to hedge their bets with a range of providers. Some favoured consortia with a lead university given responsibility for coordination – if any university would be prepared to take this on. iii. Location of delivery One of the main issues centres on where the academic part of the learning ought to be delivered. Research suggests that learning within academic institutions’ premises has better outcomes than that undertaken in training centres. However there are potential problems about whether the university or the police force ‘owns’ the learning environment, whether standards of police discipline should apply during education supplied on academic locations and, for instance, whether students should wear their uniforms. Some L&D staff talked about the need for recruits to be “embedded in policing”, especially at the outset. One solution might be to bring the academics over to the force venues to deliver their inputs in blocks within an operational environment; however, this deprives students of the immersive experience of being in a university, with access to libraries, lectures, academics and other students. Some interviewees thought that most learning would be delivered in-house, with universities providing a few classroom sessions and some ‘theoretical inputs’ in small doses. This view was driven by concerns about ‘abstraction’ issues and meeting the organisation’s demands for operational deployment as soon as possible. iv. HEI capacity and requirements HEIs face limitations around capacity to service the fluctuating demand for recruit education. There are further issues about timing, as forces currently accept intakes of recruits at several points in a year, whilst academic courses are organised around a fixed academic year. There were associated problems about forces’ co- ordination of their own staff and learning resources, to mesh with three different entry routes. PEQF could thus require a sea-change in the workforce planning and recruitment calendar across all forces. Interviewees faced the added challenge of striking the right balance between the three entry routes, in order to yield the right flow of recruits who had achieved independent patrol status (IPS). Interviewees recognised that forces with long-standing partnerships with universities were best placed to iron out all these logistical issues. Those with pre-existing partnerships had already solved many of the problems, having learnt how universities work; they also understood the time constraints that universities face in processes of course design and accreditation. Some interviewees were drawn to solutions where consortia of forces would contract with HEI consortia. Others thought that complex delivery structures of this sort could be risky as “universities are even worse than the police service when it comes to working together”. 5 1 Intellectual/pedagogic challenges Chapter 2, on the experiences of conversion courses for other professions, identified challenges associated with design, structure and shared delivery of the curriculum, assessment processes and, in particular, effective integration of theory and practice. Our interviews with L&D staff explored these issues. i. Design and shared delivery of curriculum Interviewees recognised the importance of questions about the intellectual lead in setting the DHEP agenda. They appreciated that HEIs would be issuing the degree-level qualification, and would inevitably be “in the driving seat”. At the same time, they did not think that HEIs were in a position to tell recruits “how to be a police officer”. So tensions and squabbles were expected. Generally, interviewees were unclear about managing this process. They were clear that some parts of the curriculum needed to be delivered by practitioners and experienced police officers, but had not fully worked out how other parts could be delivered in combination with theoretical inputs from academics. ii. Assessments Assessment processes were flagged up by interviewees as a potential area of concern. Whilst some forces had well established assessment processes in place, they were unsure about the nature of the three-way relationships between forces, their HEI providers and the College of Policing. There were also concerns about ensuring the functional equivalence not only of the three entry routes but also between the various providers across the country for each entry route. Several interviewees were concerned about quality assurance. They were unclear about the planned or potential role of the College of Policing here, and having an independent organisation that would be responsible for quality assurance and assessment was flagged as on possible solution. iii. Integrating theory with practice L&D leaders were aware of the need to integrate the theoretical and practical components of the recruit learning. Integration is in part a practical question of the division of labour between in-force staff and HEIs, and to some extent is dependent on the structure of the proposed curriculum. However there are also intellectual or pedagogic challenges in designing a curriculum that genuinely makes coherent links between practical skill and academic knowledge. This would involve academics and practitioners working together to deliver a joined-up curriculum while giving space for the learning to set in through opportunities for reflection and peer learning. Few interviewees had made much progress on this difficult issue. This chapter has already discussed the widely varying provision of mentors, and Chapter 6 also considers this. Interviews with PN participants showed that graduate entrants’ confidence and skills were dependent on the level of support provided to them in the force by their tutors and mentors. Some but not all interviewees were aware that DHEP recruits could require or expect tutors who operated at learning level 6 or 7, than the current level 3. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Other challenges anticipated in the implementation of DHEP Whilst interviewees had particular concerns around partnerships with HEIs, they also mentioned a range of other issues: upskilling of staff; impact on recruitment; workforce planning and costs. Upskilling of trainers Intimately related to the issue of integrating theory with practice is the challenge associated with upskilling staff current in the role, along with the tutors and mentors. Challenges centred on issues of: staff and tutor willingness to get additional education to level 6 or 7; the organisation’s ability to afford upskilling the continuous development of in-house staff in terms of time and resources; and knowing the correct level to which to upskill for specific roles. Most interviewees were aware of the need to upskill, but were well aware of the obstacles in their path. Upskilling of police learning staff and tutors was recognised as being important but interviewees were realistic enough to admit that while having a fully qualified and trained L&D cadre was ideal, it would be more of a medium to long term aspiration. Impact on recruitment Whilst some interviewees had misgivings about the principle of graduate entry, rather more expressed concerns about the impact of the PEQF on the profile of recruits from 2020 onwards. Some interviewees thought that PEQF could increase the overall diversity of the workforce, with enhanced professional status serving to attract some ethnic minority groups. Others, however, were concerned that graduate entry would bias recruitment to “the white middle classes”; it would serve as an obstacle to recruitment of people from ethnic minorities, as well as older people with life experience but little interest in getting a degree-level qualification. For this latter group, respondents were concerned that the apprenticeship route might seem especially unattractive. Workforce planning Interviewees identified three sets of workforce planning issues that would have an important bearing on how the graduate entry route would be implemented: •• assessment of recruitment requirement; •• retention rates and •• planned recruitment allocation through the three entry routes. Assessment of recruitment requirement Interviewees expressed concern about assessing force recruitment needs year on year as it had a direct bearing on setting up a working partnership with HEIs. Workforce planning would become especially difficult under PEQF if it meant matching force recruitment cycles with academic term cycles. As discussed above, most forces currently recruit on a rolling basis throughout the year. Interviewees reported a further workforce planning issue, in that operational needs would be affected when L&D staff, tutors and graduates were taken away from operational policing on learning days. 5 3 Retention rates Interviewees were worried about the changing nature of the police service, and potentially growing problems of retention as people no longer commit themselves to a full policing career. However, one of the main concerns about accurate workforce planning revolved around finances which we shall discuss below. Recruitment allocation between three entry routes Interviewees were unclear how they would manage their workforce planning to allocate recruits to each of the three PEQF routes, given the difficulty of forecasting both overall needs and the preferences of recruits – and, of course, the overall budget that would be needed. Some interviewees had started to think about the kinds of degrees their graduate entrants should have. Some thought that recruits with a social science, criminology or law degree would be able to pick up the requisite skills and knowledge much faster than someone with a fine arts degree. Costs Interviewees expressed uncertainty about the funding of the PEQF overall but particularly about how the DHEP route was to be funded. They were keenly aware that funding for the DHEP would define the scope of partnership agreements with universities. A linked issue was the geographical proximity of universities. Those outside of the force area – or distant from force HQs – raised questions about travel costs and time spent travelling. Interviewees were aware that almost all decisions governing contractual agreements with HEIs would be dominated by costs - the status of the HEI, the location of the training, the division of labour between L&D staff and HEIs, assessment loads, timing of recruitment entry routes and number of recruits per year, upskilling of in-house staff and tutors and finally, how costs would be distributed among a consortium of universities providing training. Added to this was the issue of whether the DHEP would cost the same for all graduates regardless of the discipline of their original degree. The cost of a graduate certificate for a recruit with a social science degree might be around £2,500 according to one interviewee, whilst provision for recruits with completely unrelated degrees might be three times higher or more. Overall, cost was such a dominating consideration that there was clearly a risk that forces would have to settle for less than satisfactory arrangements for the DHEP and the other PEQF routes – even if this jeopardised the overall professionalisation agenda. Training for omni-competence versus specialist skills The final set of findings relate to discussions around (a) the current graduate training and development provided by Police Now, which concentrated on developing neighbourhood or community officers, and (b) the MPS plans to introduce direct entry detectives. A majority of our interviewees saw increased specialisation as inevitable, and recognised a need for some forms of specialisation starting at recruitment. For example, they thought it would make sense to support someone with a computer sciences degree to become a cybercrime specialist from the outset. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Nevertheless some interviewees, especially from smaller forces, argued for omni-competence, on the grounds that this maximised flexibility of deployment. Although appreciating the case for training recruits as specialists in principle, interviewees felt that this might face resistance within the organisation. The main arguments against specialism for recruit officers were cultural and operational. Cultural arguments against specialism included the view that if there was a need for cybercrime specialists they could be civilian staff, why would they need to have police powers? The second aspect of the cultural argument was the impact recruit specialisms would have on the existing workforce more generally. For example, respondents speculated that direct entry detectives would be resented by their senior colleagues who had become detectives only after they had worked as ‘beat cops’ for at least five years. Discussion and implications for the DHEP Our interviews show that L&D departments were ranged along a wide spectrum of preparedness to implement the PEQF. Some forces were well aware of developments and actively engaged in the change process through communication with HEIs and the College of Policing. Others remained on the fringes of the process, with little or no awareness of the various challenges that needed to be addressed. We have identified five main areas of challenges in implementing the DHEP. These a summarised graphically in Figure 3. The interviews indicated that these were inextricably interlinked. To meet the challenges involved in implementing the DHEP, support of senior management is needed, coupled with a substantial change in organisational culture. From what we were told, not all chief officers were on board or invested in the PEQF – and organisational preparedness was closely correlated to the extent to which chief officers had engaged with the process. L&D managers themselves accepted the case for the PEQF, though not all were convinced that every entry level officer needed to have a graduate degree, a sentiment they felt was shared by a substantial section of the existing workforce. They felt that the acceptance of graduate professionals into an occupation that is traditionally anti-intellectual is likely to be fraught and will take time. 5 5 Workforce planning Training of trainers Partnership with HEIs Choice of Entry Route Costs 1. Design and delivery of training 2. Integration of theory and practice 3. Diversity of recruitment 4. Development of specialist route } FIGURE 3: COMPLEX INTERPLAY OF FACTORS AFFECTING IMPLEMENTATION OF DHEP DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Implications for PEQF and DHEP This part of our research has five main implications for introducing the DHEP: •• There needs to be a shift in organisational culture to recognise and accept the value of graduate training for recruit officers. The workforce needs to accept the case for the PEQF. For this to happen, police leaders and the College of Policing need to articulate (perhaps again and again) the rationale – and the benefits for the workforce – of professionalisation. Forces’ senior management teams need to support, and have the full support of, their L&D staff in this process. •• The daunting logistical challenges, such as getting contractual arrangements in place with HEIs, upskilling in-house training staff and mentors and bringing in these changes on budget may distract L&D units from finding solutions to the difficult pedagogical challenges in designing a conversion course that equips recruits with complementary practical and intellectual skills. •• Tight budgets increase the risks of making false economies. The cheapest and quickest routes to training officers and bringing them to independent patrol status may not be the best long-term investment, and may undermine the aspirations of professionalisation. •• Given the complexities of setting up partnerships between forces and universities and sharing the design and delivery of training, there is a need for an independent body – with expertise in both the theoretical and practical aspects of police training – to oversee these arrangements and to quality-appraise and assess the various training programmes on a national level. The College of Policing could take on this role, or an independent assessor could be found or set up. •• Finally, there is little evidence to date about the impact that making policing a graduate profession has on people’s experience of their police. The PEQF promises to achieve a radical transformation of police training, and of the nature of policing in England and Wales. Now is the time to put in place rigorous evaluation plans to track the outcomes achieved by the PEQF’s three entry routes, so that recruit training can be closely monitored and refined, as implementation proceeds. 5 7 DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Chapter 5 Motivations and appreciation of evidence based policing amongst Police Now participants 14 Almuth McDowall, David Gamblin, Jennifer Brown 14. Acknowledgement: the chapter authors wish to thank EBW Online (www.ebwonline.com) for making the EI assessments available for research purposes. 5 9 Introduction The Police Now (PN) training programme for graduates offers a unique ‘test bed’ to evaluate how a graduate conversion approach might work for bright, motivated highly pre-selected graduates who are unlike traditional policing training cohorts, both in terms of prior education but also other characteristics. The chapter draws on psychometric data collected from PN participants to assess their motivations for joining the police, their levels of emotional intelligence and their reactions to concepts of evidence based policing practice. A fuller report can be obtained from a.mcdowall@bbk.ac.uk Much of the available evidence suggests that training will only achieve intended outcomes if it addresses trainee needs and motivations: even well-designed training will ‘fall flat’ if trainees don’t like it or don’t engage in it. A second consideration is transfer of learning to the workplace: a training programme could be well received, and even may be seen as highly successful, yet be ineffective if there are obstacles in the organisational environment to applying newly gained knowledge, attitudes and skills. This chapter investigates trainee motivations and transfer of learning over a period of time drawing on and extending Kirkpatrick’s model of training evaluation (1967) to provide useful feedback not only to Police Now as the research host, but also to the College of Policing and the Home Office. The characteristics of the PN trainees We gathered background and demographic information for the 106 participants in Cohort 2 and the 223 in Cohort 3. We also collected data on Emotional Intelligence (EI) for Cohort 3. Across both cohorts, the Metropolitan Police Service (MPS) received most graduates, although Cohort 3 were spread over a wider number of participating forces (19 forces; Cohort 2: 7 forces). The gender balance was even, which is encouraging, given the long- standing difficulties in attracting gender- diverse applications to policing (the current percentage of women officers stands at 29%). The age range increased from Cohort 2 to Cohort 3, but the average age for both groups was 23. The participants therefore compare to typical graduate programme populations age- wise, as well as being highly educated. Cohort 3 indicated a larger proportion achieving 1st class degrees (36%; Cohort 2: 28%), as well as a larger proportion continuing to Master’s level (18%; Cohort 2: 13%). On the other hand, Cohort 2 trainees were more likely to have had work experience (78%; Cohort 3: 65%), policing experience (24%; Cohort 3: 14%), and exposure to the police via a family member or close friend (39%; Cohort 3: 18%). Whilst previous exposure of policing (for example as a special constable or PCSO or from a family where a member is employed by the police) played a role in attracting Cohort 2 applicants to the scheme, Cohort 3 applicants appear to have been attracted for different reasons, potentially due to the increased recruitment campaign by PN. In addition, Cohort 3 came with increased academic achievements, in line with Police Now’s own research data that shows awareness about the programme has grown, and that applications from graduates have doubled between 2016 and 2017, attracting graduates to Cohort 3 who had experience of employment before joining Police Now. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Motivations for Joining We asked the trainees to identify their top three reasons for joining PN. Overall 265 trainees provided at least one reason for joining, as shown in Table 2. The reasons for joining the scheme are predominantly intrinsic – motivations from within, such as interests, curiosity, and values (Deci & Ryan, 1980). Trainees want to make a difference and give back to the community, meanwhile maximising their self-development. Extrinsic reasons such as pay, pension, or promotion opportunities featured far less. We further mapped these reasons onto more specific motivation factors as identified by self-determination theory (SDT). SDT is a broad framework of motivation (Deci & Ryan, 2008), which acknowledges that people are motivated by an interplay of individual and situational factors. As part of this it considers people’s basic psychological needs: a) autonomy (what can they decide?), b) competence (how can they use their skills?), and c) relatedness (relationships with other people). Using this framework, our analysis shows that PN participants were drawn to the programme as they are motivated by an even distribution of intrinsic motivation, so want to join policing because they want to feel they belong, have autonomy and develop their skills (rather than for extrinsic rewards). This is an asset to any policing educational programme, and should be harnessed and reinforces through the learning process. TABLE 2: TRAINEES’ REASONS FOR JOINING: SELF-REPORTED THEMES THEME EXAMPLE(S) C’HORT 3 C’HORT 2 MOTIVATION LOCUS MOTIVATION CLASSIFICATION Giving back: people and places Helping people Making a difference 22% 19% Intrinsic Autonomy, Relatedness Policing specific Join police Respected profession 18% 9% Intrinsic & Extrinsic Relatedness Not a typical job Varied / Exciting job Not an office job Active / Practical 15% 16% Intrinsic Autonomy, Relatedness, Competence Intensity Challenge & pace Early responsibility 14% 6% Intrinsic Autonomy, Competence Scheme characteristics Development & support PN values / innovation 12% 16% Intrinsic Relatedness Competence Benefits Career & promotions Salary / Pension 11% 11% Extrinsic - Graduate specific Wanted a grad scheme Chance to use degree 8% - Intrinsic Autonomy, Competence Communities Community engagement - 9% Intrinsic Relatedness Developmental Opportunities Learn new skills Leadership training - 6% Intrinsic Competence 6 1 Emotional Intelligence Emotional intelligence (EI) is the awareness of one’s own and others’ emotions, arguably an important precursor for procedural justice and policing by consent. We measured levels of EI for cohort 3 using a well validated psychometric questionnaire designed by EBW15 (Emotions and Behaviours at Work), which compares people’s responses against a benchmark group of working professionals. In total 84 PN participants completed the questionnaire to establish a baseline measure of EI. We have measures of: Decisiveness: Willingness to make decisions, the need for control and the level of comfort with decision making responsibility. Motivation: Level of energy, passion, drive and enthusiasm for work, being optimistic and positive, the need for achievement and challenge. Influence: The drive to influence and persuade others, to be heard and have an impact. Adaptability: The desire for, and enjoyment of, variety in the workplace; the capacity to keep an open mind and be flexible with different and creative approaches. Empathy: The ability to recognise, be sensitive to, and consider others’ feelings, needs and perspectives. The need to understand, to help, and work with others. Conscientiousness (sub-scales Structure and Conformity): The need to plan and have structure, be diligent and meet deadlines; the level of comfort with conforming and following the rules. Stress tolerance (sub-scales Resilience and Emotional control): The capability to relax and deal with the day-to-day pressures of work; the level of comfort with showing and managing emotions, e.g. can control/hide temper when provoked. Self Awareness: This scale is an index of the extent to which an individual’s EBW scores is likely to correspond with the way that others would score them on the EBW scales. Table 3 (overleaf) summarises the results. The PN sample has been benchmarked against a comparable group of working professionals (EBW’s so-called ‘norm group’, which included a variety of professions and occupations), whose average scores on each dimension were standardised to a score of 5.5. Shadings in green indicate scores well above the mean, those in orange somewhat above the mean, and those in yellow somewhat below the mean. The table indicates a cohort which is preparing itself for a relatively structured and ‘top down’ training environment, and has not yet had exposure to the need to make operational decisions and take leadership roles. The PN participants are lower than the benchmark group in decisiveness and influence than comparable professionals, yet higher on empathy, emotional control, conscientiousness and structure. 15. See http://www.ebwonline.com DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Reactions to training To investigate trainee expectations for and reactions to their training as commensurate with Kirkpatrick’s model (given that regardless of any discussion about the validity of attitude measures negative evaluations of any training programme make effectiveness less likely), we utilised the Tracker Survey – the first of the bespoke diagnostic tools developed for this study (the items are available on request for anyone wanting to utilise or adapt these). The tracker survey comprised a 30 item questionnaire, capturing trainee perceptions of: Training Environment; Received Training; Information & Decisions; and Understanding the Operational Role. Full details of the tracker survey’s rationale and construction are available from the first author of the chapter on request (further publications are pending). We designed it to be administered across three time points: 1. The start of the Summer Academy – capturing trainees’ expectations for training. 2. The End of the Summer Academy – roughly six weeks following the first time point, capturing reactions to formal training. 3. The 100 Day Impact Conference – several months later, capturing assessments of training once they had had a chance to put the training into practice. The survey was piloted across three time points in 2016/17 (Cohort 2), but full analysis was only possible at the level of individual survey items as there were uneven response rates across the time points. We undertook an Exploratory Factor Analysis17 to help define a simple structure and refine the questionnaire items. This process was repeated in 2017 for Cohort 3, when we also ran Confirmatory Factor Analysis (CFA). CFA results indicated a good fit for the proposed ten-factor model, supporting a refined 30 item questionnaire, with the underlying (latent) variables presented in Table 4. TABLE 3 : AVERAGE EI SCORES PER SCALE, BENCHMARKED AGAINST A COMPARISON GROUP OF PROFESSIONALS DECISIVENESS MOTIVATION INFLUENCE EMPATHY STRUCTURE CONFORMITY 5.03 5.35 4.53 6.58 6.14 6.99 CONSCIENTIOUS- NESS ADAPTABILITY SELF-AWARE EMOTIONAL CONTROL RESILIENCE STRESS TOLERANCE 6.62 5.68 5.18 6.53 5.85 6.31 17. Exploratory Factor Analysis is a statistical technique for identifying underlying patterns in complex datasets, (one of a family of data reduction techniques). 6 3 TABLE 4: FACTORS AND EXAMPLE ITEMS FROM THE TRACKER SURVEY QUESTIONNAIRE FACTOR EXAMPLE ITEM Classroom Training My classroom training will teach me how to locate different kinds of evidence Organisational Citizenship I am prepared to go the extra mile to help my force Loyalty I have a sense of loyalty to the ward I am currently policing Work Load When analysing a lot of information, I am: Creativity When solving problems that have no obvious correct answer, I am: Autonomy When using my initiative or judgment in carrying out my work, I am: Understanding the Role It is clear to me what the goals of being a DWO are Dealing with the Public My classroom training will give me confidence to deal with the public Evidence Based Policing I do not understand the principles of evidence based policing Transfer Climate My force rewards employees for using newly gained knowledge and skills DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Expectations at the start of the Summer Academy Responses from Cohort 2 (n = 106) and Cohort 3 (n = 223) participants were received, grouped and averaged for the Factors outlined in the previous section (Table 4), with the trends reported in Figure 4. Cohort 3 trainees had significantly higher expectations than Cohort 2 for the level of training they would receive (Training Environment) which could be due to a number of factors including increased awareness of the programme itself, and that it would prepare them for Dealing with the Public. In addition, Cohort 3 trainees scored themselves higher on the Creativity self-measure, indicating that they felt more effective at solving problems. Changes over Time: 1, 2 and 3 For Cohort 2, we collected an initial baseline measure at the first time point from 106 trainees. The trainees then provided follow-up responses at the second time point (n = 29) and third time point (n = 41). The low samples mean that results from this cohort should be treated with caution, and are therefore presented as trends (Figure 5) for information, rather than statistical analysis. We managed to collect fuller data for Cohort 3: an initial 223 responses, followed by 196 and 140 responses at the second and third time points respectively. The superior response rates for this cohort allowed more detailed statistical work on the data (see Figure 6 for average scores). Our analysis shows that there were statistically significant changes across the three time points in: Classroom Training; Loyalty; Understanding the Role; Autonomy; Creativity; Dealing with the Public; and Transfer Climate. In contrast, Organisational Citizenship, Work Load, and EBP all remained stable across the three time points. Evaluations of training and learning content (Classroom Training, Understanding the Role, and Dealing with the Public) appear to decrease over time in each cohort. Some possible explanations for such a pattern might include that initial trainee expectations are (perhaps unrealistically high) at the outset due to the intensive nature of the training, or that the forces have different expectations for the training and preparation that their trainees will have received. Another potential reason is that the roles for which trainees are being prepared for vary across forces, for example, working in response units rather than neighbourhood policing, and that job tasks may be more mundane or repetitive than envisaged. On a more positive note, the highest scoring Factor, Organisational Citizenship, remained stable throughout training and into the role. The Loyalty Factor also remained stable from the start of the Summer Academy to the end, one month later. More promisingly, trainees’ self-reported Loyalty then increased once they had spent some time in their respective forces (time point 3). 6 5 FIGURE 4: FACTORS SCORES FOR COHORT 2 AND 3 TRAINEES AT OUTSET – BASELINE EXPECTATIONS FOR TRAINING FIGURE 5: COHORT 2 TRAINEE FACTOR SCORES OVER TIME. FIGURE 6: COHORT 3 TRAINEE FACTOR SCORES OVER TIME. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Learning on Evidence Based Policing Arguably, one of the most important things that entry-level training and qualifications need to achieve for any profession is to give entrants a good sense of the knowedge base that underpins their professional practice; such learning is the second level in Kirkpatrick’s model. We therefore attach considerable importance to this section of our findings. The EBP Toolkit (the Toolkit) is a multi-method instrument which combines elements of reflective learning and data collection. The Toolkit was constructed with input from serving officers, recruited as “co-researchers”, to ensure that the content was grounded in operational concerns. The Toolkit was administered online to PN trainees, as a pilot in 2016 and then again in 2017. It first assesses learning by asking participants to define EBP, and then to put the knowledge into practice and answer questions relating to a typical scenario which a neighbourhood police officer might encounter (see Appendix). Trainee Definitions of EBP We benchmarked trainee responses against definitions provided by experts. The first seven key concepts followed Telep and Sommers’ (2017) coding framework, to which were added an additional two concepts resulting from consultations with subject matter experts and the definition provided by the College of Policing.18 The final nine core concepts are summarised in Figure 7. In total, 82 PN trainees provided their own definition of EBP, 58 of whom did so before attending an EBP Skills Session, and 24 did so after attending the session. Responses were coded for mentions of the nine key concepts: trainee definitions were given a 1 for each mention of a key concept, and 0 if the concept was missing from the definitions. Table 5 details the percentages of definitions which made reference to each concept. 18. http://whatworks.college.police.uk/About/Pages/What-is-EBP.aspx 6 7 FIGURE 7: KEY CONCEPTS EXPECTED IN DEFINITIONS OF EBPi 1. What works: EBP emphasises interventions that work. This concept includes effectiveness, problem solving, and best practice. 2. Focus: Researchers highlight the benefits of focused interventions and specific problem analysis. 3. Evaluation: Definitions should acknowledge that interventions require ongoing evaluation and analysis. 4. Research: EBP interventions and strategies should be based on research/empirical evidence and studies. 5. Scientific: Building on the above, strategies should be informed by scientific research, rather than case-based, or less formal before-after type comparisons. 6. Preventative: Academic definitions of EBP make the distinction between reactive and proactive strategies, with proactivity regarded as more effective than response-only strategies. 7. Statistics: Definitions should also emphasise the utilisation of data and statistics, in contrast to anecdotal evidence. 8. Resources: EBP interventions should be efficient, and make best use of police resources. 9. Understanding: EBP should strive to understand the underlying causes of the issue. TABLE 5: PERCENTAGE OF TRAINEE DEFINITIONS WHICH INCLUDED EACH OF THE NINE KEY CONCEPTS – COMPARING TRAINEES WHO HAVE AND HAVE NOT ATTENDED AN EBP SKILLS SESSION. KEY CONCEPT % OF DEFINITIONS CONTAINING THE KEY CONCEPT Before Skills Session After Skills Session Total (n = 58) (n = 24) (n = 82) What Works 62% 63% 62% Research 38% 38% 38% Evaluation a 19% 46% 27% Statistics 28% 25% 27% Scientific 19% 25% 21% Resources 16% 17% 16% Focus 7% 8% 7% Preventative a 2% 21% 7% Understanding 7% 8% 7% a X2 , p < .05 i. Note: Concepts 1-7 follow Telep & Sommers’ (2017) framework; Concepts 8-9 follow consultation with subject matter experts. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME By far the most common concept in the trainees’ definitions was that of What Works, which featured in over 60% of answers. The concepts of Research and Statistics appeared relatively frequently. The Evaluation concept was also mentioned in over a quarter of definitions. Perhaps unexpectedly, considering the emphasis on focused interventions and the policing of ‘hot-spots’, in both the literature and training, the Focus concept featured relatively infrequently. We were also interested in whether the EBP Skills Session had an effect on trainee’s definitions. Trainees who provided their definitions before attending a Skills Session mentioned an average of 1.97 key concepts; Trainees who provided definitions after attending a Skills Session mentioned an average of 2.50 key concepts. This represented a trend towards providing a more complete definition, although this fell slightly below statistical significance. This lack of significance indicated that the frequency of many core concepts (e.g. What Works, Research) did not change between the two groups of trainees. However, on a concept level, Evaluation and Preventative both featured more prevalently in the answers from trainees who had attended a Skills Session. Overall, some tentative positives can be drawn from findings. Firstly, the prevalence of key concepts across both trainee groups appeared higher than for a cohort of US Police Officers (Telep & Sommers, 2017), and are more in line with more senior practitioners such as those noted in Yesberg & Dawson (2017). As this applies to both of the trainee groups it suggests that the training that they receive during the classroom lessons had a positive effect in transmitting some definitional knowledge. Secondly, the frequency of two of the core concepts (Evaluation and Preventative) in definitions was higher amongst trainees who had attended an EBP Skills Session, again suggesting a positive effect on knowledge transmission. Despite this, even in the group of trainees who had attended the additional Skills Session, an average acknowledgement of 2.50 core concepts out of a possible 9 is low. In addition, the percentage of definitions featuring the core concepts of Focus and Understanding were both low (7%). This is in spite of both concepts receiving attention during the Skills Session itself – and there was no uplift in their relative frequencies amongst the post-training group. Nonetheless, coding in such an objective way may miss ‘implied’ meaning from the trainee definitions – one may know a definition without being able to effectively articulate it. Furthermore, knowing the definition of EBP does not necessarily predict effective application. Therefore, in the next section of the Toolkit, the trainees were tested with a typical scenario in order to examine their problem-solving and critical thinking skills. 6 9 Toolkit – The Critical Thinking Scenario Whilst knowing the definitions gives a good indicator of knowledge transmission, it does not necessarily predict effective application (learning facilitation; Kember & Gow, 1994). Therefore, follow up testing of the trainees with a typical scenario was undertaken in order to examine their problem-solving and critical thinking skills (or “deep smarts”, Leonard & Swap, 2004) to give an indication of how learning might transfer into actual behaviours in a policing context (level 3 in Kirkpatrick’s model). The abbreviated scenario is presented in Box 1. Consistent with the pilot data from 2016, the 2017 trainees identified Dealing with the Boys’ Behaviour and Diffusing Community Tension as two of the most prevalent priorities. Both Cohorts (2016 and 2017) also mentioned that they would involve other agencies, as well as try to improve confidence in the police. There were also key differences between the two Cohorts. Firstly there was the emergence of a new theme related to protecting the V family. This theme was present in over half of the trainees’ priorities, and indicated an awareness of the family’s vulnerabilities (elderly mother, minors involved) as well as the risk of them being targeted by the community. A second new theme – Gaining Greater Understanding – illustrates that the 2017 cohort were more likely to attempt to identify the underlying causes of the problem, or pose alternate hypotheses (e.g. could be cultural misunderstanding, potential racism). BOX 1: SAMPLE EBP SCENARIO STRUCTURE On your ward, you have regular contact with the parents of the teenage boys x and y who live in local council accommodation. The neighbourhood is very mixed and ethically diverse. Recently there have been complaints from…… and there are community tensions. You recently attended a community meeting where…… Your sergeant is now asking you to resolve this situation. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Pulse Check Survey As an additional measure of learning from the training as all as an indication of behaviour on the job, we distributed an additional survey amongst PN trainees. This survey was completed at the end of the 2017 EBP Skills Sessions (held in Manchester & London), and asked trainees to answer the following three open-ended questions: 1. Thinking about everything you’ve learned, what stands out about Evidence Based Policing? (89 responses) 2. What are your main takeaways from today’s Skills Session? What will help you in practice? (92 responses) 3. In general, which aspects of the academy and beyond helped you most to do a good job? (82 responses). Responses were collected electronically and analysed into key themes, as presented in Table 7. The Focus theme emerged as the most prevalent concept for EBP. There were also general positive comments in 16% of responses that were too broad to classify, such as that EBP was “worthwhile”, “sensible”, and “beneficial”. However, there were also negative comments in 20% of responses, including critiques. The main issues identified were: that EBP was “common sense anyway”, that it was too demanding on time and resources, and that it was more for academics than for officers on the ground. This kind of characterisation of EBP is perhaps a reflection of its absence from the discussion and practice of front line police. Future research would do well to further entangle how policing culture influences such views; we also refer the reader to Chapter 6. TABLE 6: PRIORITIES SELECTED BY TRAINEES, BEFORE AND AFTER EBP SKILLS TRAINING. PRIORITY % OF ANSWERS CONTAINING EACH PRIORITY Before Skills After Skills Total Experts Students Session Session Cohort 3 trainees (n = 31) (n = 12) (n = 43) (n = 9) (n = 25) Deal with the Boys’ Behaviour 52% 58% 53% 44% 36% Protect the V Familyb 55% 50% 53% 11% 12% Diffuse Community Tensionb 39% 67% 47% 11% 12% Consultation & Communication 32% 33% 33% 56% 40% Gain Greater Understandinga, b 13% 42% 21% 56% 8% Resolve the Housing Issue 16% 25% 19% 11% 16% Improve Confidence in the Policec 16% 17% 16% 0% 0% Involve Other Agencies 16% 7% 16% 33% 4% Involve a Translator 10% 17% 12% 11% 4% a X2 Differences between before vs after training groups (p < .05) b X2 Differences between trainee, expert, and undergraduate groups (p < .05) NB: whilst we recognise that Ns are low in the expert group, this is in line with typical expert consultations c X2 Differences between trainee, expert, and undergraduate groups (p = .052) 7 1 TABLE 7: RESPONSES TO PULSE CHECK SURVEY Theme Example % Q1 – KEY THINGS LEARNT ABOUT EBP Focus / Specification “Define your problem – and make it much more specific” “How important it is to narrow down and hone in on key issues” 17% Research / Data “It’s a scientific academic approach to making society safer” “The ability to use the database to research previous police based activities/studies” 15% What Works ““It allows you to know what tactics work and what don’t” “…trying methods that have worked elsewhere first” 13% Before acting / Proactive “Proactive can be much better than reactive” “Stepping back from an issue and properly considering it rather than rushing to a solution” 11% Learn from past (mistakes) “Being able to use experience and other existing knowledge to implement strategies for improvement” “Failure is evidence in itself” 10% Understanding / Causes “That you need to understand the issue before tackling it” “Make sure you analyse your problem” 8% SARA model “How to use the SARA model” “(I) understand the SARA model a lot better from the day” 6% Innovative “Don’t take the easy option, try something different” “Creative and innovative” 4% DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME TABLE 7: RESPONSES TO PULSE CHECK SURVEY Theme Example % Q2 – MAIN THINGS LEARNT TODAY Focus / Specification “It is important to fully define the problem that needs solving and I need to be as specific as possible” 30% Research / Data “Make time to get to the baseline data and current crime level before I do anything” 16% Finding evidence “Studies and websites where you can find help” 14% Understanding / Causes “Think about the mechanisms to a problem” 12% Models / SARA “Use SARA, rather than thinking it obvious ‘I'll just do this’” 8% Evaluation “To check if implementation has worked” 8% What Works “Ensuring to check what has worked or not before” 5% Q3 – WHAT HELPED YOU DO A GOOD JOB? Force immersion / Field training “The immersion period back in force” 16% Experts / Guests “The advice and interaction with the guest speakers” 16% Colleagues / Leads “The belief from my syndicate lead that I was capable gave me belief in myself” 15% Personal “The academy allowed me to become more confident when speaking in public” 15% Practical / Role play “Role-playing stop and searches, booking into custody… were also really useful” 12% Law / Legislation / Theory “Understanding/being taught my powers and different ways to use them” 12% EBP / Problem solving “The encouragement to think critically and ask “why” 9% 7 3 Impact presentations Following the Summer Academy training event, the trainees join their designated force to start their operational role. Several months later, the trainees are asked to report on their progress by presenting a problem which they have been working on over the last 100 days. We observed several of the trainees’ presentations and coded for the presence of key EBP concepts as outlined above in Figure 7. We randomly selected 57 presentations for analysis (all occurring after EBP training). These covered a broad range of topics covered with anti-social behaviour and drugs problems being the most frequent. Problems were highly localised and were often triggered by residents or business people’s complaints. Hot spot analyses and force wide statistic also drew attention to particular problem areas. Use of SARA and the problem analysis triangle to focus the problem was noticeable as were conscious attempts to consult the community and engage with multi-agency partners. Many of the problems were focused by default as they were problems caused by one offender or group of offenders, and often limited to single premises or street. There was a creditable awareness of the need for sustainable solutions. This was often achieved by working in partnership with local authorities or voluntary sector organisations. The main weakness of the presentations was a general absence of use of existing research to design their interventions or relate their findings back to the original research, their use of a basic before-and-after design and the total absence of any tests of statistical significance for any quantitative data reported; whether this is due to a need for more training, a lack of time or opportunity or both remains to be investigated further. As they stand, the outputs conform to what Anderson, Herriot and Hodkinson (2001) call the ‘transfer of learning’ model of evidence use. Studies of this type have high practice relevance but lack scientific rigour. This means that conclusions drawn from the data tended to be overstated, and the contribution that the presentations could make to building a corpus of reliable knowledge was minimal. Helping and Hindering EBP Application We investigated trainee perceptions of factors which help or hinder the use of EBP on the job, firstly as a pilot with Cohort 2 (2016) and then again with Cohort 3 (2017). The trainees were asked their views about what helped or hindered the application of evidence-based policing during their working day. These factors included the trainees’ own attitudes towards EBP effectiveness, the attitudes of those around them, and intrinsic factors such as time and resources to appropriately conduct EBP. For each of the ten factors, a Net Helper Score was calculated by subtracting the number of trainees (total n = 45) who felt that the factor was a hindrance from the number of trainees who felt that the factor was a help. Therefore, positive scores indicate that the factor helps application of EBP, whilst a negative score indicates that the factor hinders application. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME As shown in Figure 8, the biggest helpers for applying EBP in the workplace were the trainees’ own attitudes, and their beliefs that EBP improves what police officers do. The attitude of the trainees’ sergeants was also identified as a helper towards applying EBP by some but not all participants, perhaps reflecting variations in the sergeants’ own attitudes. The three largest hindrances to using EBP were not knowing where to find evidence, not understanding what is meant by evidence, and the demands of day-to-day duties. There also appeared to be a positive shift for the majority of the factors from 2016 to 2017. Firstly, although three hindrances were identified, these were not reported to be as problematic. For example, in 2016, 50% of trainees reported that day to day demands “hindered a great deal”, compared to a reduced 20% in 2017. Similarly, peer attitudes were seen as an overall hindrance in 2016, whereas in the current year, an increased percentage of trainees found that peer attitudes had helped them, resulting in the overall neutral net helper score above. However, the largest improvement was for trainee beliefs that EBP was necessarily the best way to help their communities. In 2016, this factor was perceived to be a hindrance to the application of EBP, whilst in 2017 this was perceived to help. FIGURE 8: COHORT 3’S NET HELPER SCORES FOR THE APPLICATION OF EBP IN THE WORKPLACE. My own attitude about policing using EBP My belief that EPB improves what police officers do My beliefs about whether EPB is the best way to solve work issues Attitude of my sergeant My beliefs about whether EPB is necessarily the best way to help my community My beliefs about whether EPB is only one way to clarify the problems I have to deal with Attitude of my peers Not understanding what is meant by evidence required The reality of the demands of my day to day duties Not knowing where to find evidence -20 -10 0 10 20 30 7 5 Conclusion The key learning points from the analysis in this chapter are as follows. First, we offer an evaluation framework for understanding learner motivations, personal characteristics and transfer of learning into practice by drawing on Kirkpatrick’s model (1967, 2018) of training evaluation. Our data shows that trainees are highly motivated and receptive which offers a significant opportunity and resource for forces taking in graduate trainees, provided receptivity and support in the work environment. Emotional intelligence may offer a useful framework for assessing future officers’ capacity to engage in procedurally just policing principles and which aspects may require developing through education and continuous development on the job. The data shows that more targeted inculcation of EBP reaps rewards by increasing knowledge of basic principles and increased likelihood of a more focused approach to addressing real-life policing scenarios. The latter are habitually highly localised and contextualised. Efforts were made to review in-force data and consult the affected community to define and limit the problem but these would benefit from more systematic approaches such as the problem analysis triangle. Enabling trainees to frame stronger research designs together with a greater appreciation of some basic tests of statistical significance, or appropriate qualitative approaches, would allow them to draw stronger conclusions from their data. There was little evidence to suggest that having undertaken some more systematic problem solving these ideas were taken up elsewhere within force or the learning exploited for trialling elsewhere. This finding may be beyond the scope of PN itself but raises the broader question of embedding EBP in practice through a better understanding of how this is enacted, or not, by local officers once they have been trained to instigate a learning culture. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Chapter 6 View from the shop-floor: perspectives on the Police Now programme Tiggey May and Gillian Hunter 19. The Metropolitan Police Service was the only police force to participate in Cohort 1. Thereafter, the PN programme extended to include 7 forces in 2016 and 19 forces in 2017. 7 7 Introduction Using Police Now as a test bed, we have explored the experiences of 45 graduate entrants to policing. We describe how they see the strengths and weaknesses of the programme, how they have coped with the learning and their neighbourhood placements and how they understand, value and apply research evidence to their practice. Our findings complement the findings from Chapter 5. We use our findings to draw out key learning points to help inform the development of the DHEP, focusing in particular on the integration of the learning and practice elements of a graduate programme. We address four questions: •• How do graduates experience the short intensive policing course? •• How do graduates learn the essential craft of policing? •• How do graduate police officers apply their critical skills, gleaned from both their university study and the summer academy, to the demands of their neighbourhood police work? •• And what is the role of evidence based policing in their work? Methods The research was conducted in 2016/17. The main components were: •• 60 in-depth interviews with 45 PN participants, (30 from Cohort 2 and 15 from Cohort 3) •• Seven observations of PN training days and 100-day impact events; •• 13 interviews with police line managers of graduates; •• Eight interviews with PN staff; •• Two focus groups to examine perceptions and usefulness of training on evidence-based policing; •• And content review of 30 PN Graduate 100-day impact presentations. Cohort 2 interviewees comprised 15 female and 15 male graduates who were interviewed six months after the Summer Academy 2016 (n=30) and again at one year (n=15). They were aged between 21 and 32. Seven were from Black, Asian and Minority ethnic (BAME) groups; the remaining 23 were White British. They came from a range of academic disciplines, including: law, criminology, the natural, social and political sciences, music, foreign languages, engineering, war studies, history and geography. Six graduates had relatives currently or previously in police employment and five had previous policing experience themselves. Fourteen were subsequently employed by the MPS,19 six by West Midlands Police, three by Lancashire, two respectively by Cheshire, Surrey and Thames Valley Police and one by Northamptonshire Police. Cohort 3 interviewees comprised six female and nine male graduates who were interviewed between six and 12 weeks after leaving the Summer Academy in 2017. They were aged between 21 and 26; 13 said they were White British, one was Asian, and one did not say. Like Cohort 2 they had degrees in a range of subjects including: law, geography, psychology, chemistry, politics, sports science and war studies. Five had family who were currently or had previously been in the police service. A fuller report can be obtained from t.may@bbk.ac.uk DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Findings on experiences of the Intensive Learning There are obvious expectations of graduates and the skills they bring to PN; their degrees imply ability for higher-level learning and critical thinking. They undertake a shorter six-week introductory course to policing in contrast to the average of twelve weeks for recruits coming in via the standard route and are expected to undertake some of that learning independently. PN graduates are required to complete an on-line, pre-learning course, the content of which is predominantly on the law and on officers’ legal powers. On the first day of the Summer Academy the graduates sit an exam, which they must pass to progress onto the six-week academy. Most of Cohort 2 and 3 passed the exam at the first attempt. None of the interviewees reported the pre-learn to be too difficult and all were happy to cover these basics independently - most had recently completed their final university exams, so were used to pressurised learning and exam conditions – and viewed the pre-learn as a good foundation for entering the policing profession and starting the Summer Academy. The few who failed the exam blamed too short a period of time between finishing their university exams and starting the Academy. As far as I can remember the majority of the learning was legislation. There were also modules about values and ethics, I think there was one about styles of neighbourhood policing. The vast majority was factual stuff that you could just learn by rote almost. Anything that was a bit more nuanced was stuff that we got taught in the classroom, as it were, when we got to the academy. [PNG 26] The pre-learn was primarily preparation, for the exams. It was helpful, it meant when we went in we weren’t completely dithering like, “What am I doing?” I thought it was good, it was reasonable. It certainly did what it was meant to do. [PNG 19] I didn’t do much prep work due to going away, exams and finishing essays, I failed the first exam, but passed the re-take 48 hours later [PNG 10] The Summer Academy comprised three distinct but complementary training strands: lectures and classroom-based activities; officer safety training; and field training. Lectures are delivered by both guest speakers and serving police officers and cover topics including: stop and search; roads and breath testing; values, legitimacy and ethics; female genital mutilation; fraud and cybercrime; humility, toxicity and communication; evidence-based policing; mental health; sex offences and domestic abuse. Lecture length varied from one to four hours. Graduates were placed in a syndicate for the duration of the Academy. A syndicate was comprised of a lead, who was an experienced operational police officer, and 10-12 graduates. Syndicates sat as a group for lectures, worked through exercises in the classroom, and completed their field training together. In late 2017, MOPAC published their final evaluation report detailing the experiences of the first PN Cohort (Yesberg and Dawson, 2017). The views of Cohort 1 officers on their experience of the Summer Academy were similar to those of our Cohort 2 and 3 interviewees. Like Cohort 1, our interviewees spoke positively about the training they received at the Summer Academy, commenting positively on the fact that 7 9 most of the speakers were ‘serving’ police officers: The session organisation was great, I thought the depth was quite good, they got across the points they needed to and they brought in people who were still doing it… current serving officers, serving sergeants, serving inspectors, people who are top of their field who know everything that is current. That was the most important thing because it showed that what we were learning is the best working practice used by people who are still practising and working [PNG 05]. However, like Cohort 1, many of the interviewees were dissatisfied with the craft/knowledge-based balance of the intensive learning. A key theme in our interviews was the importance for the graduates of learning the craft of policing. They valued the sessions on practical or craft skills more than those which focused on the so called ‘soft skills’20 or theories of policing, often worrying that there was too much emphasis on the latter. (This is in contrast to the findings of this report, and may reflect more the lack of craft on the street and lack of credibility that formal knowledge has on the front line.) Importantly, our first interviews with graduates, and therefore their reflection on the intensive learning, was done after they had been in force for several months. Their views about their own competency when starting in force were generally low (which is not unique to PN participants), in particular they stressed their lack of confidence in policing skills (see also Chapter 5; and Yesberg & Dawson 2017): They put a lot of emphasis on the soft skills, as it were. We had a four-hour lesson, I think, on resilience which I didn’t really take a whole lot away from. Compare that to, say, a one-hour lesson on sexual offences, not that we’re going to be dealing with a whole lot of that in our capacity as neighbourhood officers. I felt it was sometimes lacking in terms of giving us a rounded view. [PNG 32] In terms of the content, there was a lot of debate amongst us that there wasn’t enough time being spent on particular things, some things were being sacrificed for other things. So I think our input on traffic law was probably two or three hours, whereas we would talk about resilience and soft skills for days on end. And that would be a constant thing, but when you start the job, you think back and wish that the practical aspects and stuff such as traffic law and stop and search, you wish that had been covered more…I think they really pushed soft skills, you know, reflection, the management toolkit they call a ‘resilience toolkit’, they really pushed that...[PNG 28] Like Cohort 1 (Yesberg and Dawson, 2017), most interviewees found the six- week course intensive and exhausting, some noting that the days were too long, and the balance of learning biased towards lecture-style teaching rather than discussion and field training which they largely preferred: The days were so long; sometimes you were in one room for five hours with hardly any breaks. Your brain just can’t concentrate for that long. A lot of stuff you don’t even remember that you had a talk about it because there were just 20. Officers tended to refer to sessions such as legitimacy, procedural justice, empathy, resilience as ‘soft skills’ sessions – presumably reflecting their focus on interpersonal skills. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME so many talks, so many lessons and so much sitting and listening that it just all kind of merges into one sometimes. We didn’t get much reflection time at the Summer Academy. [PNG 09] The field training was very good, which is the bit where you actually go out, in [police] cars. You learn things that you just can’t learn in a classroom. The lectures were good, predominantly good, but long. The days were long and for my money eight weeks would be a more appropriate time just for making everything easier... It was a bit like death by PowerPoint for quite a lot of it to be honest with you. [PNG 16] I would have rather had more classroom-based discussions, I thought the hall was really good as that’s when you’re getting the experts to come and talk to you, that’s where you’re getting the knowledge being passed to you. But for me, I think it’s partly the way that I learn, it’s not really hammered home until I get into a group and just... talk about it. So what I would have liked is to have slightly shorter hall sessions and slightly more classroom sessions. [PNG 18] Chapters 2 and 3 examined the relationship between theory and its application to practice and how best to integrate the two in conversion courses. Interviews with Higher Education Institution (HEI) professionals running courses for teaching, social work and law suggest that courses which neglect this relationship ultimately limit a student’s professional development and in the long term compromise their professional trajectories (See Appendix A). Some of our interviewees showed limited understanding about the rationale for including each element during the Summer Academy and how the theoretical learning informs policing practice; and they frequently expressed a preference for more opportunities to discuss issues raised in lectures. Different learning styles – as well as the different perspectives on whether participants had enough time to reflect and to acquire craft skills – must be addressed through the collaborations of HEIs with police forces/Police Now when preparing for the DHEP. From graduate to police officer – experiences within the home force After completing intensive learning, officers are deployed to their home force to start a 28-day immersion. Each force is expected to organise a force-specific training input (usually one week in length), to acclimatise graduates to local procedures, policies and ICT systems. For the remaining three weeks, graduates are expected to accompany a tutor/mentor to continue their professional development, including practical skills, and to learn about their local area and the priorities of their team. This transition is of particular interest in order to examine the coordination of learning and practice between PN and the host forces. We sought feedback about: •• The information provided to local neighbourhood teams from PN and Force Learning and Development Departments; •• The arrangements put in place to ensure a smooth transition from the SA to neighbourhood policing teams; and •• The support provided by the host force, including the reception from police colleagues. 8 1 Interviewees from Cohort 2 and 3 had mixed experiences of transition. While most reported a relatively smooth transition, in that their new colleagues were aware they were arriving in force, others described their experience as poorly managed in force and their team and line managers as ill-informed about PN. Our interviews with police line managers supported this view, in that several complained of having very little knowledge or information about the graduate scheme or understanding about the skill-set their new colleague would arrive with: Yes, my team were aware, which is always a good thing; in fact most of the station was aware that a Police Now graduate was arriving, my colleagues were aware of what Police Now is but I think we might have been miss-sold to them slightly. They were expecting this all-singing all-dancing, ready-to-go, problem-solving, impressive, best human there is in the whole planet. We arrived, and we literally couldn’t do anything in terms of policing, which is quite an essential part of the job; but I guess that’s something you learn through experience and actually doing it. We weren’t the all-singing all- dancing problem-solving extraordinaires they were expecting. [PNG 17] Yes my team were aware. I was really lucky because I had had a previous conversation with the Inspector who organised my arrival, so that really helped. I got to know my Sergeant and the officer that would be tutoring me as well. It was all really, really helpful. I said to my own HR team here that, “I think it would be really good if future officers were given that experience as well.” [PNG25] Some of my team knew I was arriving some didn���t, same with the senior management, some knew, some have only just come to terms with it, literally. [PNG 27] I don’t think anyone was aware I was arriving… My new sergeant is very anti Police Now. [PNG19] PN has made various efforts to improve communication with the host forces, for example by inviting line managers to the Summer Academy and introducing Leadership Development Officers to act as a point of contact between graduate and force and to troubleshoot problems. Police forces need to think about how to draw in the wider support systems for graduate recruits when the DHEP is operational. Mentors Graduates are expected to be supported in neighbourhood teams in the first instance by a trained mentor/tutor selected and provided by their force. As Chapter 2 discussed, mentoring is an important element of graduate conversion courses. Unfortunately, not all graduates were assigned a mentor, and some reported only limited access to one; for example, available only on some shifts or for a limited number of shifts. It was also unclear what training – if any – forces provide to mentors to undertake this important role. Line manager interviewees also raised this issue, noting that they were restricted by who they could ask to mentor the PN graduates because of limited staff availability and resources. Most mentors were neighbourhood colleagues. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Our interviewees’ descriptions of mentoring varied. Some reported very positive experiences, often highlighting the support, pastoral care and neighbourhood knowledge which their mentor possessed. Others were less enthusiastic, complaining about their mentor’s lack of interest in PN or understanding of their supporting role: My mentor was brilliant, she was really good. She just showed me the ropes of what we were doing. Youth engagement is very different from beat management so she took me out to the schools and she showed me how we operated, which was good. I know one of my friends had a really bad time because his mentor wasn’t about for the entire induction period. I felt pleased that mine was, and she was so good. [PNG 17] They were welcoming, but it would have been nice to have a mentor, or someone who pointed out what I was doing wrong, or spent a little bit more time with me explaining things. I felt quite frustrated in the first few weeks, because I didn’t really know how to use any of the systems on the computer. Almost, you feel like a bit of a burden continuously asking questions. You don’t want to ask the same person questions so many times, you are a bit useless or whatever. Yes, it would have been nice to have a bit more there. [PNG 41] We came in and we said, “We’re supposed to have this 28-day period. What do you want us to do in it?” they were like, “Oh, I’m happy for you to start in your roles.” At the time we were like, “Okay, that’s fine. We’ll start in our roles as neighbourhood officers.” Probably in retrospect, it would have been more useful to me to request to be put on a different team for a bit, or maybe just be on response for a little bit. [PNG 33] Line managers’ experience of the Police Now Programme A graduate’s experience of policing, regardless of entry route, is inevitably shaped by the dominant culture of their team and immediate line manager. Line managers who had visited the Summer Academy and met their recruit beforehand, unsurprisingly, had a much better understanding of the skills and competency of their recruit and what additional support would be required. Line managers’ accounts of supervising PN graduates tended to focus on the need to increase practice competency to achieve independent patrol status. Being abstracted from neighbourhood work to attend PN events or further training was sometimes perceived as an impediment to achieving these core policing skills and they were generally considered by line managers to be less practice-ready than other recruits. However, line managers were also mainly of the view that PN graduates were quick learners and highly competent individuals who showed initiative. Our line manager interviewees were critical both of the PN team and their own force Learning and Development Departments for a lack of clarity or communication about how the transition and mentoring process should be managed and what was expected of them as line managers.21 Most were also ill-informed about the focus and aims of abstraction days, highlighting disconnect between key parties involved in managing the programme. 8 3 Integrating into the neighbourhood team A key theme in our graduate interviews was how they integrated into neighbourhood policing teams. In some cases, they described how they sought to manage the perceived stigma of being a graduate entrant and to underplay any sense of difference between themselves and other colleagues: The stigma that comes with graduate is the idea that you think you know everything and you think you know how to be a police officer. I was so conscious that I didn’t know how to be a police officer, I didn’t want to give them any way of being able to say, “Oh well, she just comes in and thinks she knows everything. She can’t even do a proper stop and search. [Participant, Focus Group, 1] Yes, they knew they were getting someone from Police Now, but they didn’t know what Police Now was, no-one had ever told them what it was. I think they were sceptical to start with, but they said I wasn’t how they expected me to be, they expected someone who thought they knew everything, a typical graduate… Yes, just as I expected them to be really old men stuck in their ways. [PNG 21] Most were keen to emphasise and defer to colleagues’ professional experience and to acknowledge the importance of that experience for informing their day- to-day work and their PN impact assessments (discussed below) of what interventions might be implemented and tested locally. [with] Police Now, we were taught, “You’ve got these ideas. Go and do it,” but, equally, I think it’s important to reflect on the fact that our colleagues have probably done that before or have experiences of [place], that problem or people. [PNG25] They’ve got the experience of trying it in the field so if I read something and I’m having a think about something and then I chat to them about it, it’s given me much more knowledge and understanding of how that could possibly work in practice. [PNG17] Fundamentally, I might be better educated than anyone else in my office, but they’re better police officers than I am right now, and it would be arrogant of me to think otherwise. [PNG04] When you come out of university you know how to be a Uni student, not a police officer, not a community leader, you just know how to work efficiently; work, study and write things efficiently. [PNG 05] There were some who expressed more confidence in their ability to challenge or question colleagues, and of note was the autonomy that most interviewees had in responding to different neighbourhood issues – though sometimes this was thought to signal a lack of interest from line managers: I’ve probably got quite a lot of autonomy to be honest. I mean we normally, as a team, will discuss what’s going on, what problems we’ve got. Therefore, you get the advice of your colleagues and their input and stuff. Generally, if it’s something that I’m taking the lead on, as long as I can argue my case so to speak, my sergeant is very supportive with how I want to deal with that. [PNG 04] 21. Line managers who had not attended the SA had less knowledge about both the PN Programme and what needed to be put in place for the continued professional development of their graduate. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Yes I do, I feel like I’m given the choice to tackle it however I want. To be honest, I’ve felt, maybe at times, I’ve felt like I had a bit too much. I felt a bit under-supervised, in terms of like, “Well go on then, solve the problem.” You’re like, “Oh, I don’t know how.” [PNG 26] I’ve got a lot of autonomy in terms of how I want to approach problems. I could go to my sergeant with any kind of idea, as long as it’s somewhat sensible they’d probably be happy for me to crack on with it. I do feel like I have autonomy, and my input is taken in. However, obviously, the more experienced neighbourhood officers do get the final say. But, as a team, I think we get a lot of autonomy. [PNG 41] Most graduates described neighbourhood colleagues as generous with their time, keen to explain processes and systems and happy to help. A minority described a more cynical and cautious reception from colleagues, which they worked hard to turn around. Further, there was little opportunity for critical reflection within teams either because this was not normal team practice or because work pressure meant time for reflection was difficult to achieve in any meaningful or routine way. No. I must say I’ve only spoken to my sergeant once or twice properly, because she’s so busy. I feel like I’m almost being managed by this colleague of mine. Or not really being managed at all. Which is quite nice as well. So not really. [PNG 40] Not really. Most of the time if I’m out with my colleague and we’ve dealt with something, or given out a warning or whatever, we come back, we speak to the sergeant and say, oh we did this and that, and explain what happened, and that’s it. There’s not really a discussion with the team. Everybody does their own thing, so there is no kind of team discussion about anything. [PNG 33] Unfortunately I’m mainly left to my own devices in my neighbourhood. So ultimately I’ve got to pull the strings and try and pull everything together, the problem is I don’t always know what strings to pull. So I’m always feeling that I can only respond to certain situations with the tools that I’ve been given and I know there’s a much wider range of tools out there that I could be applying to community engagement, it’s just that I’m missing out on them and ultimately what I could achieve. [PNG 05] The role of evidence-based policing We examined how PN participants understand, value and apply research evidence to their practice during neighbourhood postings, through assessed impact presentations. The broader context of police professionalisation is pertinent. So too is the College of Policing’s role in setting professional standards and promoting good practice based on evidence – including the aim to create a more autonomous workforce, who can exercise judgment and knowledge of what works (Neyroud, 2010; Tilley and Laycock, 2017; College of Policing, 2015). Police leaders are supportive of evidence-based practice and are expected to encourage reflective practice and engagement with research (National Police Chiefs Council, 2016; College of Policing, 2015). 22. Projects entail graduates identifying problems in their neighbourhoods, designing an intervention approach and measuring the impact of that approach. 8 5 The What Works Centre for Crime Reduction – hosted by the College of Policing - is developing and promoting the professional knowledge base and is a repository for quality research evidence on what works and of good professional practice. However, positive orientation to evidence informed policing is not yet evident at all levels of the service, despite support from senior staff (Hunter, May and Hough, 2018). This makes the quality of the practice ‘learning environment’ for graduates highly relevant to this research. Chapter 2 stressed how important it was for conversion courses to integrate theoretical elements with practice skills, with mentoring and supervision as a key part of that integration process. The 100-day impact presentations22 are intended to demonstrate graduates’ problem-solving and critical thinking skills and are a means of applying evidence-based practice to address local neighbourhood issues. Each graduate is expected to complete up to five presentations over the course of the two-year programme in a variety of formats at ‘impact’ events. This is a core assessed element of PN and is used to demonstrate to host forces, the potential added value and ‘impact’ of a PN graduate to the neighbourhood team. Perceptions of evidence-based practice The SA initially included only one half-day session on evidence-based practice, which was augmented in 2017 by a further full day of skills training, focusing on defining evidence, showing how research can inform practice and different evaluation methods. This also allowed graduates to discuss their impact presentations with an ‘expert’ panel and encouraged them to champion the use of evidence and evaluation amongst their colleagues. As Chapter 5 shows, this extra training was found to deepen graduates’ understanding of evidence-based practice, to make clearer how to identify problems and underlying causes, and where to find different sources of relevant research evidence to inform and test interventions. Our focus group interviews conducted after this training supported these findings, with participants showing good understanding of, and support for, evidence-based practice – one participant described the session as “preaching to the choir”- but they also reported various barriers they faced in promoting research in their force. The following quotes highlight competing priorities for their attention and the cynicism expressed by some colleagues and managers about research or new ideas for age-old neighbourhood problems: It’s all very well saying, “you do have the time, make the time.” Well actually, you can’t make time, and when you’ve got your skipper going, “You’re doing this, then this, then this,” actually you can’t make time unless you stay on after work. So, I thought that was a little bit glib and slightly pie in the sky. Not understanding the realities of modern policing. [Participant, Focus Group 2] I think [EBP] is really important, because that’s the only way that workplaces change. If everybody only goes by what our predecessors do, it does end up being quite a backwards organisation that doesn’t really take on new things. I think, it is good to let new people have a say, from a fresh, outsider perspective of what needs to change, or what needs to be done. But, it is difficult being vocal in that sense, because you are new. [PNG 41] DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME I love evidence based policing. My whole Master’s was pretty much that. So when we had the lecture on it, I was just loving life and everyone was like, “Okay.” (Laughter) Yes, I just think there’s not time for it. Like we were saying before, officers probably wouldn’t have time to read and research, but I’m really keen on that, and learning [PNG 40] Interviewees’ perceptions of the value placed on evidence-based practice by their neighbourhood colleagues often underlined the disconnect between the learning and practice environments: [EBP] is not something that’s really, from what I can see, is used at a local level. And actually trying to do it involves time and resourcing commitment that’s not easy to do. But it’s something that I would like, or I can see that it would be beneficial to do. [PNG 08] The feeling I got was that most people in policing don’t know what it is [EBP]. That it is difficult in a day-to-day policing world to use it, because sometimes when we have our Police Now events, they are so far removed from everyday work. Yes. It is quite difficult to sometimes put it into practice in your everyday. [PNG 09] No one – which amazes me – no one is particularly strongly subscribed to evidence based policing, which when I was at the Summer Academy and we were talking about evidence based policing I was like, “why are spending so much time on this? It’s such an obvious concept. Obviously, we should be doing things that are proved to have worked. Why would we be doing anything else?” But when you come into [neighbourhood team], it’s not like that. [PNG 18] However, our interview findings also suggested that in some force areas and with some police managers, the interest in evidence-based practice had taken root and was being supported both in practice and through additional training given to graduates, there was just no overall consistency in the PN graduates’ experience in force: I know it’s something that the force has moved a lot more in the direction of using. We’ve had training from the force using an evidence-based approach to solve problems. I think that’s a cultural thing where, over time, it’s going to take a while, for people, I suppose, to come round to that approach. It’s certainly used a lot more than I had perhaps expected when I first started. [PNG 11] A lot of people pooh-pooh it. Quite a few say, “It’s more about experience than it is about academic research.” [but] I think there is a wave of change coming in [PNG 02]. 8 7 The impact presentations The impact presentations are an important link between learning and practice in that these presentations are an assessed aspect of the graduate programme, but they are also informed by and developed during everyday practice. The participants had mixed views about the aims and usefulness of the impact presentations. On the whole, they were seen as a distraction from enhancing practical or core policing skills; a view that echoed some of the police line managers we interviewed. Based on their experiences of undertaking presentations and some wider reflection on their purpose, interviewees offered some common suggestions for how these presentations could be more useful. This also raised questions for us about how the impact presentations could be better integrated with the wider aspirations of policing to develop the professional knowledge base. The current short-term focus on 100 days was thought by most interviewees to negate any impetus to review impact over the longer term or to use findings iteratively in order to develop and adapt interventions that may have worked elsewhere. One focus group participant raised the issue of legacy and a lack of clarity about how the research undertaken for impact presentations will be useful in the longer- term, which may simply be a reflection of how PN might better communicate the purpose of the impact presentations, as well as the way forces use the example presentations of problem-solving locally. You know, if what they’re encouraging us to do is to look for long-term solutions and not just look for a quick fix... If you try something, and it doesn’t have any impact for your first 100 days, the temptation is to be like, “Right, I’ll move onto a different project, then, because I need some impacts. I’ll find something that I can fix quickly.” That’s completely the antithesis of what they’re trying to teach us, in terms of long-term problem solving…, there’s some paradox there [PNG 26] Something that kind of demotivates me when I’m doing these projects is that I know that when I leave in July it will all revert back to normal, nothing will get carried on. So the legacy that we’re creating now, no one’s going to carry that on unless there’s new [graduates] replacing us. [Participant, Focus Group 1] A potentially misleading emphasis on innovation in solving neighbourhood problems and a perception (not necessarily supported by PN) among interviewees that documenting the failure of an intervention to address a specific problem would result in lower marks for the presentation, may have reduced the value of the exercise in the eyes of our interviewees. I think one of the falling down marks for policing in general is the fact that we can’t accept failure, and I think that we really need to do that. I think the NHS do that really well, and I think that we do that very badly, and I would love to see that change, because I think that my presentation would be great if I felt comfortable going up there and saying, “Here’s what I did, here’s why it didn’t work.”[PNG 18] DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME The pressure to succeed was also commented on by one of the police line managers we interviewed, who described being unconvinced by some of the impacts reported by graduates - “little bit of artistic license there”. However, he was also interested in their ability to be innovative and had high expectations of those he was supervising. Coming up with new ideas was perceived as difficult in the context of neighbourhood policing: These are people with no previous policing experience, who haven’t been tarnished - perhaps that’s the wrong word - but my hope was that they would look at [problem] with a complete fresh set of eyes and what I gave them were long-term issues. They weren’t coming up with the new ideas I’d hoped for. [Line Manager 03] It’s a pressure to come up with something innovative and new, and sometimes it’s to do with the job. There’s not always something amazing that’s going to come up. [Participant, Focus Group 2]. In our view, too much stress on innovation diminishes the impact presentations’ potential contribution to building professional knowledge, highlighting the need for greater coordination from either Police Now or the College of Policing about how graduate presentations like these might be better used to develop and refine the policing evidence base. For example, alongside new approaches and ideas generated through such assessments, there is huge potential for presentations to further test or adapt to ‘local’ conditions, interventions that have proved effective elsewhere. In the extract below, the perceived emphasis on innovation has actually served to reduce interest in What Works evidence: I think because Police Now is so focused on innovation and doing things differently, which is great, it’s a good thing to encourage but there’s such a heavy focus on it that it makes you feel quite reluctant to look to the College of Policing because obviously that’s all the tried and tested things. I think there’s so much pressure to try to do something different that it sort of counteracts all of the resources that are suggested in those lectures because they’re sort of saying, “Oh don’t do what everyone else has done”. [PNG 14] As noted, some participants were assisted by line managers and colleagues in the design and implementation of their presentations and these were sometimes linked in to current neighbourhood operations or based on neighbourhood priorities; two of the line managers we interviewed, reported that they determined the focus of the impact presentations. Line managers also assisted in most cases, but not all, by allowing time away from other duties to focus on presentations. Additionally, some of the LDOs were credited with helping participants develop presentation ideas. Overall, based on interviewees’ accounts, there was no standardised way in which participants were helped to complete their impact presentations: I think it is fair to say that [LDO] bounces questions to make us consider other options and make sure that we are considering evidence- based practice in what we are doing and that kind of thing. [PNG 11] 8 9 [LDO] is great to talk about what would be suitable for a 100-day impact event, where to look for the evidence base that I need to work on that sort of stuff. [PNG 19] Assessment and feedback are important elements of learning and development. A few of the interviewees mentioned the quality of feedback they received on their impact presentations. This centred on two issues –the transparency of the marking protocol - so understanding how the project was being assessed - and the usefulness of the feedback:23 It was anonymous feedback and I had no sense of how the [project] was being assessed. [PNG 02] I don’t know who does the marking, but I think if you have a teaching background it would be a lot more of an effective marking strategy. [PNG 18] It was useful to an extent but, this sounds awful, you know when you do well, and they don’t give a huge amount of useful feedback because it’s all mainly like, “Oh yes that was a good idea.” It was useful in saying I’d done a good job but… [PNG 17] The above demonstrates that assessment criteria for the DHEP should be clear, and probably a collaborative assessment between the force and an HEI. Content of impact presentations We reviewed the content of presentations conducted by our PN cohort 2 interviewees. The two most common topics for presentations were anti-social behaviour (e.g. street-drinking and drug use, noise, rough-sleeping) and improving community engagement to build confidence and improve police and community relationships. Other issues included the misuse of police time, reducing traffic offences, targeting street drug and sex markets and theft. The rationale for selecting the topic ranged from neighbourhood priorities to trends in statistical data, current operations and the time and interest of the participant. Most, however, provided some rationale for why the problem tackled was important in the longer term, such as building positive relationships with young people, increasing public confidence and increasing the flow of intelligence. Participants used a range of evidence to assess and scope the problem – usually force statistics on crime, call-outs and complaints, but also local authority data. They sought advice from colleagues, line managers and sometimes other forces. Some referenced national policies, including government and National Police Chief Council guidance. Less frequent were references to academic research in building the case for, or design of a project. The Scanning, Analysis, Response and Assessment model (SARA), for identifying issues, designing and assessing responses was mentioned regularly as were the concepts of hot- spot policing and target-hardening. 23. PN has since developed a more detailed marking scheme for the impact presentations but this had not been implemented at the time of our interviews. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Only three of the 30 graduates who discussed their impact presentation mentioned using the Crime Reduction Toolkit or other College of Policing resources in the development of their presentations. The weakest element of the presentations, unsurprisingly, was the assessment of impact, which in part reflects the stage in the development of evidence based practice across the police forces today. Part of the reason for the lack of impact assessment was the shortage of time or resources to undertake any assessment of impact. In some cases, the presentations were works in progress, so may have reported impact measurement at a later stage. Police call out statistics and incident reports, both before and after an intervention, were commonly used to show change. In some cases, verbal anecdotal feedback or testimonies from residents or colleagues were presented. We reviewed only one presentation that had attempted to compare data from the intervention site with a comparison site. However, most showed a good understanding of the difficulties of measuring impact and also the likely short-term nature of change and some consideration for how any positive impact might be sustained over the longer term. 9 1 In summary Our interviews with PN participants and their police line mangers about their experiences of this programme, highlight some important learning points for developing the DHEP. •• While our interviewees could largely cope with learning over a condensed period of time and were happy to complete some of these elements independently, there was a general call for more discussion and reflection alongside lecture-style learning during intensive learning. This might also help reinforce connection between the more theoretical or ‘soft-skills’ aspects of the training curriculum and policing practice. •• Common concerns about a lack of practical knowledge or skill on completion of intensive learning will need to be addressed through good coordination between HEIs and forces about the immersion and mentoring period and clear plans for how graduates’ practice will be developed in force and though top-up learning. •• More generally, communication between headquarters’ learning and development staff and those front-line staff who provide graduate participants with support and management needs to be effective, with clear roles and responsibilities and two-way lines of communication. •• DHEP participants will be expected to apply and promote evidence-based practice. It is important to recognise that this research indicates that they may meet with various obstacles, including the lack of support for EBP in some local areas. •• The Police Now impact presentations aim to enhance graduates’ problem- solving skills, to apply research to challenge ‘received wisdom’ and to test out innovative approaches to neighbourhood problems. There is clear scope for thinking further about how these presentations are done, improving their research elements and feeding the results into the police knowledge base. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Chapter 7 Insight from the research 24. Detailed reports containing full research findings upon which each chapter draws are available, and links to these can be found at the start of each chapter. 9 3 The findings presented in this report are intended to support the College of Policing and the police service as they develop the PEQF Degree Holder Entry Progamme (DHEP). The two-year programme will result in a Graduate Diploma in Professional Policing Practice, meaning that this route into the police service is compliant with the requirement that all police entrants are trained to degree level. DHEP is intended to ‘go live’ in 2020. Clearly these findings also have lessons for the other two police constable recruit entry routes – the apprenticeship degree and the pre-join police degree. The project has drawn on the available literature on police training and that on graduate conversion courses for other professions, and it has also mounted original research into the experience of participants in the Police Now (PN) programme. Whilst this is not formally a conversion programme, it shares features with DHEP that make it a suitable ‘test- bed’ from which to develop practical lessons for the latter. We have also interviewed other key stakeholders in police recruitment training, notably those in Learning and Development Departments, as well as people with experience of conversion courses for recruitment into the social work, law and teaching professions. The six substantive chapters presenting results all contain summaries of findings.24 In this final chapter we offer a synthesis of our findings, summarising in a few pages what lessons the project has for the DHEP. It is clear that there are significant challenges in implementing the PEQF overall, as well as specific challenges for each of the three routes into the police service. These fall into three categories: •• Pedagogic or educational challenges in equipping entrants with the increased level of skills and professional judgement that is now expected of police officers. •• Logistical challenges in substantially overhauling the entire recruit training system over a two-year period, and establishing new partnerships with universities. •• Force-wide change programmes in the context of austerity which have the impact of pushing learning and development outside the heart of the police service. The key findings of this project underscore the importance of striking the right balance of theory and practice in police learning. Therefore, finding the right approach to the pedagogic challenges is critically important to the entire ‘professionalisation’ enterprise. We shall consider these first, before turning to the logistic challenges and the impact of other contemporaneous change programmes. We urge learning and development leaders to keep these educational challenges to the forefront even though we see how the financial and logistical pressures too often take centre stage. These practical challenges are more immediate, but the pedagogic ones are probably more important, as the new training structures amount to a completely new way of working. We see a real risk that rushing to grasp what looks like a solution – commissioning an HEI provider and getting on with validating a curriculum – may bring more problems down the line without a firm grasp of the transformation necessary to meet the uplift in the educational approach. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Without a roadmap that includes the educational transformation, practical and process tasks often take precedence: validate a curriculum, recruit an entry cohort, and hope for the best. Whilst the academic and practical dimensions of DHEP training programmes might be in place, they could fall down in their delivery simply because policing theory and the practice remain largely unconnected with each other. As we very clearly found in the research, integrating formal evidence based knowledge and practice skills is one of the key pedagogic challenges facing DHEP and must be accomplished as a co-produced and cooperative venture. Combining theory and practice: the pedagogic challenge for DHEP At the risk of oversimplification, the existing learning approach for police recruits, recognised to be at Level 3, aims to ensure that officers know what they need to do when confronted with a given situation, but are not expected to understand why this is the right course of action. Recruit police officers are expected to possess specified skills, but do not need to have any theoretical justification to drive their decision making and interventions. The underlying assumption of the PEQF – also articulated in the Policing Vision 2025 – is that in the 21st Century a police constable has to be able to operate with greater autonomy than hitherto, exercising more discretion, to develop the right solutions to novel problems as these arise and evolve over time. This requires a theoretical grasp of the nature of policing, victimisation, crime and offending and their context, a commitment to continuous learning and some engagement with the evidence about what constitutes good policing practice. An agreement between police forces and their partner HEIs on ‘what good practice looks like’ cannot be taken for granted. Whilst ‘a good officer’ is envisioned in the Policing Vision 2025, each local force will have a view about what this means in their area. We recommend therefore that police force engagement with HEIs begins with explicit discussions about ‘what a good police constable looks like’ locally. Setting out a vision will enable the HEI to ground its theoretical input into the daily delivery expectations of the public and the force. Stating learning outcomes If the aim of PEQF (including DHEP) is to ensure that officers are intellectually equipped to act in a more autonomous and professional way, any entry learning programme needs to articulate what sort of police officer it aims to produce. A striking finding of the systematic review on police training was that few courses actually stated with any clarity what a good police officer looks like. Indeed, what Police Now (our test bed research site) found is that this expectation of ‘the good officer’ differs in different police forces. Whilst the College of Policing sets out a high level vision in its curriculum, any precise formulation of the ‘good police officer’ will always be contested and contestable and should, we believe, be locally driven. Moreover, other bodies have a view of what constitutes the ‘good officer’. Her Majesty’s Inspectorate of Constabulary and Fire & Rescue Services, for instance, leads force inspections taking a view whether forces as a whole are delivering efficient, legitimate and effective policing for local people. There is probably scope for a variety of such statements, but HEI/police partnerships need a clear shared understanding about this in order to provide a stated outcome for its co- operative and collaborative work. 9 5 To our minds, any such statements need to contain not only the attainment of cognitive or intellectual skills, but also the development of a values-based, ethical perspective that permits some empathy with all community members, including those who have enduring and extensive difficulties dealing with their lives. Building the bridge between formal knowledge and practical skills It was clear from our research with PN participants that practical competence when out policing the streets did not feel like it had much connection with any more formal knowledge and the ‘soft skills’ to which they had been exposed. What really mattered was having enough basic competence to do the jobs that had to be done, without losing all credibility in the eyes of their more experienced colleagues. To us as police researchers the links may seem obvious between the practical skills in, say, stopping and searching suspects and the theoretical frameworks about policing, such as procedural justice theory. However, PN participants made a clear distinction between the practical craft skills they valued and softer ‘people skills’ exemplified in the procedural justice perspective. This says less specifically about the PN programme, and more about the trickiness in blending craft with knowledge. In our judgement, one of the most important things to get right in the DHEP is to close this gap, so that recruits regard formal academic knowledge as relevant to practical policing skills and their practice as police officers. Translating theory into practice is also a skill, that will require close collaboration between HEIs delivering academic content and the mentors, line-managers and in-force trainers and learning and development staff equipping recruits with practice skills. It also will require an active socialisation of mentors to the importance of evidence-led practice, too. The academics need to appreciate the perspectives of the in-force staff, who in turn need to ‘buy into’ the academic perspectives that are being taught. If this doesn’t happen, the DHEP runs the risk of replicating the existing system of (level 3) training, albeit with an academic ‘bolt-on’ component which can safely be discarded once recruits have found their feet in the realities of front-line policing. We cannot overestimate the kinds of effort this collaboration will take – in spirit as well as in creating cultures within both the academic and policing worlds to do so. Evidence-based policing practice Part of the process of forging proper functioning links between formal knowledge and practical craft skills is arriving at a coherent specification of what is meant by ‘evidence-based policing practice’. It is fairly clear what constitutes the academic knowledge base on policing. It is wide-ranging, embracing extensive research in disciplines of sociology, criminology, social and organisational psychological and socio-legal theory. Such academic perspectives enable a widened and influential narrative within which course learning content should be aligned in order to enable recruit officers to know what and why such practices are important. Within this academic eclectic body of knowledge sits a small amount of experimental research addressing questions about effectiveness of particular tactics or approaches. It is probably important to be clear from the outset that evidence on policing practice DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME is drawn from a wider base of information than that informed through experimental research, and that many of the best insights into police working practices derive from non-experimental research. But equally, our research suggests that recruits do need to be clear about the role of empirical research in professionalising the service; they need to understand the research process, and to be open to challenge their own practices from research evidence. Styles of learning: group work and reflection It is clear from the experience of conversion courses for other professions that graduates have specific learning styles and that skilled educators also need to take into account the well- established fact that there are wide individual variations in these learning styles. Graduates bring to their learning both the capacity to absorb rapidly large amounts of information and ideas, and a range of analytic and intellectual skills honed through their university studies. Many recruits may also bring additional life experience from other jobs. It may not be the best use of student contact time in HEIs simply to ‘lecture at’ these graduates. If the aim is to generate genuine thoughtfulness, critical judgement and learning about the role for which recruits are training, it may make more sense to ensure that a good part of this time is spent in group work, and in providing opportunities for articulating and testing ideas about what constitutes effective policing practice is about. It is also well-established good practice that in level 6 training of this sort, additional time for reflection is equally critical. HEI providers may find it difficult to move from traditional styles of undergraduate teaching to adapt to the learning styles of this – unusual and potentially demanding – DHEP group. Learning and development staff have a role to play in requiring HEIs to consider how they will adapt their university based approach to learning to an approach that is about good practice and the continuous application of academic knowledge to consider new and better ways of policing. Motivation and emotional intelligence It is clear that there is a pool of highly qualified graduates available for recruitment. Extrapolating from the PN participants, they are characterised by intrinsic (value-based) factors such as doing a fulfilling job and ‘making a difference’ rather than extrinsic (or instrumental) factors. The aggregate profile of emotional intelligence for the PN graduates also seems suited to the demands of police work. The cohort seemed suited to a relatively structured and ‘top down’ training environment. The trainees are lower than the benchmark group in decisiveness and influence than comparable professionals, yet high on empathy, emotional control, conscientiousness and structure. Logistical challenges Interviews with key police stakeholders, as well as those with expertise in the provision of conversion courses for other professions, left no doubt that getting DHEP in place by 2020 will be challenging. There are several interlocking difficulties. The most pressing is the need to settle on HEIs that will serve as partners and providers – because progress on many subsequent issues needs a partner in place with whom to negotiate. Some forces, or 9 7 consortiums of forces, have pre-existing relationships with HEIs, and plenty of experience of joint working. Others may not have obvious HEI partners, and may lack experience in joint-working, and in negotiating the terms of any partnership. Procurement and contractual issues are inevitably complex – and may also carry implications for estates planning. It is critically important for police leaders to bring the discussions about the choices of HEIs and the assessment of how well the collaborations are proceeding into the heart of their senior management decision-making. We found wide variations in our discussions with learning and development staff in the kinds of senior support and champion support they are receiving from leaders in their forces. Divisions of labour between HEIs and L&D staff Once HEI providers are in place, there are choices to be made about the division of labour between HEI staff and in-force staff in the delivery of learning, and in the geographical location of these activities. These obviously mesh with the pedagogic challenges discussed above, especially in relation to ensuring full integration between HEI and in-force learning and development staff. There are additional issues to be addressed in some forces, to ensure that in-force staff are themselves equipped to operate at level 6 – which may require these staff to be qualified to Masters level – so that they are able to take an appropriate share of the delivery of level 6 material. Assistance with the uplift may be a legitimate part of the negotiations police forces have with HEIs as part of the design of the collaborative contract. Communication Both the experts on other professions’ conversion courses and our own research point to the need for effective communication between staff in HEIs, L&D staff in forces, front-line staff and their senior management, within the context of clearly defined roles and responsibilities. The way in which new recruits are supported by police forces and mentors when they first land ‘on the street’ is likely to continue to be variable. Some PN participants found their immersion in their home force a very positive experience, others had a much tougher experience, partly the result of mis-communication within forces. Some mentors and line managers were highly praised, whilst others had been poorly equipped to do their jobs and provide effective support for new recruits. Staff at all levels need to buy into the professionalisation agenda, including PEQF, and if they will not do so, they must at least be encouraged not to subvert it. Time to learn: abstractions and the continuing demands of demand Beyond all these practical challenges, there is a need to take full account of the abstraction demands of DHEP training – when the assumption that will readily be made is that demands for ‘boots on the ground’ will take precedence over educational activities, especially when these involve self-directed learning and time for reflective learning. This means that new and creative ways of blending time on and off the job for learning and reflection is needed. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME And finally, some considerations for the College of Policing in developing the DHEP There are key findings that the College may wish to consider in the iterative development of the DHEP: •• To articulate clearly the desired outcomes of graduate entry education with explicit criteria for becoming a ‘good police officer’ and for achieving ‘good policing’ •• To articulate the aspired model of professionalism that underpins the PEQF by drawing on learning from other professions •• To review the published curriculum for the PCDA and concentrate on fusing theory and practice in the different modules including explicit consideration of applying EBP principles to practice •• To provide specific guidance for forces on working collaboratively with HEIs – especially in creating new ways of co-producing learning integrating HEIs’ contributions to this new curriculum •• To give more consideration to the candidate (learner) perspective in policing education from their initial selection; including the consideration of personal motivation and other characteristics, and through ongoing support and monitoring •• To highlight the receptivity of the Service as a whole to the creation of a graduate profession of policing as an ongoing issue •• To support the undertaking of an on-going major evaluation of the DHEP, alongside its evaluation of the PCDA. 9 9 DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME References 1 0 1 Belur J., Agnew-Pauley W., Tompson L. (2017), Rapid evidence assessment on graduate conversion courses in other professions, PIF Project Belur J., McGinley B., Agnew Pauley W., and Tompson L. (2018) Systematic review of recruit police training, PIF Project. Belur, J., Agnew-Pauley, W., and Tompson, L. (2017). Rapid evidence assessment (REA) on graduate conversion courses, Police Practice and Research (under review). Berry-Lound SD, Tate S. and Greatbatch D. (2016), Social work teaching partnerships programme pilots: evaluation Final Research Report, HOST policy research, Social Science in Government. https://www.gov.uk/ government/uploads/system/uploads/ attachment_data/file/563261/Social_ Work_Teaching_Partnership_Programme_ An_Evaluation.pdf (Accessed 20th December 2017) Bruns, D., and Bruns, J. (2015) Assessing the worth of the college degree on self- perceived police performance. Journal of Criminal Justice Education. 26:121-146. Chan, J. (1996). Changing police culture. The British Journal of Criminology, 36(1), 109-134 Charles, M. T. (2000) Police training: breaking all the rules. Implementing the Adult Education Model into police training, Charles C Thomas Publisher College of Policing (2015) The Leadership Review: Recommendations for delivering leadership at all levels. College of Policing: London. Croisdale-Appleby D. (2014) Re-visioning social work education: An independent review, Department of Health, UK Government. https://www.gov.uk/ government/uploads/system/uploads/ attachment_data/file/285788/DCA_ Accessible.pdf (Accessed 20th Dec 2017) Hallenberg, K., and Cockcroft, T. (2017). From indifference to hostility; police officers, organizational response and the symbolic value of in-service higher education in policing. Policing 11: 273- 288. Heslop, R. (2011) Community engagement and learning as ‘becoming’: findings from a study of British police recruit training. Policing & Society, 21(3), 327 - 342 Heslop, R. (2013) Police pre-employment training the United States of America: (A study conducted under the auspices of a Fulbright Police Research Scholarship). Wakefield, UK: West Yorkshire Police Training and Development Centre. Available online at: http://www. richardheslop.net/PDFs/Heslop%20 Fulbright%20USA%20Police%20 Training%20Report.pdf Hunter, G., May, T., and Hough, M. (2017) An Evaluation of the ‘What Works Centre for Crime Reduction’ Final Report. London. ICPR http://www.icpr.org.uk/ media/43599/icpr_final_evaluation_ wwccr.pdf Hunter, G., May, T., and Hough, M. (2018) Are the police embracing evidence- informed practice? A view from England and Wales. London: ICPR Working Paper. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Kirkpatrick, D.L. (1967). Evaluation of training. In R.L. Craig and L.R. Bittel (Eds.), Training and Development Handbook. New York: McGraw-Hill. Kirkpatrick Partners (2018). The New World Kirkpatrick Model. Retrieved from https://www.kirkpatrickpartners.com/ Our-Philosophy/The-New-World- Kirkpatrick-Model Laycock, G. and Tilley, N. (2018 in press) Special issue paper Lee. M., and Punch, M. (2004) Policing by degrees; police officers’ experience of university education. Policing and Society. 14:233-249. Mark, R. (1977) Policing a perplexed society. London: George, Allen and Unwin. National Police Chiefs Council (2016) Policing vision for 2025.http://www.npcc. police.uk/documents/policing%20vision. pdf Neyroud, P. (2010) Review of police leadership and training. London: Home Office. https://www.gov.uk/government/ publications/police-leadership-and- training-report-review Pearson S. (2007). Preparing for inclusive education: the pre-course experiences of prospective secondary school teachers on a UK-based training course. Journal of Research in Special Education, 7(2), pp.124-135. Paoline, E., Terrill, W., and Rossier, M. (2015) Higher education, college degree major and police occupational attitudes. Journal of Criminal Justice Education 26: 49-73. Police Now (2016) Police Now Information and Briefing Pack. London. Police Now. https://www.policenow.org.uk/wp- content/uploads/2016/09/Police-Now- Programme-Booklet.pdf Ritchie, J., & Spencer, L. (2002). Qualitative data analysis for applied policy research. The qualitative researcher’s companion, 573(2002), 305-29. Paterson, C. (2011) Adding value? A review of the international literature on the role of higher education in police training and education. Police Practice and Research, 12: 286-297. Tang S.Y.F, Wong A.K.Y. and Cheng M.M.H. (2016). Examining professional learning and the preparation of professionally competent teachers in initial teacher education. Teachers and Teaching: Theory and Practice, 22(1), pp.54. Tilley, N., and Laycock, G (2017) The why, what, when and how of evidence-based policing. In Knutsson J and Tompson L (Eds). Advances in Evidence Based Policing. Routledge. London. Tong, S., and Hallenberg, K. (2018) Education and the police professionalization agenda a perspective from England and Wales In Rogers, C., and Frenel, B. (Eds.) Higher Education and Policing. Springer (pp17-24) Wilson G., Kelly B. (2010). Evaluating the effectiveness of social work education: Preparing students for practice learning, British Journal of Social Work, 23, pp. 1-19 1 0 3 Wimshurst, K., and Ransley, J. (2007) Police education and the university sector: contrasting models of Australian experience. Journal of Criminal Justice Education, 18:106-122. Wood, D. A., & Tong, S. (2009). The future of initial police training: a university perspective. International Journal of Police Science & Management, 11(3), 294-305. Wong A.F.L, Chong S, Choy D, and Lim K.M. (2012). Investigating changes in pedagogical knowledge and skills from pre-service to the initial year of teaching. Educational Research for Policy and Practice, 11(2), pp.105-117. Yesberg, J and Dawson, P. (2017) Police Now Cohort 1: final evaluation report. London: Mayor’s Office for Policing and Crime. https://www.london.gov.uk/sites/ default/files/mopac_police_now_ evaluation_report.pdf Zhang L.J. (2004b). Reforming a teacher education program for PRC EFL Teachers in Singapore: Sociocultural Considerations and Curriculum Evolution. International Journal of Educational Reform, 13(3), pp.223-252. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Appendix A Stakeholder interviews about graduate conversion courses W. Agnew Pauley and J. Belur 25. UK Government Teachers’ standards: https://www.gov.uk/government/publications/teachers-standards 1 0 5 The rapid evidence assessment reported in Chapter 2 identified specific gaps in knowledge about conversion courses. Unanswered or partially answered questions included: 1. What basic assumptions are made about graduates’ abilities and skills that form the basis of the intensive and condensed curriculum design? 2. How are theory and practice integrated? 3. What is the rationale behind particular course structure and delivery method? 4. How are the assessments of the various theoretical/conceptual and practical components designed and distributed between organisations/ employers and HEIs? 5. How much involvement and input should employers/end user organisations have in the course content, design, structure and assessment criteria for conversion courses? To address these gaps we interviewed seven academics across a range of Higher Education Institutions (HEIs) who oversee or coordinate graduate conversion programmes for social work, teaching and law. Methods and findings are described more fully in a full report (Agnew Pauley and Belur, 2018) and is available from j.belur@ucl.ac.uk This appendix summarises findings. Description of graduate programmes Table 8 below summarises the different routes into the three professions. Our interviewees provided additional material on the different programmes. 1. Teaching The graduate conversion course discussed by Interviewee 3 was the Postgraduate Certificate of Education (PGCE) which is a popular route into entry level teaching for graduates who hold a Bachelors degree. The PGCE combines university-based theoretical learning with practical training through placements in schools and leads to the attainment of Qualified Teacher Status (QTS).25 This course is one year (full time) in duration. There are a number of different routes into teaching which fit into two broad categories: university-led or School Direct route. For the university-led route the student applies to the university directly and follows the “programme designed by the university, provided by the university in partnership with schools where the student does their training placement” (Interview 3). For the School Direct route, students apply for a position at a school (or a cluster of schools), who (often) are in an established partnership with a university. Once students are accepted by the school for a teaching position, they are then interviewed by the partner university. Both institutions must agree that the student is to be accepted. Students on the School Direct route then study along-side the university-led PGCE students to attain their PGCE and QTS. A slight variation in the Schools Direct route is a salaried route, where students apply to the school and are accepted for a salaried post, prior to being a qualified teacher. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME The school provides most of the training in-house, however the student is released for 15 days a year to attend a partner university. Students on this route only qualify for QTS, and are not awarded the Level 7 PGCE, since it is perceived that 15 days in the HEI is “not enough input from ourselves [the university] to qualify for a PGCE”. According to our interviewee, this route supports the training provided by the school and also “provides our own quality assurance that they can be awarded a QTS at the end of the year”. However, the interviewee acknowledged that the quality assurance process of students through this route is not as rigorous or thorough as the university or other Schools Direct routes. This is partly driven by the fact these students are often hired by schools where there has been a shortage in the supply of teachers in key STEM subjects (science, technology, engineering and mathematics) for example, and therefore face difficulties in recruiting. Such schools are eager to fill posts and train teachers quickly, to meet the demand. Tuition fees can be covered either by the school, or through scholarships and bursaries available for particular subjects, or by the student (Interview 3). 2. Social work Graduates who wish to become social workers, and who hold a degree in a subject other than social work have two routes into the profession: they can either enrol with a university in a postgraduate qualification (either a Diploma or Masters degree), or train within a number of ‘fast- track’ social work programmes that operate in the UK. Courses discussed are two years in duration, with a national requirement of a minimum of 170 days in practice (Interview 1a; Interview 3). Requirements for both routes specify a 2:1 or higher undergraduate degree (in a subject other than social work), Grade C or above in GCSE Maths and English and some relevant practical experience. Students must pass a selection process, which includes an interview, to be offered a place. The fast-track programmes are fully funded, and students receive a bursary for living expenses. For graduates pursuing the postgraduate route (i.e. not a fast-track programme) social work bursaries are available from the government. 3. Law For law, the graduate conversion course discussed in Interview 2 was the Graduate Diploma in Law (GDL) or Common Professional Examination (CPE), which is the standard qualifying programme for law in the UK. The course duration is one-year full time or two years part time. Conversion courses in law differ to other professions examined in that they do not include a compulsory practical element. The GDL is an intensive, theoretical content based course, which provides students with a knowledge base of the law. Students complete the Legal Practice Course (LPC) which is a separate programme also run by a university for fulfilling the practical component. To qualify, students must then secure a two-year training contract in a law firm during which time they have no connection with a university. There is no bursary or Government funding available for those who wish to study the GDL, however it is possible to receive sponsorship from an employer to cover the costs if a student has secured a training contract with the employer prior to beginning the GDL. The law programme referred to in this report relates only to the GDL, not the LPC. 1 0 7 PROFESSION ROUTES LENGTH FUNDING ACCREDITATION EMPLOYMENT Teaching Undergraduate degree Three- four years Funded by student or through bursary / scholarship Bachelor of Education (Bed) or Bachelor of Arts/Science (BA or BSc) and QTS Post-qualifying University-led/ PGCE One year Funded by student or through bursary / scholarship PGCE and QTS Post-qualifying Schools Direct Unsalaried - Funded by student or through bursary / scholarship PGCE and QTS Pre-qualifying Schools Direct Salaried One year Funded by the training school QTS Pre-qualifying Fast-track teaching programmes Two years Funded by the fast-track programme PGCE (with available top up to Masters) and QTS Pre-qualifying Social Work Undergraduate degree Three years Funded by student or through bursary / scholarship BA or BSc (Hons) Post-qualifying Postgraduate (Diploma or Masters) Two years Funded by student or through bursary / scholarship Masters in Social Work (MSW) or Postgraduate diploma Post-qualifying Fast-track scheme Two years or 14 months Funded by the fast-track programme Masters in Social Work (MSW) and/or Postgraduate diploma Pre-qualifying Law Undergraduate degree Three years Funded by student or through bursary / scholarship Bachelor of Laws (LLB) Post-qualifying Graduate Diploma in Law (GDL) and Legal Practice Course (LPC) 9 months for GDL, followed by 9 months LPC Self-funded or sponsorship from an employer GDL and LPC Post-qualifying TABLE 8: ROUTES INTO TEACHING, SOCIAL WORK AND LAW IN THE UK DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Assumptions about graduates’ abilities and skills Interviewees were asked about the assumptions they think are made about graduates’ abilities that underpin the design and length of the curriculum. According to the interviewees, graduates should be, by nature, more mature, with (potentially) more professional or life experience, and possess a general understanding of how a university programme operates. It is also assumed that their working habits are more organised and structured, compared to a typical undergraduate (Interview 1a; Interview 2b). Furthermore, according to one interviewee, attainment of an undergraduate degree indicates that they should possess: “critical analysis and independent analysis… they have to have strong academic literacy, as well as communication skills, both verbal and written… and resilience, patience to be able to cope with incredibly stressful work situations” (Interview 3). A key quality necessary for the profession identified in the teaching and social work interviews was resilience. For these professions, the assessment of an applicant’s resilience was conducted at the recruitment stage and was considered a necessary criterion for being accepted on the programme. Resilience was referred to as the ability to cope with the demands of the course and the profession. Students in social work, for example, are asked to identify their own strengths and weaknesses and to give examples of how they can demonstrate being able to cope with a highly intensive course. Applicants are also expected to demonstrate high levels of motivation for joining the profession. They are expected to come to the course with a detailed understanding of the profession and evidence behind their reasons for joining, according to Interviewees 2a, 3 and 4b. This is tested at the interview stage for teaching where, for example, they are asked to demonstrate a “real interest in working with young people and children, demonstrably so” (interviewee 3), or in social work, students will be asked why they want to join the profession, what they know about it, and what they might have done to prepare themselves for the degree (i.e. through relevant practical experience, or through what they have read about or seen in current affairs) to ascertain “how motivated they are” (Interviewee 4b). The assumption that graduates were more familiar with the academic requirements of a degree programme had implications for curriculum design and length. Our interviewees admitted that the graduate conversion curriculum, delivered generally over one year, was virtually the same as that for a three-year undergraduate degree programme, just delivered in a more condensed form. As outlined by one interviewee: “The syllabus might look exactly the same [as the undergraduate programme]… but what you actually teach from that syllabus can be different… [for example] I tend to look at the same number of subjects but focus on one or two topics in much more detail [for the conversion course]” (Interview 2a). Notably, interviewees expressed an understanding that graduates had to be ready and able to work at a heightened level from the commencement of the course: “what they are presumed to bring with them is the ability to study at that heightened level” and “that they are capable of getting to Level 6 within 9 1 0 9 months” (Interview 2a). They have to be ready to work at a “faster pace and more intensely” (Interview 4a). This assumption was referred in some interviews as a justification of slightly front-loading a course with an intensive academic introductory period, as graduates can cope with more in-depth and academic work from commencement, ensuring they do not miss any time preparing for their first experience in practice. Additionally, graduates tend to be introduced into practice or placement earlier on in the course and at an intensive pace that “builds up over time”, as compared to undergraduate students (Interview 3). The graduate conversion courses discussed in these interviews had a rigorous and competitive selection process to ensure they attract graduates who are ready and able to work at this heightened level and who demonstrate the required qualities for the profession, namely critical thinking skills, problems solving skills and resilience (Interview 1a and 1b; Interview 3; Interview 4a). Mechanisms for integrating theory and practice Interviewees were asked about the underpinning rationale for the delivery and structure of the course, particularly in relation to the balance between classroom and practical content, which identified a number of mechanisms for integrating theory and practice. The integration of theory and practice is a key component of any practice-oriented profession. As identified in the REA: “the integration of the practical elements of training and the theoretical elements are achieved in order to ensure that the practical element, as it is considered to be indicative of job worthiness, is not privileged over the theoretical elements, which underpin the ethos and conceptual underpinnings of the profession relevant to the context in which it is practiced” (Belur et al 2017). Whilst the importance of this integration has been acknowledged in the academic literature on graduate conversion courses (see for example Tang, Wong and Cheng 2016; Pearson 2007; Wong et al 2012; Zhang 2004), specific mechanisms for achieving this are not always explicit. The three interviews relating to teaching and social work indicated that the integration of theory and practice was essential to ensure that training moves beyond acquiring technical skills necessary to fill a particular role, and leads to the development of a professional (Interview 1a; Interview 3; Interview 4a). For law, the distinction between theory and practice appears to be more distinct as theoretical and practical elements are taught as two discrete programmes. However, in the interview it was acknowledged that greater integration would be beneficial for the profession and there might be a move towards this in the future (Interview 2b). The two main mechanisms for integrating theory and practice were course structure and the role of tutors, both of which require coordinated and collaborative partnership between the HEI and practice organisation. Interviewee 3 stated that integrating theory and practice begins with: “The design of the programme as a starting point and the resistance to the idea that there is a theory-practice divide… it’s about the education of the tutors, it’s about programme philosophy and it’s about the way in which the programme is structured”. These collaborations were long standing in the teaching and social work professions and were constantly being reviewed and updated. DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME Course structure One of the key findings from the interviews was that the integration of theory and practice needed to be embedded in the design and structure of the course. Teaching and social work programmes for graduates are slightly frontloaded with an intensive, academic block of learning at the beginning of the course (Interview 1a; Interview 3; Interview 4a). Within this first block it is important that themes, or “threads” are made explicit so students understand the importance of why they are learning something at a particular time and how this will benefit them later in the course as well as providing a clear structure during this intensive, academic period (Interview 1b). This could be a key teaching or learning theory in Teaching or keeping client welfare foremost in Social Work. Following this initial block, most courses discussed introduced their students into their first placement early on in the course to maximise their time learning in the field. Courses are structured so that placement gradually increases over time, but days spent returning to the HEI to reflect and develop on their practical experience remain consistent throughout the programme (Interview 1a; Interview 3; Interview 4a). Thus, reflective practice played an important role in the courses in order to provide students with the space to reconcile experiences or problems faced in the ‘field’ with theoretical inputs provided in the classroom. Interviewee 3 explained how this is operationalised in a graduate teaching programme as follows: “we work at the beginning to help our trainee teachers understand that there is a theory of teaching that underpins what we do, but we don’t give them lectures on learning theory… what we do instead is every single lecture that we do… is underpinned by theory, everything is practice that is then theorised – but most of our work is not through lectures”. An example then given of a typical day spent in the HEI for a trainee teacher might be a lecture in the morning, followed by practical individual and group workshops spent with tutors and peers, so that students have the opportunity to draw on practical experiences from placement and to look at the theory behind why they did this (Interview 3). One interviewee explained that as a profession, social work education has grappled with the integration of theory and practice . As a result of a number of evaluations of social work education that identified a significant gap in graduates’ ability to apply what they had learnt to the practice environment, there have been a number of initiatives developed to introduce a new model of social work training to make the link between theory and practice more explicit (Interview 1b; Interview 4a). This forms the underpinning basis of the course: “We see the integration of theory and practice… as the line through the middle that we’re always trying to hold them together in a creative tension and we are committed to doing things in a relational or group way” (Interview 4a). This was understood to mean that theory and practice is embedded in the course design, and then further explored throughout the course through group discussion with peers or tutors. In this course there are distinct modules that are explicitly about, and even titled, theory and practice in social work. This module takes place while students are in placement and are returning to the HEI for one day a week, and time is specifically spent drawing examples from their practice environment and applying 1 1 1 and discussing theory with a tutor and other peers. Peer learning and collaboration through group work is vital here (Interview 4a). Another example used in social work is the Blended Learning Model, which is about “delivering the curriculum in the practice environment” (Interview 1a). Students are placed by their practice institution in units of four and spend most of their time in placement, with allocated teaching days at the HEI throughout the year. They also meet regularly with their tutors and practice specialists, both individually and with the rest of their unit. Each week there is a formal consultation group with four members of the unit and their consultant social worker, where cases are presented, linked to academic work and reflected upon: “A lot of the academic work is very much practice based, so they have to do a lot of reflective writing, for example” (Interview 1a). The teaching methodology is about “looking at the theories that are going to be relevant in that situation, sharing learning, peer support… reflective practice is core to social work” (Interview 1a). These two social work examples demonstrate the importance of building an integration of theory and practice into the design and structure of a course as well as the importance of collaboration and cooperation between the HEI and practice institution as both need to work together to deliver these models. The role of tutor Key to each of the models described above is the role of the tutor/mentor in assisting students develop the link between their academic and practical learning and thus understanding the theory/practice cycle. This emerged as a key finding within both teaching and social work as students have a range of different tutors, each with a clearly defined role. In both professions there is an academic tutor from the HEI who takes responsibility for assisting with the academic components of the course including coursework and assessments (Interview 3). As an example from social work: “in the academic context, the person who runs the [theory and practice] module is also their academic tutor… this is their first port of call in relation to anything related to the course whether it’s practice or academic” (Interview 4a). This tutor also visits them midway during the placement. There is also a practice tutor/mentor that is selected from the organisation within which the student conducts their placement to assist with guiding their practice development. Providing education, training and support to both academic and practice tutors/ mentors so they are clear on their role in aiding students to link their practice to theory, and to develop professional practice, is crucial. Both social work programmes discussed had three tutor/ mentor roles with an additional practice educator or practice specialist who performed an enhanced supervisory role within the placement institution to provide a further level of theoretical and conceptual support (Interview 1a; Interview 4a). Having clearly defined roles and responsibilities for both types of tutors is imperative as emphasised by Interviewee 3 who gave an example of a graduate programme within teaching: “The role of the people who were supporting the development [practice mentors] was importantly different from the university tutor role… they will be mentoring [the] development of the practitioner, having conversations with them and professional dialogue”. Assessment of learning Apply in practice Academic / theoretical learning Reflect on learning with tutor / mentor and peers Apply in practice DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME The practice tutor/mentor’s role was to support the practitioner to engage with the programme and oversee the student’s development as a practitioner, but not to deliver the academic workload. According to Interviewee 3 these tutors required specialised mentor training for the role, however this did not necessarily mean they needed to gain a higher qualification. It was felt more broadly that it is also very important that all tutors understand that they need to work collaboratively and coordinate through the partnership that exists between the HEI and practice institution (Interview 1a; Interview 3). Additionally, central to the role of the tutor/mentor is their involvement in the assessment process. Assessments for graduate conversion courses essentially tie in the integration of theory and practice with the role of the tutor. Assessments also act as an important quality assurance mechanism to ensure that students are making the connection between theory and practice and are consistently developing throughout the programme (Interview 1a; Interview 3; Interview 4a). For example, in the teaching programme described, academic assessments were assessed by academic tutors but also required the student to incorporate elements from their practice. For the practical components, students “had to be observed on at least five occasions per placement by the school mentor… plus one observation per placement by the institute [academic] tutor” (Interview 3). Similarly Interviewee 4a explained that in social work, students’ practical training had to be signed off by their practice tutor, in addition to which they had to pass a practice assessment panel consisting of both academic and practice tutors, and submit a portfolio of work that demonstrates their academic learning: “assessments are read by both academic tutors and practice educators so we get a quality assurance mechanism that the standard of practice education is consistent” (Interviewee 4a). Common to these interviews was evidence of a theory/practice cycle whereby students throughout the course combine academic learning with practical experiences, and spend time reflecting on this with tutors and peers to ensure depth of learning. This cycle is demonstrated in Figure 9. FIGURE 9 – THEORY / PRACTICE LEARNING CYCLE 1 1 3 Importance of strong partnerships An overarching finding that emerged across the teaching and social work interviews was that the mechanism that ensures the effectiveness and efficiency of a graduate course, including the delivery of integrated theory and practice, is a strong, working partnership between a HEI and practice organisation. For these courses, there needs to be mutual agreements and understandings that are formalised. Relationships need to be built so that there is coordination and cooperation during all stages of the programme from the course design, recruitment and selection of students, delivery of programme and assessment of the student: “from beginning to end of the process, from when students apply to when they graduate, we’re working closely the whole time with our agency partners” (Interview 4a). A number of different partnership models were described in the interviews, particularly for teaching and social work. Both teaching and social work in the UK have formalised teaching partnership models that set out the relationship and governance between HEIs and practice institutions . These partnership agreements outline the collaboration between both institutions at all stages of the course meaning that practice institutions can be involved in the admission process, contribute to teaching by having practitioners give guest lectures or workshops and provide opportunities for academics to re- integrate in practice to update their knowledge (Interview 1a; Interview 3; Interview 4a). An example of this was given in a social work programme where academics: “go out and immerse ourselves back in practice to make sure we are familiar with what practice looks and feels like” (Interview 4a). These partnerships are maintained formally through memoranda of understanding and agreements, management boards and working groups, as well as informally through the relationships that are built between the institutions through regular consultation (Interview 1a; Interview 1b Interview 3; Interview 4a). This ensures that there is common understanding from both sides of the arrangement, and all parties are clear in terms of their respective responsibilities and mutual objectives. Problems arise between partnerships particularly when HEIs and practice institutions “do not have a shared understanding of professional learning” (Interview 3). Quality assurance Quality assurance mechanisms are routinely built into these graduate programmes as part of the partnership agreement. These can take a number of forms. Within graduate courses, the HEI acts as an important quality assurance mechanism both for recruiting the right students on to the programme and as support for students who are out on placement. To ensure that the training being delivered within the practice institution is following the guidelines and does not become insular: “that’s where the quality assurance and the recruitment process comes in” (Interview 3). Having an academic tutor visit students during placement as well as develop a supportive relationship with the student ensures that if any problems arise, such as poor support within the placement institution, inadequate mentoring, too much responsibility, or not enough time allocated to course requirements, the academic tutor can bring this to the attention of the HEI and address the issue through the partnership agreement (Interview 3). DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME “Practice specialists have a really important quality assurance function… their job is to provide coaching and mentoring to the consultant social worker… oversee the unit and manage local relationships” (Interview 1a). HEIs also have existing student support systems in place that can be used by students (Interview 1b; Interview 3). Formalised and transparent partnership agreements that involve a HEI also ensure some level of consistency across different contexts. Within teaching and social work, the role of the overseeing professional body is highly important. For example, the Health and Care Professional Council (HCPC) which validates social work programmes and the Office for Standards in Education (OfSTED) inspections ensure consistency across teaching programmes (Interview 1a; Interview 3; Interview 4a). Maintaining professional standards The final key finding to emerge from the interviews was the importance of maintaining professional standards in training. Interviewees from all three professions – teaching, social work and law – articulated similar concerns surrounding the importance of the integrity of the profession. The justifications given for each of the above themes: acknowledging graduates’ abilities, integrating theory and practice and maintaining strong partnerships was to address and maintain professional standards (Interview 1a; Interview 2a; Interview 3; Interview 4a). One of the risks of neglecting the academic or theoretical side of any of these professions, particularly teaching and social work, and over-privileging the practical or technical sides, is that students become trained for a job rather than a profession. Training students to simply fill a gap, be it during a recruiting crisis or to “fit a mould” for a particular post is ultimately to the detriment of the student as it limits their professional development in the long term and compromises their future professional trajectories (Interview 3). The importance of the qualification as a learning and development experience cannot be underestimated, and should not be viewed as a “tick-box” process (interview 2a; Interview 3). This was identified as particularly important within law, as a lot of the commercial graduate conversion course providers offer more of a “polytechnic course designed to get you a job” rather than focus on “the professional development of a lawyer” (Interview 2a). This has been a previous issue in teaching where is has been recognised that: “It is very important to emphasise that teaching is an intellectual profession as well as a practice and that the two go together… very important to us as a stance as there is such an anti-intellectual move in teaching at the moment” [as a result of the crisis in teacher recruitment] (Interview 3). This is a further reason for ensuring students keep returning to the HEI with their peers, so that they form a sense of being part of a professional community (Interview 3). Furthermore, it is important to build in a research component to graduate courses, for their professional development, to develop an understanding of evidence-informed practice and to also continue building the evidence base of the profession. It is recognised in some programmes that graduate conversion courses do not have the time to allocate to an empirical research project or dissertation. One compromise provided by Interviewee 4a was that students complete a literature based dissertation (for example, over their summer term) in an area of the profession they are interested in to demonstrate a depth of understanding. 1 1 5 Implications for the DHEP Overall, the key components of an effective graduate programme are interrelated, and must have quality assurance mechanisms built in at all stages to ensure the course is achieving its intended outcomes. Beginning with capable and resilient graduates who have been accurately selected through a rigorous application process, ensures that any assumptions made in the course design and delivery are fairly met. The design of a graduate course should be underpinned by the integration of theory and practice, achieved through the structure of the course, within regular built in days spent in the HEI as well as in the field. This integration is aided by both academic and practice tutors, who have adequate training to understand their role in supporting students to link practice to theory. Assessments act as a further quality assurance mechanism to ensure students are making this link. Essential to effective tutoring and assessment processes, as well as the overall design and delivery of a programme, is a strong, working partnership between the HEI and the practice organisation that is maintained through formal agreements and governance, as well as informally through continual contact and collaboration at all stages of the programme. At all stages of this process the overall objective of developing professional practice and maintaining professional standards should be at the forefront of the programme, which means maintaining programme consistency across contexts, embedding research components for students and framing course design and delivery with evidence-based practice. Figure 10 indicates the relationship between core graduate abilities being harnessed by a programme that is designed and developed in partnership mode with the overall objective of integrating theory into practice to encourage the development of an evidence based and reflective professional police officer. FIGURE 10 – RELATIONAL MODEL FOR EFFECTIVE GRADUATE PROGRAMMES Integration of theory and practice Strong partnerships Graduates’ abilities Design and structure of programme Role of tutors and assessments Resilience Critical Thinking Selection & recruitment Formal agreements Oversight by professional body Professional Standards Link research to practice DEVELOPING AN EVIDENCE BASED POLICE DEGREE-HOLDER ENTRY PROGRAMME For policing, the development of a graduate-entry route into the profession forms a key part of the professionalisation process. Drawing on professions that have existing graduate entry routes and have been through a similar process, namely teaching, social work and law, it is important that this learning is drawn upon. With this in mind, the findings of this report come to a number of policy implications for the development of a graduate-entry programme for police recruits. Initial selection and recruitment 1. Graduates can be assumed to possess a number of key qualities, including critical thinking, problem solving, resilience and high levels of motivation to join the profession. Police forces should include and assess these qualities in recruitment and selection criteria, so that the programme is targeted to the right individuals, reducing the risk of student drop-out and improving retention within the profession. 2. For policing, although recruits apply to forces, involvement of the HEI at some stage of the recruitment or application process is likely to ensure that individuals possess the required qualities and abilities. Integrating theory and practice 3. Graduates’ abilities and experiences should be integrated and built on in the design of the course. 4. Frequent opportunities for students to reflect upon and integrate their practical experiences with theoretical content need to be built into the course design. For policing, it is vital that days spent in the HEI remain an integral and integrated part of the course throughout its duration, despite the difficulties involved with abstraction of newly sworn officers. These days in the HEI provide the opportunity for students to discuss, reflect and theorise about their practice with their peers and tutors as well as form part of a professional community. 5. Regular time must be allocated with both academic and practice tutors/ mentors, with clearly defined roles for what these tutors are responsible for. Their role in aiding students in linking their practice with theory should be made explicit. Tutor constables need to be willing and able to work with academic tutors and with the HEI. Similarly, academic tutors should be willing to be part of the field training and understand the context of operational policing. Despite being resource intensive, tutors need to be provided with adequate mentor training and ongoing support and be clear on their role in supporting the development of a professional police officer. 6. Assessments need to involve the integration of theory and practice as well as be assessed collaboratively by the HEI (academic tutor) and the tutor constable. 1 1 7 Strong partnerships 7. For policing, it is vital that partnerships are formed, developed and maintained through formal agreements and memorandums with in-built quality assurance mechanisms. Collaboration and coordination between HEIs and police forces ought to be maintained at all stages of the process, from course design, student recruitment, delivery, and assessment, with clear ongoing governance arrangements. 8. Course consistency is a key concern for policing as a profession, given the varied contexts of policing, size and resources of police organisations, and quality of HEIs across the country. It is important that areas draw from their context to make the course relevant to the community, but at the same time are also broad enough to develop general skills to be able to work across contexts. 9. Any conflicts that arise between the HEI and the practice institution, in regard to professional standards, should be resolved and reconciled through the partnership. 10.	The professional body, in this case the College of Policing, will play a key role in maintaining course consistency and professional standards. City Hall, London SE1 2AA Enquiries: (020) 7983 4100 Minicom: (020) 7983 4458 www.london.gov.uk © May 2018 p ,QLJ ,CE MAYOR OF LONDON OFFICE FOR POLICING AND CRIME A Problem Profile of Violence, Gangs And Young People MOPAC Evidence and Insight September 2022 METROPOUTAN ■ ■ ■ ■ ■ ■ . ■ ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Executive summary – the key context • The Police and Crime Plan sets out the key priorities and objectives for policing and community safety in London over the next three years (22-25). • Reducing and preventing violence in all its forms is the Mayor’s first priority for this term. • Violent crime takes an appalling toll on our city, inflicting terrible harm on victims, devastating families and traumatising communities. • Police recorded violence against the person began to increase both in London and across E&W in 2014/15 largely related to improvements in recording practices. • Some violent crimes in London had started to fall before the pandemic – for example knife crime resulting in injury and gun crime. • However, the number of teenage homicides in 2021, the highest on record of 30, shows how much more there is still to do to ensure Londoners are safe. • An evidence-based approach is a core ingredient in tackling violence. ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ .ONDON IG AND CRIME Executive summary – the key findings The profile updates performance insights, such as the long term trends of youth violence, borough variation & change, locations, public perceptions, the role of robbery and change post CV-19. • We see disproportionality across youth violence. Over half of weapon enabled robbery (59%) and homicide (65%) suspects were Black. Victim profiles were more diverse (i.e., 56% of weapon enabled robbery victims were white) but the majority of youth homicide victims were male (93%) and Black (61%) (ethnicity data uses police observed rather than self reported). • Pre-pandemic increases in knife and serious youth violence were wholly driven by robbery. • Deprivation metrics (i.e., IMD, food insecurity), school suspensions/absence, and low youth employment were all predictors of most serious youth violence. • The cultural sites related to weapon enabled robbery and most serious violence were parks (with Skate park) and Legal graffiti walls. As the numbers of these increased, the number of victims increased. The current profile is the largest product of its type to be published – covering a multitude of topics. These highlight a wealth of insight, many of which confirm what was already known, many are new. Data indicates the offending history of many victims and suspects of violence as well as the importance of issues such as education, gangs and weapon type. • Approximately 23% of all serious youth violence suspects were already known to police. This proportion increases as the offence escalates in severity. • 63% of victims and 87% of accused of teenage homicide had been previously arrested by police. • From available data, 23% of victims and 35% of accused of violence were not in full time education, BUT this increases to 56% of accused when only looking at under 18s. • Analysis showed 20% of those accused for teenage violence were RUId for an offence in the six months prior to the violent offence. For homicide accused, this was 25%. • Gangs are related to violence especially at the more severe end, with over half of all teenage homicides in the last 3 years had links to gangs. • 82% of teenage homicides committed in the last two years involving the use of a combat/machete/Rambo style knife. Deep dive coding was undertaken to explore typologies, motivations and attrition through the system of 300 cases of youth violence. • Two main typologies of violence emerged: ‘single suspect’ and ‘group of suspects’. Single cases were less likely to use a weapon, be brief, be between acquaintances and more in a private dwelling. Group cases were more likely between strangers, outside, with a weapon and be prolonged in nature. • Where suspects or the victim had been a previous perpetrator and there was a larger number of suspects were significant predictors of moderate to severe injury. Gang suspicion was the strongest predictor of there being multiple suspects. • Half of the victims of violence withdrew from the investigation and just over a quarter of these withdrew on the same day as the case was reported. • Witnesses/bystanders being present, injury and weapon usage were all predictors of the case progressing through the system. • A significant predictor of the police NFA-ing a case was when they expressed doubt over victim credibility (6x more likely to NFA in these case). ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Background and aims This project seeks to improve current understanding of gang and group violence in London. This has been led by the MOPAC Evidence and Insight (E&I) team and involved a wide-ranging research package covering: 1) Updating the picture of Youth Violence • Updating the literature and how this can be communicated to practitioners. • Refreshing the problem profile of youth violence across trends, places and people (i.e., public perceptions, the long-term trends of youth violence, borough variations & change, geography, wider ‘drivers’ associated with violence, as well as the individual characteristics of victims and suspects). 2) Understanding the joins and gaps across the violent group • The aspect explored the joins and gaps of a specific youth violence group across MPS data sources to explore issues relating to intelligence and disproportionality. 3) Exploring the context behind the violence • Coding of a sample of the violent cohort to develop a richer understanding of violent incidents (drivers/needs/situational dynamics) and aspects such as attrition and types of violence. ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Data sources and methodology Updating the picture on youth violence • Crime Reduction Toolkit | College of Policing – a review of what works in policing, filtered for youth violence. • Youth Endowment Fund Toolkit – a review of most available evidence of interventions for the prevention of youth violence. • Locally developed tools and responses | London Councils – a collection of the best evidence for youth violence prevention as applied by individual Local Authorities in London. • Evidence and Gap Map | Youth Endowment Fund – an evidence gap map on youth violence prevention research, highlighting the current quality of research and field of focus. The aim is to guide better research and a focus on underserved aspects of youth violence. Understanding the joins and gaps within the violent cohorts • Cohort Analysis - Base data - All CRIS records between January 2019 and February 2021 inclusive, with the following parameters: current classification is Violence with Injury, Domestic Abuse Incidents Excluded, At least one victim, suspect or accused is U25. • Matrix datasets - Gangs Violence Matrix data from September 2021. - Priority Robbery Offenders, Habitual Knife Carriers, Priority Firearm Offenders data extracted - July 2021. • Police Recorded Crime data – review trends over time by offence types, victim and offender demographics and geographical differences in offending. - Offenders: this refers to those charged with a serious violence offence against a young person aged 1-24. - Ethnicity: the ethnicity data used in this report is observational data recorded by the police in the relevant crime reports. Although the police do also collect self-reported ethnicity data, this is not sufficiently complete for a population analysis. • ONS crime data – used to analyse national and Police force level trends and differences. • Public Attitude Survey – analysis of Londoners perceptions of violent crime over time and by demographic. • London Datastore – was used to assess the demographics and landscape of London. These include Central Trend Population Projections, Indices of Multiple Deprivation aggregated to a ward level and locations of Cultural Venues across London. This data was used to provide further context in addition to police recorded crime. Exploring the context behind the violence • Coding – 282 cases of SYV offences taken from police systems (CRIS) were coded across 95 variables in 5 areas. Cases were randomly generated from Jan 2019 – Feb 2021. Cases were selected from Tower Hamlets (95), Southwark (103), Croydon (84). The selection was determined by considering the crime counts for GBH & GBH with intent, along with a measure of public health deprivation. The work also draws from a specific MPS problem profile of teenage violence. The Probation Service undertook a review of cases and provided data to support the research. MAYOR OF LONDON OFFICE FOR POLICING AND CRIME 1. Updating the picture on youth violence - Updating the literature and how this can be communicated to practitioners - Refreshing the problem profile of the nature of the violence (trends, places and people) MAYOR OF LONDON Serious Youth Violence: Signposting established & emerging interventions. ......,,_ __ ,_,_., D CRIME uty ~ (MOPAC & lht V11UI ••t comrrutted to m.1kq ew;ienct~COff\fMSIOnSIC ¥Id~ dfoslons, pthfma btst prxoct 10 lflform future tundifl . PM1 of thl1, IS lo undtfsl.Nid 'WNl worts' on tht ground, CommtSSiOnlfll lrllfflffltlons md PfoCf ~mtS wrth both ¥1 HU~ •nd nowt M:ttnce bast, to undtnUnd if good pr.MtlCe an be rtplc•ttd in difftftnl SftbngS o, If f'mfflCe pps un bf filtd throu&h ~ more 1olluM ...,.,,,.. .... Thsdocumtrt imilo wgn ~l arsbl1 toolt.rtsMld Mdfflct ,awwsaf Strious VoulhVIOltntt Miri them•bC MUS¥ld ~.u,~fd, Wfth the1nttobonolduKtlf11prXbtionNstolht~•of~tsUIIIMed¥ldntW/nowA11tffl'tf\tlOmwhK:h~ollllteJm.ln~boncommoo 'fis.t" prOlffl>Yt lxtors' chwn from the ~ltfoitUff Mt hi£hliahttd MAYOR OF LONDON OFFICE FOR POLICING AN D CRIME Serious Youth Violence: Signposting established & emerging interventions . Risk Factors factors thousht to put an lndl,..ldual at an Increased risk of committing vlolence. They can occur In Isolation, but are often multiple & complex In Below is an adapted list from the Early Intervention Foundation systematic review on risk factors for youth participation in violence . The coloured dots indicate which intervention theme(s) address corresponding risk factors. Links were determined by the primary in u.(s) addressed by each intervention, meaning the link is not exhaustive as to which risk facto r is addressed by each intervention category. The primary purpose is to highlight risk factors which currently appear 10 be unaddressed in a focused manner by current interventions. Risks are split across five areas below. 5c>u,-u; taTN!Officefunded Efr1y 1ntny,;npon fgyndf!i9n (Elf )prcjectgn,pkfpmlrotr£ftd2!l Community Norms of acceptability of Yio1ena, • Hou<ln,:lnstabillry Poorn,,,rt,bourhood< Gancactivitv • • Peer o..linQ.....ipttr, • CommltmNltto dielinquer,I Pff1'S l'oorrNtion!IWithpttr, e Individual Agrnsion . lmputsiveoe,s e e R~, e Key: Intervention Colours Aversion Progr;,mmes • DiversiooProgrammE'S After School School Based Care • PositiveRoleModels e ReshapingBeh.avioun e Public Health Approach e f orging Networks family & Parenting Programmes • Emerging Risk Factors Tbesearr the topfiwfadorslinkrd ta)Qlthlliolenaidenti~byUKlocal Autharifies (118 &un,eyed}as part of their )Qlth Vlokncea5Sewnents. ChildrcnwithpoormNlt.lhealth(96] Childrcne•~$Ub<tance mi<u"'(84) Childrcninallousmoklwithl)llreni.isubst..-.ce fflGU"'(6B) Childrcnin a llousmokl~rw:lnf:dcime<rlcw>lence(66) Childrcninallousmokl~ncin1""11ectl6S) otherri<binduclehouwc,nnabiity{SO),bei111outsidl!of,.,.;nstr""'Khool(as), pr..vk>u<Yicliml<a!lonor aMi<l<of e~rc"'xualexploitationj47).ml<<lrc lncl­ dcnt< (36),prMou<wtlmisa!lonotat-lidof e,cperlcnl;ln,:crimlnalexploitM>On(25I NKl<harinc•hou<mO!dwithapr-.lyconvicte<loffendtt(ll) Schoo/ Truancy e e Family ~ family Runnlrca-from ~ e Ant½oca.l/crimi,,_11)11,cnts Umbrella term A tl'rm which MCompaSSl'S indMdual riskfoctanforchlldll!n&youngp«1plf', partirolarlydurlngeorly df'wlopmf'nt. Ad'ttna!ChiklhoodE,cpe,iencnj.-.cb) ACE<a,e u«.d toa«n<heallhrl<bunder the public health •ppro«l',-J)Mticula,ly .....,.<i,.,.rt<l<ofyouthlllolence(asavlctlm andperpei..torl 1!lli. presenting the main thematic findings and links. The product updates the literature on What Works evidence as well as risk and protective factors. Work was undertaken to refresh the literature pertaining to youth violence and emerging interventions. The aim was to both update the literature as well as design a product that can be shared with front line practitioners to serve as a guide Updating the literature and sharing learning Click here for the product: The main groups of evidence: • Diversion & Deterrence (i.e., diversion or wider support) • Safer Neighbourhoods (i.e., interventions supported by community / schools) • Social Skills & Mental Health (i.e., role models, mentoring) • Partnership Working (i.e., public health, networks) • A supportive home (i.e., family and parenting) As well as an updated knowledge of risk and protective factors: • Individual (i.e., aggression, drug use) • Family (i.e., disrupted, poor supervision) • Peer (i.e., delinquent peers) • School (i.e., truancy, poor attainment) • Community (i.e., norms, crime) Against the Person 233,450 ! 6.0% Metropolitan Police 2,017,307 &13.3% England and Wales 1,783,857 ! 14.3% England and Wales (excl MPS) ! 32.6% Most Similar Forces - --· - LONDON OFFICE FOR POLICING AND CRIME The long term violence picture in London is improved compared to England and Wales and Most Similar Forces Violence There was a 18% reduction in Robbery in the MPS, which remained stable in the rest of England and Wales but increased by 5% in the MSF. Total Theft Offences remained stable, in line with the rest of England and Wales, however this increased in the MSF (+5%). In the year to December 2021, there was a reduction of 7% in knife crime. This is more than the other National Data Quality Improvement Forces and MSF - both remained stable. In the year to December 2021 there was a reduction of 18% in firearm offences. This is a larger decrease than the rest of England and Wales and MSF, which both remained stable. Violence increasing more outside of London and across similar forces In the year to December 2021 the MPS recorded 233,450 Violence Against the Person (VAP) offences, which accounted for 12% of all VAP recorded offences in England and Wales. VAP increased by 6% compared to the previous year, while the rest of E&W saw far larger increases (+14.3%) with the Most Similar Forces (MSF) recording the largest increase (+ 32% increase). Violence with Injury – Rolling 12 month trend (Normalised Figures) England and Wales (excl MPS) Metropolitan Police Most Similar Forces 3.5 2.5 1.5 0.5 -0.5 -1.5 -2.5 -3.5 1st National Lockdown Jun-16 Sep-16 Dec-16 Mar-17 Jun-17 Sep-17 Dec-17 Mar-18 Jun-18 Sep-18 Dec-18 Mar-19 Jun-19 Sep-19 Dec-19 Mar-20 Jun-20 Sep-20 Dec-20 Mar-21 Jun-21 Sep-21 Dec-21 Source: Crime in England and Wales: year ending December 2021 - Office for National Statistics (ons.gov.uk). - 77% 66% 64% 59% 57% 57% 54% 56% 45% 40% -------- Q3 20-21 Q4 20-21 Ql 21-22 Q2 21-22 Q321-22 Q4 21-22 - worried about crime --Gun crime problem -- Knife crime problem --Gangs problem I f 100% 80% 60% 40% 20% 0% 11 16to 24 25 to 34 ■ Worried about crime ■ Gun crime problem ■ IMD Group 1 {MOST deprived 25%) IMD Group 2 ■ IMD Group 3 ■ IMD Group 4 (LEAST deprived 25%) I Worried about crime Gun crime problem 35 to 64 65+ ■ Knife crime problem ■ Gangs problem Knife crime problem Gangs problem MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Many adult Londoners are worried about violence, and concerns have been increasing *Across both FY 20/21 and 21/22 the MOPAC Public Attitude Survey (PAS) interviewed 12,800 London adults per year. The proportion feeling violence-related issues were a problem has recently increased – although worry about crime overall has remained more stable. Discrete quarterly data In Q4 21-22, three-quarters of Londoners felt knife crime was a problem locally, while two-thirds felt gangs were a problem. Londoners from Asian Ethnic Backgrounds are more worried about crime overall, while those from Mixed Ethnic Backgrounds are most likely to feel knife crime and gangs are a problem. The table below shows gaps compared with the MPS result: Younger residents (aged 16 to 24) report similar levels of concern about knife crime and gangs as other age groups… …while older residents (aged 65+) tend to be least concerned about such issues. However, area deprivation appears more influential in shaping concerns. Londoners living in more deprived Wards are far more likely to feel violence related issues are a problem. This holds true even when controlling for differences in age and ethnicity*. Weighted MPS Result White British White Other Mixed Asian Black Other Worried about crime 58% -4pp -4pp. -4pp. +10pp. -5pp. +6pp. Gun crime problem 43% 0pp. 0pp. 0pp. -2pp. +3pp. -1pp. Knife crime problem 73% +1pp. -2pp. +6pp. -2pp. -2pp. +3pp. Gangs problem 60% -1pp. -2pp. +5pp. +2pp. +3pp. +1pp. * The PAS upto and including 21/22 sampled a representative sample of 12,800 adult Londoners per year. Data above combine ‘major’ and ‘minor’ problem. Please note that the PAS was moved from face-to-face interviews to telephone interviews as a result of COVID­ 19: results here are from the telephone methodology. For more information see Taking part in MOPAC's surveys | London City Hall. *Results from regression models predicting issues a ‘problem’ from Deprivation Groups whilst controlling for Ethnicity and Age groups. Deprivation Groups relate to IMD scores at Ward levels grouped into quartiles. - ' I , il Hate crime People carrying knives People joining gangs People being violent People stealing things People using or dealing drugs ■ ... In your local area ■ ... At school MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Concerns about violence were also prominent amongst young Londoners MOPAC’s Youth Survey 2021-22 gathered the views of nearly 12,000 young Londoners aged 11 to 16 on the crime, safety and policing issues that affect them. Exposure to violence impacts on young people’s safety. Less than half of those who had been ‘pressured to carry a knife’ or who had been ‘threatened or hurt with a knife’ said they felt safe locally. 88% …of young people feel safe at school 76% …feel safe in the area where they live 10% … is in a gang 14% … has carried a knife Do you know anyone who … Although most young people felt safe, many were concerned about violence – particularly in their local area (see below). Exposure to such issues increased with age. For example, those aged sixteen were nearly three times more likely to know someone who had carried a knife (21%, vs 8%) and five times more likely to have been threatened or hurt with a knife (11% vs 2%) than those aged eleven. Smaller differences were seen by Ethnicity. Young people from Black (15%) or Mixed (14%) Ethnic Backgrounds were more likely to know someone in a gang (vs. 10% of White and 6% of Asian), but no differences were seen for having personally carried a knife. 2% … asked or pressured to carry/hide a knife 5% … threatened or hurt by someone with a knife Have you ever been … 43% …agree the MPS is an organisation they can trust 31% …agree the police do a good job locally Young people’s trust and confidence in the police are low. Confidence remains in line with levels seen in MOPAC’s 2018 Youth Survey (31%) but results are below those seen for adult Londoners (Confidence: 49%, Trust: 73%). A minority of young people said they knew others involved in violence or had personally been involved themselves. 3% … carried another type of weapon 2% … carried a knife 1% … belonged to a gang Have you ever … Youth Survey was hosted online between November 2021 and January 2022, with a total of 11,874 young people living and/or going to school in London taking part. *Data for adult Londoners drawn from MOPAC’s Public Attitude Survey for FY 21-22. However, please note that methodological differences between the PAS and MOPAC’s Youth Survey limits direct comparisons for these measures. Question wording for good job local was on an agree/disagree scale for the Youth Survey, and analysis here includes those who said ‘don’t know’. % feeling ‘big’ problem Young people feel tackling gun and knife crime should be the TOP PRIORITY for police in London - - - MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Youth homicide offences are higher than pre-COVID levels Key Categories Volume Compared to (% change) Volume Compared to (% change) 2021 2020 2019 Q3 2021/22 Q2 2021/22 Q3 2020/21 Q4 2019/20 Homicide Offs 132 0.8% -11.8% 34 -25.0% -6.3% -3.2% Youth Homicide Offs 37 17.9% 3.1% 9 -14.3% -33.3% 20.0% Attempted Murder Offs 170 -7.5% -4.2% 45 -29.8% -25.0% -13.2% Serious Youth Violence 5,898 -11.7% -33.6% 1,496 4.4% -5.9% -26.4% Youth Violence 14,305 -0.4% -23.1% 4,082 10.0% 3.4% -12.3% Knife Crime Offs 10,638 -8.5% -30.4% 2,510 -3.8% 6.1% -22.4% Knife Crime With Injury Offs 3,163 1.0% -18.8% 799 -14.8% 7.0% -2.1% Knife Crime with Injury Pers Robb Offs 421 -17.0% -31.8% 95 7.2% -11.1% -7.1% Knife Crime Injury Victims 1-24 1,355 -0.2% -25.6% 343 -15.1% -2.8% -10.4% Knife Crime Injury Victims Not DA 1-24 1,250 -0.2% -27.3% 320 -18.9% -7.1% -11.9% Hospital Admissions 945 -6.9% -22.2% 210 -26.5% -12.2% -14.3% Hospital Admissions (U25) 485 -5.4% -26.3% 105 -26.1% -19.0% -22.7% Knife Possession Offs 4,328 -11.2% -17.5% 950 -7.4% -15.1% -18.5% Gun Crime Offs 1,545 -20.2% -34.8% 332 -4.0% -9.5% -29.1% Gun Crime Lethal Barrelled Discharged Offs 260 -30.3% -20.7% 56 -24.6% -18.3% -35.5% Non Domestic Abuse VWI Offs 45,448 8.3% -8.5% 13,070 -4.0% 25.7% 12.6% Poss Drugs Offs 44,779 -18.8% -7.2% 9,950 4.3% -17.6% -17.9% Robbery of Personal Property Offs 22,009 -17.4% -43.5% 4,940 7.4% -1.5% -28.6% Increase in Youth Violence in the latest quarters but still less than pre COVID. Decrease in Knife Crime with injury victims in the latest quarter yet similar to pre COVID. Assault with sharp object hospital admissions (victims u25) have decreased when compared to the previous quarters and pre COVID. Youth (1-19) homicide has increased by 18% in 2021 compared to the previous year and by 20% when compared to the pre-COVID period. 23% of homicides were for teen victims – a higher proportion that 2020 and 2019. In Q3 21/22 Weapon-enabled offending decreased compared to the previous quarter across most areas, with volumes still less than pre-COVID. Gun crime lethal barrelled discharged decreased by 25% this quarter. Youth violence has recorded an increase in recorded victims in the latest quarter, but remain below pre-COVID levels. Personal Robbery increased compared to the previous quarter yet less than the previous two years; it remains lower than the same period pre-COVID. Between January and March 2022 there have been 3 teenage victims of homicide, compared to 10 in the same period in 2021 and 7 in 2019. Source: MetStats2 (MPS Database) – Police Recorded Crime, Hospital Admissions (https://digital.nhs.uk/data-and-information) and MetStats2 (MPS Database). ----- ----- ----- ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Gang involvement would appear to be decreasing in youth violence • Examining gang involvement 2017 through 2021 within youth violence suggests that generally gang involvement appears to have been declining. • 15% of all homicides in 2021 were believed to be linked to gangs - this is significantly lower than 2017, where 27% of homicides were linked to gangs. • However, as seen later the MPS profile indicates gang involvement in teen homicides is far larger (e.g., over half of teen homicides with links to gangs). • Lethal barrelled discharge (40% - 34%), knife injury under 25 (21% - 12%) also saw a reduction in gang involvement from 2017 to 2021. • Whether this decline is an actual finding or an artefact of terminology, definitions and/or data capture is unknown. 2017 2018 2019 2020 2021 All Homicides 136 133 150 126 131 Gang related 27% 34% 22% 13% 15% Homicide (Non-DA & Non-CT) 115 105 134 104 107 Gang related 32% 44% 25% 15% 19% Lethal Barreled Discharge 354 433 214 298 186 Gang related 40% 49% 38% 59% 34% Lethal Barreled Discharge - Victim Shot 144 142 75 101 92 Gang related 49% 63% 52% 68% 43% Knife Injury under 25 (Non-DA) 2,138 1,827 1,271 1,198 1,342 Gang related 21% 19% 15% 11% 12% Homicide Firearms Knives Crime Type Source: Data provided by MPS which includes Gang flagged crimes on CRIS, together with data from the Gangs Violence Matrix and GRITs. - - - - - MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Personal robbery offences drive trends of youth violence and overall knife crime Trends in MPS recorded violent crime - Normalised Rolling 12 month Offences of knife crime, personal robbery and 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -3.0 Gun crime offences and victims of non-domestic knife crime with injury (aged under 25) were already on a downward trend before Covid-19 related restrictions commenced. Lockdown restrictions begin serious youth violence were increasing pre­ pandemic. During the past three years 60% of all knife crime and 46% of all serious youth violence offences were recorded as weapon enabled personal robbery. Excluding robbery victims from SYV and knife crime shows these crimes were on a downward trend before pandemic restrictions were applied. Normalised rolling 12 trends – Excluding Robbery Knife Crime excl. Robbery 2.0 1.5 Serious Youth Violence excl. Robbery 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 Dec 16 Feb 17 Apr 17 Jun 17 Aug 17 Oct 17 Dec 17 Feb 18 Apr 18 Jun 18 Aug 18 Oct 18 Dec 18 Feb 19 Apr 19 Jun 19 Aug 19 Oct 19 Dec 19 Feb 20 Apr 20 Jun 20 Aug 20 Oct 20 Dec 20 Feb 21 Apr 21 Jun 21 Serious Youth Violence Knife Crime Offs Knife Crime Injury Victims Not DA 1-24 Gun Crime Offs Robbery of Personal Property Offs *Identified with either a GA flag, a member of the current GVM as the suspect or victim or feature within GRITS. Source: MetStats2 (MPS Database) – Police Recorded Crime. - 0 1 2 3 4 5 6 8 9 11 12 13 14 15 16 17 18 19 Weekday Weekend Weekday Weekend 20 21 22 23 Homicide Mon - Tue - Wed - Thu - Fri - Sat - Sun - MSV Mon- - Tue - Wed - Thu - Fri - Sat Sun - Robbery Mon - - Tue - Wed - Thu - Fri - Sat Sun - ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Youth violence peaks during weekdays and late afternoon/early evening Higher volume Lower volume Homicide MSV Robbery Across the selected offences, the offending peaks predominately during weekdays and rises from 3:00pm onwards. Homicide peaks later in the day. This pattern is consistent across boroughs. Nearly a third of all offences happened within a four hour window between the hours of 3pm-7pm. Robbery and MSV are especially focussed within these times. Nearly 40% of all group offences (36%) and repeat offences (37%) happened during this period. Source: Violent Cohort CRIS data extract. Borough ranking over time Rob bery of P<>rsonal 511:!rious Youth Knif <> Crime Injury Property Offe nc<>s Knit.. Crim<> Off@nces Viole nc<> Gun Crime Offences Victims Not DA 1-24 OCU Name 201 9 2020 2021 201 9 2020 2021 2019 2020 2021 2019 2020 2021 2019 2020 2021 Westminster ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Haringey ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Southwark ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Hackney ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Newham ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Enfiel d ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Lamb<>th ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ C.omde n ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Tower Haml<>t s ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Isli ngton ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Croydon ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Br<>nt ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Lewisham ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Ea li ng ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Barn<>t ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Barking & Dagenham ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Gr~enwich ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Walt ham Forest ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ Wandswo rth ■ ■ ■ ■ ■ ■ ■ ■ Redbridg<> ■ ■ ■ ■ ■ ■ ■ ■ Hounslow ■ ■ ■ ■ ■ Hamm<>rsmit h & Fulha m ■ ■ Kensington & Ch<>lsea ■ Hillingdon ■ Harrow ■ Hav<>ring Broml@y ■ Merto n Rank of volume Bexl@y Rich mond upon Thames 10 Sutton Kingston upon Thames • 20 • 32 MAYOR OF LONDON OFFICE FOR POLICING AND CRIME The borough offence landscape has changed over the last two years, although some boroughs remain consistently high * Ranking borough offence levels over time show a shift in where some offence types are most likely to take place Robbery offences: Westminster remains the highest ranked and always has been – even with the current reduction in volumes. Knife crime offences: In 2021, Tower Hamlets recorded the largest upward move (+8 places) in ranking and was the highest volume borough. Serious Youth Violence: In 2021, Croydon was the highest ranked and recorded the largest increase (+4 places). Westminster saw the biggest drop (-10 places) since 2019. Gun crime: Lambeth recorded the highest volume in 2021 and has always been highly placed. Newham has moved up 5 places and Haringey down 6 since 2019. Knife crime with injury victims not-DA (1-24): Within 2021, Tower Hamlets recorded the largest upward move (+11 places) and was the highest volume borough. * The borough ordering on the table reflects the total volume across all of the described offence groups. Source: MetStats2 (MPS Database) – Police Recorded Crime. Higher volume Lower volume Chart Legend: ■ ■ ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Decreases in weapon enabled robbery have impacted on SYV in some boroughs more than others Over 90% of all serious youth violence (SYV) were recorded as either personal robbery or serious wounding during the last three years. Reductions in personal robbery (post­ pandemic) have impacted on some boroughs more than others. Boroughs experiencing the largest decrease in SYV volumes i.e., Westminster, Enfield and Haringey have been driven by reduced levels of weapon enabled robbery (2020 vs 2019). SYV has decreased in Enfield – driven by reductions in weapon enabled robbery; however, serious wounding offences have actually increased between 2019 and 2020. Tower Hamlets has recorded an increase in victims of serious wounding. Weapon enabled robbery, total Serious Wounding, total Source: SAP BI (MPS Database) – Victim data. Deprivation Levels ■ Very High ■ High Medium D Low ■ Very Low MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Youth homicide is largely unpredictable, however deprivation levels are a driver Wards with at least 1 youth homicide victim (April 17 – March 21) Youth homicide offences occur across a number of wards Between April 2017 and March 2021 there were 221 homicides recorded in London with victims aged under 25; occurring in 30 boroughs and 169 wards. Newham (18), Greenwich (17), Southwark (15), Croydon (15), Haringey (13) and Enfield (11) had the highest number of homicide victims in that period. Greenwich West (a ward in Greenwich) recorded the most homicide victims aged under 25 in this period, with 4 homicides. Links to deprivation There is a clear difference in where these homicides have taken place, as it is more likely to occur in more deprived areas (62% in Very High/High v 23% in Low/Very Low). There was no correlation between homicides with cultural venues or schools. A logistic regression was used to better understand the relationship between youth murders and weapon-enabled crime, in the wider context of local demographics, deprivation and other area characteristics. Only youth population and deprivation in a Ward were associated with increased odds of seeing a youth murder. Source: SAP BI (MPS Database) – Victim data (April-17 to March-21); IMD (ward level) – London Datastore (GLA); Cultural Venue London Datastore (GLA). MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS analysis replicated these findings focussing upon certain wards • To identify key locations for offending data was obtained from MPS Crime records and London Ambulance Service (LAS) call outs that featured teenaged victims. • Over the three year period 2019-2021, MPS crime data shows that Croydon has suffered the highest volumes of teenage violence (141), followed by Lambeth (137), Tower Hamlets (135), and Southwark (125). • In 2021, Tower Hamlets recorded the highest volume of teenage violence (49), followed by Lambeth (47), Croydon (46), Lewisham and Southwark (both 35). • The wards with the highest volume of offences in 2021 were Fairfield (Croydon, seven offences) and Edmonton Green (Enfield, seven offences). Coldharbour (Lambeth), Brockley (Lewisham), Broad Green (Croydon) and Woolwich Riverside (Greenwich) - all recorded six offences. • Overall 31% of total offending is committed in locations that feature within the most deprived 20% of Lower Super Output Areas (LSOAs) in the country. 5% of teenage violence is committed in the most deprived 10% of LSOAs. • Six wards feature in the top ten volume wards for both LAS and MPS Crime records; Leyton (Waltham Forest), New Cross (Lewisham), East Acton (Ealing), Woolwich Riverside (Greenwich), Brockley (Lewisham) and Fairfield (Croydon). • Two wards, Bishops (Lambeth) and New Cross (Lewisham) both feature as the second most common location for LAS call outs but are not in the top 10 for MPS recorded teenager related crimes. Source: MPS analysis. MPS Crime Data – Teenage Violence LAS Data – Teenage Violence "'O "'O ro Q) E 0 I ~ Q) u ro a. V') u ~ :, ~ • e / ; \ offen bei corn type. for ho pronounc horn cides ( loca ·ons. Offences co are most p . 2090 street (5 19 eenag otal teen in 2021. Offences co · ere ·g 19 was an incr 2 o 2020 (21-12-18). or commons, incl ho cides (17%). e e :, C Q) > L. ..c. ... 0 Overall there have been 29 teenage ·o1en offences committed ,n other enues. This includes one t~eh ·de. 5 of these offences have been committed in shops, rangi g rom convenience stores, supermarkets and shops four of these offences have been committed in Hospitals. Three offences have been Three o ences have been c ·ned ,n schools. MAYOR OF LONDON OFFICE FOR POLICING AND CRIME The majority of teenage homicides occur in public spaces Over the last three years (2019-2021) Croydon (7) and Newham (6) have suffered the highest volumes of teenage homicide. Haringey, Southwark, Greenwich, and Westminster have all suffered five offences over this time period. When homicide offending in 2021 is looked at in isolation Croydon (5) has had the highest volume of teenage homicides, almost double the volume of the next highest volume boroughs (Greenwich and Haringey, both 3). ~Q All teenage violence thematic map with 2021 teenage homicides shown on top. Source: MPS analysis. Factors associated with rates of victimisation and offending in London boroughs Benefits I Heath I Food I School I Employment Percentage of those claiming 'out-of-work' benefits Risky health behaviour (as indicated by area rates of chlamydia detection) Percentage of secondary school children suspended from school Low employment rate of 16-24 year olds Area has deprived IMng envirorvnent (as measured by the IMD) • Victims Uvmg Environment • Offenders Percentage of those claiming 'out-of-work' benefits Food Insecurity (as Indicated by percentage of people struggling to access food) Percentage of those claiming ·out-of-work' benefits Overall absence rate in secondary school . E ~ • u , ! < ,3 0-10% (most dtpriv•dl 11-20% 21-30% 31-40% 41 -50% 014 001 Ill 015 51-60% 013 001 102 010 61-70% 71--80% 81-90% 91-100% O••st dep,ivod) ~ .i, ,f ,: = c i -= " ► ~ 037 056 001 067 019 043 002 054 ~ , ! 2854 903 238> HJ/ d 0 • a. . .. Ii .. 078 059 ,: i ► 161 142 ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Analytics illustrate the social drivers of youth violence Analysis conducted demonstrates the linkages between deprivation, wider social factors and serious violence. On the left are analytics conducted by GLA City Intelligence and MOPAC E&I reporting structural factors. Rates of offending are strongly associated with different measures of deprivation and poverty, including areas where people are most likely to struggle to access food. Modelling found absence rates from secondary school were also a significant factor in predicting the Boroughs with the highest rates of offending for serious violence. On the right is a pure focus on levels of deprivation, showing the rates of violent crime split by deprivation (for the year of 2021) with the chart illustrating clearly the association between deprivation across violence. Lethal Barrel Discharge Violence Against the Person For more information on data sources – see - Serious Youth Violence Regression Modelling (GLA City Intelligence Unit, 2021), Understanding serious violence among young people in London - London Datastore. # Cinemas or Museums or Theatres Ward vulnerability decile # Private indoor music venues # Schools # Music or Nightime or Pubs Ward population -0.375 -0.345 -1 Impact of 1 unit of *variable* on# of *victim* victims ■ ■ ■ ■ ■ ■ 0.75 0.919 1.029 0.532 - 0.45 - 0.239 0.144 ■ 0.189 0.001 lo 1.61 2.349 MSV ■ 5.409 - ■- ■ ■ ■ ■ ■ LONDON OFFICE FOR POLICING AND CRIME Exploring more detail - certain cultural sites were more important for youth violence The GLA publishes data on London's’ cultural infrastructure including a detailed map of the locations of sites across London (https://maps.london.gov.uk/cim/index.html). Data was used to identify those with the strongest association with youth violence. The cultural sites related to Weapon Enabled Robbery and Most Serious Violence (MSV) were: • Park (with Skate park) • Legal graffiti wall As the numbers of these sites 0.00 increased, the number of weapon enabled robbery or MSV victims 0.00 increased *. Weapon enabled robbery and MSV model - Linear Regression - Coefficient used ­ Data covers Apr 20 to Mar 21. 0.00 As the number of community centres and Archives/ Libraries increased the number of weapon enabled robbery or MSV victims decreased *. * This is not to say these sites cause or that all those individuals that attend these sites are involved in criminality – only that the number of sites is related to the overall volume of offences. Cultural sites include: cinemas, museums, parks, music venues, pubs, art centres, community centres. All categories of site were included in analysis. The above excludes non-significant ones i.e., ‘Park’, ‘Arts or Creative or Culture”. Sites are taken from the https://data.london.gov.uk/dataset/cultural-infrastructure-map. Weapon Enabled Robbery MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights on teenage violence - offending higher in the summer months Slides 22 – 30 draw from analysis conducted by the MPS for all incidents of Many violent offences usually see peaks during the summer months (i.e., total stabbings, lethal barrelled discharges and homicides where the victim Violence with Injury, Knife Crime with Injury and Gun crime). of the offence that received injuries was aged between 13 and 19 in the previous three years*. For the purposes of the report, figures relate to a victim count and not an offence count. Category 2019 2020 2021 Total Period Total Offences/victims 907 683 678 2,268 Homicides 26 15 30 71 Lethal Barrelled Discharges 12 9 17 38 Stabbings 845 636 625 2,106 Stabbings (Att. Murder & GBH) 517 350 332 1,199 Robbery Injury 110 83 69 262 Fatal 26 15 30 71 Serious Injury 343 232 234 809 Moderate Injury 223 189 166 578 Minor Injury 316 247 224 787 Repeat Victims 44 Accused 271 204 194 669 2013 2014 2015 2016 2017 2018 2019 2020 2021 Violence with Injury Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 4517 3943 4537 4510 4700 4793 5345 4847 4475 4881 4945 5109 4971 4553 5578 5036 5826 6268 6269 5937 5975 6092 5989 5856 5544 5077 6004 5768 6390 6384 6580 6164 5952 6384 6247 6323 5947 5529 6031 6044 6901 6542 6984 6484 6262 6294 5753 6232 5675 5616 6736 6382 7048 7104 7168 6306 6289 6897 6376 6419 6014 5230 6354 6187 7086 7108 7554 6204 6204 6427 6320 6544 6162 5758 6621 6336 6807 6847 7126 6730 6294 6274 6473 6067 5912 5658 5849 4670 5595 6181 6547 6731 6209 5581 5250 5135 4383 4311 5277 5712 6489 6910 7044 6207 6523 6650 6333 6087 2013 2014 2015 2016 2017 2018 2019 2020 2021 Knife crime with Injury Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 255 219 254 268 258 287 304 315 273 257 257 268 258 235 224 254 297 307 281 336 339 294 320 272 286 281 314 283 336 362 270 360 302 312 303 287 308 244 308 304 378 348 403 393 351 393 342 339 384 352 460 423 384 477 433 438 396 430 385 339 342 297 394 410 413 430 425 384 313 347 315 323 319 291 330 344 375 370 347 381 309 326 316 280 269 254 271 207 248 294 303 362 275 269 225 232 213 242 272 252 261 313 324 298 291 261 249 267 * This time period and focus slightly differs from other sections included in the profile and so results will not align. Source: MPS analysis. tt Age of Teenage Violence Victims 13 - 14 1S 16 17 18 19 Ethnicity of Teenage Violence Victims 0 100 200 300 400 SOO 600 Unknown, 165 Otller,8 • • ~''""I Black, 1,083 MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights – victim profile and risk factors Victim Profile – Teenage Violence Victim Risk Factors – Education and Exploitation Gender In the last three years the vast majority of victims of teenage 90% 10% violence have been male, with 90% of the victims being males. Age Those in their late teens are more likely to be the victims of violence with those 18yrs of age most likely to be victimised for both homicides and overall violence. Ethnicity Black victims are more prominent within the data than other ethnicities, with 48% of the recorded victims being described as black. Overall 44% of the victims of teenage violence are black males, with this increasing to 73% when homicides are looked at in isolation. This is higher than for overall homicide offending where 39% of victims are described as being black males. Source: MPS analysis. Covers time period 2019 to 2021. * Young people under 18 not in full time education would include those subject of exclusion from school, but would also include those on an apprenticeship/traineeship or those to spend time working whilst in part time education. Not in full time education* When looking at all victims of teenage violence, 90 are shown on the crime records as being not in full time education. This is 23% of the total number of teenage victims in 2021. However, if just those aged under 18 are looked at then this proportion increases to 40% of the teenage victims of violence identified in 2021 (further work is needed to confirm the educational status of these victims). When teenage homicide is looked at in isolation then there are 11 victims shown as not being in full time education on CRIS where further searches have not identified other forms of education they may be utilising. This is 37% of all teenage homicide victims and 61% of those under 18. Exploitation In the last three years there have been 295 teenage victims of violence that have also featured on an exploitation Crime report. This is 13% of the victims in this time period. Six of the 71 teenage homicide victims have featured on these reports, 8% of the total. In 2021, 3 of the 30 victims of teenage homicide had previously featured on an exploitation Crime report ­ 10% of the victims. { { { tttttt tttttt MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights - victim offending history There have been 390 victims of homicide, stabbings or lethal barrelled weapon (LBW) discharges in 2021 where the injury degree is shown as moderate or above, of these: When teenage homicide is looked at in isolation, 60% (18) of the 30 victims were arrested in the twelve months prior to the offence taking place and 63% (19) have been previously arrested at any point before the offence took place. 34% (152) have been arrested of an offence in the 12 months before the offence took place. 50% (194) have been arrested if the entire offending history of victims is considered. …(63) of teenage victims have been arrested for violence in the 12 months prior to the offence taking place. This increases to 16% 20% (79) if all arrests are considered. Violence 30% …(9) of teenage homicide victims have been arrested for violence in the 12 months prior to the homicide taking place. This increases to 37% (11) when their entire arrest history is looked at. …(56) of teenage victims have been arrested for possession of offensive weapons in the 12 months prior to the offence taking place. This increases to 14% 29% (114) if all arrests are considered. Weapons …(6) of teenage homicide victims have been arrested for possession of offensive weapon in the 12 months prior to the homicide taking place. This 20% increases to 40% (12) when their entire arrest history is looked at. …(47) of teenage victims have been arrested for robbery in the 12 months prior to the offence taking place. This increases to 27% (106) if all arrests are 12% considered. Robbery …(3) of teenage homicide victims have been arrested for robbery in the 12 months prior to the homicide taking place. This increases to 40% (12) when 10% their entire arrest history is looked at. ...(78) of teenage victims have been arrested for drugs possession/supply in the 12 months prior to the offence taking place. 44 of these have been 20% arrested for supply and 58 for possession. This increases to 30% (93) if all arrests are considered, 93 for possession and 60 for supply. Drugs …(8) of teenage homicide victims have been arrested for drug offences in the 12 months prior to the homicide taking place, 4 for supply and 5 for 27% possession. This increases to 30% (10) when their entire arrest history is looked at. Source: MPS analysis. Covers time period 2019 to 2021. r ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights – victimisation history Repeat Victims Victim and Accused Previous victimisation In the last three years there have been 44 people who have been the victim of an offence on more than one occasion. Five of these individuals have been the victim of three offences over this time period. Two individuals have featured as a victim in each of the three years analysed: Three homicide victims in the last three years were previously the victim of a teenage violence offence. Two of these were the victims of homicides committed in 2021. …One of these was the victim of a fatal stabbing in 2021 having suffered minor stab wounds in 2019 and moderate stab wounds in 2020. …The second suffered minor injuries in 2019 and 2020 but suffered serious injuries when he was stabbed in 2021. Source: MPS analysis. Covers time period 2019 to 2021. Prevalence 39 …individuals feature as both a victim and an accused for teenage violence within the last three years… Educational risk-factors 15 of these nominals are shown as being ‘Not in Full Time Education’. With 14 members of this cohort having no status recorded this is 60% of those where an educational status is shown on CRIS. …Together these individuals have been accused of 42 offences. Age of Accused (Teenage Accused Only) 13 - 14 15 16 17 18 19 0 20 Asian, 81 40 60 80 100 Ethnicity of Accused Other/Unknown, 26 tt 120 MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights – accused profile and risk factors Accused Profile – Teenage Violence Accused Risk Factors – Education and Exploitation Overall there have been 669 people accused of committing acts of violence against teenage victims since 2019, 194 in 2021. Gender Males are significantly more prominent in accused data, with 94% of accused in the last three years being male. 94% 6% Age and Ethnicity Teenage violence is largely a peer on of victims which was 16.9. Overall 61% (384) of the accused are black males. and then 18. Although the average age of all accused is 18.5 if just teenage accused are looked at then the average age of accused is 16.5, slightly lower than the average peer 17 age the accused Unlike accused Source: MPS analysis. Covers time period 2019 to 2021. offence, with 77% of the also being teenagers. for the victim data the are younger with 16 being most common age, followed by The current education status of those accused of homicide has been analysed for those who are known to be of school age (17 and under). Of the 48 accused within this age group 20 are recorded as not being in full time education (42%) 2021 (further work needed to confirm the educational status of these victims). years there have been just 14 Not in full time education * For all accused in 2021 there are 68 shown on a Crime Report as not being in full time education. This is 35% of all accused but when just those that are aged under 18 are looked at this is 56% of the 116 accused. Exploitation In the last three accused of teenage violence offences that have also featured on an exploitation Crime Report. This is just 2% of the accused in this time period. Overall there have been 44 people in the last three years who have featured as both a victim and an accused for a teenage violence offence. This includes 16 of the accused in 2021. * Young people under 18 not in full time education would include those subject of exclusion from school, but would also include those on an apprenticeship/traineeship or those to spend time working whilst in part time education. { { -- -- - tttttt tttttt MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights – accused offending history There have been 152 nominals accused of committing a violent offence where the victim is recorded as a teenager in 2021 and a moderate or above injury has been suffered. 82% of these offenders have previously been arrested by police, and… 68% were arrested in the 12 months prior to the incident being committed. When teenage homicide is looked at in isolation 78% (49) of the 63 accused were arrested in the twelve months prior to the offence taking place and 87% (55) have been previously arrested by police at any point before the offence. …(45) of those accused of all teenage violence have been arrested for violence in the 12 months prior to the offence. This increases to 56% (85) if all arrests 30% are considered. Violence …(14) of those accused of teenage homicide have been arrested for violence in the 12 months prior to the offence taking place. This increases to 52% (33) 22% when their entire arrest history is looked at. …(49) of those accused of all teenage violence have been arrested for possession of a weapon in the 12 months prior to the offence taking place. This 32% increases to 40% (60) if all arrests are considered. Weapons …(22) of those accused of a teenage homicide have been arrested for possession of a weapon in the 12 months prior to the offence taking place. This 35% increases to 59% (37) if all arrests are considered. …(33) of those accused of all teenage violence have been arrested for robbery in the 12 months prior to the offence. This increases to 49% (74) if all arrests 22% are considered. Robbery …(17) of those accused of a teenage homicide have been arrested for robbery in the 12 months prior to the offence. This increases to 51% (32) if all arrests 27% are considered. …(44) of accused of all teenage violence have been arrested for drugs possession/supply in the 12 months prior to the offence taking place. This increases to 29% 45% (68) if all arrests are considered. Of these 22% (33) have been arrested for supply in the 12 months prior and 32% (48) for supply if all arrests are considered. Drugs …(31) of those accused of a teenage homicide have been arrested for possession/supply of drugs in the 12 months prior to the offence taking place. This 49% increases to 70% (44) if all arrests are considered. Of these 22% (14) have been arrested for supply in the 12 months prior and 43% (27) if all arrests are considered. Source: MPS analysis. Covers time period 2019 to 2021. ' I , il -------.. aF ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights – accused position in CJS Accused Offending – Position in the Criminal Justice System Police Bail Accused Involvement in the Criminal Justice System The 152 nominals accused of violence with a teenage victim suffering moderate or above 21 of those accused of violence against a teenager were injuries have been reviewed in terms of their position within the criminal justice system at the given police bail at the time the incident took place, time the incident took place. For the purposes of RUIs the six months prior to the offence 14% of the total accused. 10 of these were subsequently have been reviewed. accused of a teenage homicide, 16% of homicide accused. Ten of those that were given police bail had been bailed by a different force. Overall, These include Hertfordshire, Kent, Essex, Norfolk, Suffolk and Dorset. 40% (61) of total accused…. In the case of three of these nominals they were given curfews as part of their bail conditions. …and …were found to have been in the Criminal 49% (31) of homicide accused… Justice System Released Under Investigation 30 of the accused had been released under investigation (RUI’d) in the six months prior to the Court Bail incident taking place, 20% of the total accused. 25% of homicide accused. 22 of the accused were on court bail at the time the incident took place, 15% of the total accused. Of these, eight had a curfew as part Accused had been RUI’d for a range of offences of the bail conditions imposed by court. Eleven of these were from drug supply, knife possession and theft. subsequently accused of a teenage homicide, 17% of homicide accused. Of those on a curfew four are shown on PNC as having breached the curfew in the time period before the incident. Source: MPS analysis. Covers time period 2019 to 2021. ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights – the role of gangs Gangs and Organised Crime Groups The volume of gang related offences has decreased in each of the three years analysed (176-120-100). Over this time period the proportion of gang related offences has also reduced (19%-18%-15%). Overall, there have been 396 teenage violence offences in the last three years that are assessed as being gang-related. 17% of the total number of offences that have been committed in this time period. When teenage homicide is looked at in isolation the proportion of gang related offences increases. In contrast to the overall violence picture, 2021 saw an increase in the volume of gang related homicide offences (14-5- 18). In both years with higher teenage homicide offending gangs are more prominent in the data, implying that within teenage violence the more serious harm is caused by gangs. Overall, there have been 37 teenage homicides that have links to gangs. This is 52% of those committed in the last three years. Source: MPS analysis. Covers time period 2019 to 2021. Further detail on gangs as a driver of homicides Overall there have been 43 separate gangs linked to homicides in the last three years. 28 of these networks have been linked to the committal of teenage homicide offences whilst 21 have had members or males affiliated to networks feature as victims. Gangs have been involved in homicides to varying degrees in 2021, from the offence being specifically linked to tensions to those with links to gangs members being involved in the offending. In some cases victims have been on the periphery of gangs and despite not being full members targeted as a result of tensions between networks. There are eight gangs that have been linked to more than one teenage homicide in the last three years. Two of these are based in Croydon with all of these offences having been committed since July 2021. MAYOR OF LONDON OFFICE FOR POLICING AND CRIME MPS insights from teenage violence problem profile – weapons and drugs as drivers Knives, and in particular the increased use of combat/machete/Rambo style knives, are a key driver of teenage violence, in particular homicide. The use of these weapons appears to be a key factor in teenage homicides with the nature of these weapons meaning that fatal and serious injuries are more likely. As previously noted, the carriage of weapons is a key signal offence for offenders in teenage violence with 59% of accused having previously been arrested for the possession of weapons. Knives In the last three years there have been 2106 teenage victims injured as a result of a knife attack. However, the volume of knife enabled offending has decreased each year (845-636-625). 64 teenage homicides have been knife enabled (90%). This has been highest in the most recent twelve month period (23-16-25) although the proportion that have been knife enabled is actually lowest in the most recent 12 month period (96-94-90). Combat/machete/Rambo style knives The use of such knives appears to be increasing in London with the proportion of knife injury offences involving these weapons increasing in each of the three years analysed (34%-36%-44%). It should be noted that this may be partly down to improved awareness and recording of these knives on reports. This change in weapon use could be a driving factor behind teenage homicide with 82% of teenage homicides committed in the last two years involving the use of a combat/machete/Rambo style knife. This is significantly higher than for homicide as a whole, with the volume of teenage homicides involving these weapons having more than tripled in 2021 being a 50% increase in the volume of offences involving these weapons within teenage homicides in the last twelve months (5-16). Source: MPS analysis. Covers time period 2019 to 2021. Firearms and Other Weapons There have been 99 teenage victims of a firearm discharge in the last three years. As with knife offending the volume of offences have decreased each year (37-33-23) In the last three years 4 teenage homicides have been firearm enabled (6%). However, in contrast to overall teenage violence, firearm enabled homicides have increased with three of these four offences having been committed in the most recent twelve month period (1-0-3). Two of the four are linked to gangs (50%). The remaining two teenage homicides have been committed using ‘other’ weapons. These offences consist of an arson and a believed blunt object offence. Drugs There are difficulties in establishing those offences that are motivated by drugs within wider teenage offending. In 2021 there have been 2 teenage homicides where drugs has been cited as a motivational factor, 7% of the total teenage homicides committed. However, if the wider home office definition relating to drug related is used then there have been 19 teenage homicides with links to drugs, 63% of the total homicides committed. If the wider home office definition for homicide is also used for teenage violence as a whole then there have been 148 drug related offences in 2021, 38% of those where the victim received a moderate or above injury. Probation data adds an extra layer of insight Work was undertaken to assess the linkages of victims and suspects of all 2021 teenage homicide victims to London Probation data. At the time of the review*, 61 individuals had been charged in connection with these homicides. Gender and Age 9 1 There were 9 males and 1 female. All were aged 18 to 25, the youngest 2 having recently transferred to Probation from the Youth Offending Service (YOS). Supervision 6 were subject to Licence and 4 to a Community Order. . There was no breach proceedings initiated at the time of the homicides so the assumption being they were all compliant. 10 of the charged individuals were current to London Probation. Of these... Convictions Accommodation Risk All 10 had convictions for violent offences inc. assault, robbery, possession of a blade. 4 had Possession with Intent to Supply, although a number of others also had previous convictions for drug supply too. 8 had previous convictions for carrying knives. 3 were assessed as high risk 5 as medium risk 2 as low risk. Accommodation was regarded stable; no-one they lived with known to London probation. 3 were looked after children and lived independently, 6 lived with family and 1 was resident in an Approved Premises. Safeguarding Safeguarding concerns were highlighted in 2 cases due to identification of DA concerns. 5 had known associations with gangs. Of the ten, half had diagnosed or suspected neurodiversity concerns, specifically Asperger's, ADHD, Autism Spectrum Disorder – evidence suggests that neurodivergent individuals are increasingly part of the make up of people who are gang affiliated. Of the 30 teenage victims murdered in London in 2021, one was known to London Probation. He was aged 19 when he died in March last year and was subject to an Unpaid Work order (UPW) imposed as a condition of a Youth Rehabilitation Order (YRO). The YRO had expired but the UPW remained due to this element having been suspended as a result of the pandemic. He was a looked after child, diagnosed with Autism Spectrum Disorder and assessed as high risk of harm. He lived alone in semi-independent accommodation. He was murdered by a 22 year old female – this was his partner, although this was not known initially as the victim told his probation officer he was single and not in a relationship. She was known to London Probation and was on license for a previous offence. * March 2022. - - – – MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Juvenile robbery First Time Entrants are increasing against an overall reduction, with community sentences the most likely outcome for juvenile offenders 56% of those sentenced for robbery were Juveniles (10-17) – this compares to 9% of total cohort. 50% increase in juvenile FTE for robbery - over the last five years. Against a 28% reduction in all juvenile FTEs. 13% of all sentenced VAP offenders in London (CY2020) were aged 20 or under. 75% of those under 20 VAP offenders were sentenced for one of the following - Assault of an emergency worker, Wounding with intent to cause grievous bodily harm or VAP offenders aged under 20 were 3.5 time more likely to be defendants for murder, within their age group ­ compared to the adult cohort. Weapon enabled robbery Most Serious Violence Assault occasioning actual bodily harm. Offenders aged under 20 were twice as likely to be an offender of Wounding with intent to cause grievous bodily harm ­ compared to the adult cohort. Juvenile ABH offenders were over 7 times more likely to be given a community sentence than either an adult (20+) or a young adult (18-20). 29% of VAP juvenile offenders were reoffenders nationally within the period Oct - Dec 2019, remained relatively stable since Oct - Dec 2018. Homicides Knife crime sentencing In 2020/21 there were 3,340 knife crime sentencing disposals, a 12% reduction from the previous financial year. 28% were of individuals aged under 18. Ages 16 17 accounted for the highest rate (271 disposals per 100,000 population). Black males aged 16 17 were 5.3 times more likely to receive a knife crime disposal compared to White males. Only 14% of those with no previous convictions/cautions received an immediate custody with a third receiving a community sentence these increase to 69% for those with previous knife related convictions. Community sentence use for juvenile robbery offenders is 3 times higher than for VAP, higher than other crime type including drugs. Use of caution 6 times less likely for Robbery than for VAP*. 40% of robbery juvenile offenders were reoffenders nationally within the period Oct - Dec 2019, increasing since Oct - Dec 2018. *Regression model Source: MoJ, unless stated data used covers 2020. https://www.gov.uk/government/statistics/criminal-justice-system-statistics-quarterly-march-2021 MAYOR OF LONDON OFFICE FOR POLICING AND CRIME 2. Understanding the joins and gaps within the violent cohorts - Analysis across intelligence and disproportionality G,VM 34% '-.,/ ,,-....., 69,775 URNs X 15,840 PNCIIDS 23% X 1,429 On a matrix 9% PRO 63.62% .....__,,, I-IKC PFO 45.19% 36.73% '-.._,/ .......__,,, "- l / ---------1+ = MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Official Sensitive internal use only – The majority of young people impacted by violence were not on Police systems Violence cohort make-up Source: Violent Cohort CRIS data extract Our Violence Cohort has been created using CRIS records from January 2019 to February 2021, with the following parameters: • Current classification Violence with Injury • Domestic Abuse Incidents Excluded • At lease one victim, suspect or accused is under 25 VHI These 3 Matrices will be replaced with the new VHI matrix Of those on any matrix, proportions on each: Where details were known, the Violence Cohort was 69% male and 31% female. In terms of ethnicity the make up was 48% White, 32% Black, 17% Asian and 3% other. 70,000 individuals that have been suspected of over 80,000 offences during this period.. 15,840 (23%) have been identified as having a PNC ID, indicating how the majority of individuals are not known to police (at least formally). 1,429 (9%) individuals matched via their PNCID to at least one of four different Metropolitan Police matrices we were able to access. 43% of all individuals on one of the four matrices were present within our violence cohort. 93% of all individuals matched to the matrix were the suspect of a violent offence. With 7% being both suspect and victim. Matrix criteria • Gangs Violence Matrix* (675 of cohort matched /1,985 total individuals on matrix) – (i.e.,) someone identified as being in a gang identified from two or more intelligence sources. • Priority Robbery Offenders (495 /778) - (i.e.,) suspect coming up 3 times in 2 years for personal robbery offence. • Habitual Knife Carriers (459 /1,016) - (i.e.,) suspect coming up twice in the last 2 years for possession or Injury involving a blade/Knife. • Priority Firearm Offenders (137 /373) - (i.e.,) Individuals coming up twice in last 5 years for possession of a Gun offence. Going forward the PRO,HKC & PFO Matrices are going to be replaced by the Violence harm index (VHI) matrix. Whilst this cohort has similarities to the SYV cohort in the next section, this dataset is more encompassing including a wider age range, covering 20 offence classification types. *The Mayor’s review of the Gangs Violence Matrix was published in late 2018, since this time the Deputy Mayor for Policing and Crime has continued to carry out oversight of the steps taken by the Metropolitan Police Service to implement the recommendations made by the Mayor. ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Splitting robbery, most serious violence and homicide cohorts is needed to better understand the landscape of violence impacting youths Differences emerge between weapon enabled, most serious violence and homicide. Weapon Enabled Robbery Most Serious Violence Homicides Total Cohort Victims Suspects Victims Suspects Victims Suspects All Number 9,141 7,396 29,577 27,124 160 634 69,775 Male 84% 91% 60% 68% 93% 93% 69% Gender Female 16% 9% 40% 32% 8% 7% 31% Black 18% 59% 27% 39% 61% 65% 32% Asian 4% 11% 3% 15% 2% 9% 17% Ethnicity Other 23% 3% 17% 3% 8% 5% 3% White 56% 28% 53% 44% 28% 22% 48% Age Mean 26 21 26 28 27 25 26 Median 21 19 23 25 25 23 23 Mode 18 18 17 18 20 21 17 • Across all cohorts, a majority of suspects and victims were male. • For weapon enabled robbery - over half (56%) of all victims were White, with 59% of suspects recorded as Black. • The majority of homicide victims were male (93%) and Black (61%). 61% of victims of homicide were recorded as Black. • Suspect profile is slightly younger than the victim cohort. Ethnicity: the ethnicity data used in this report is observational data recorded by the police in the relevant crime reports. Although the police do also collect self-reported ethnicity data, this is not sufficiently complete for a population analysis. Data has been cleaned in order to remove duplicated demographic categories. Of the total cohort there 11,195 individuals with no identifiable ethnicity. Also, 3,116 had no identifiable gender information. Source: SAP BI (MPS Database) – Victim data, unless stated data used for profiles covers 2020 calendar year; WhatDoTheyKnow School Data; IMD (ward level) – London Datastore (GLA); Cultural Venue London Datastore (GLA). ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Homicide suspects more likely to be in groups, rated higher harm and more likely to be identified on a matrix For those on any matrix 85% of violent offence suspects identified on a matrix were repeats. Nearly 20% of individuals matched to a matrix were found to be on more than one matrix. 59% of offences on the matrix had more than one suspect or victim. Weapon Enabled Robbery Most Serious Violence Homicides Total Victims Suspects Victims Suspects Victims Suspects Victims Suspects Number 9,141 7,396 29,577 27,124 160 634 38,611 33,491 More than 1 Suspect / Victim 14% 40% 13% 16% - 89% 13% 22% Offence Crossover * 4% 30% 5% 12% 3% 33% 4% 13% Repeats individuals 1% 22% 4% 8% 0% 2% 9% 18% Repeat Average offences 2.8 3.5 2.5 2.9 1.2 3.2 2.3 2.7 % against violence/harm tools 0.3% 11.6% 0.3% 3.2% 1.3% 24.1% 0.3% 4.3% 1 MATRIX 21 689 65 693 2 106 85 1150 2 MATRIX 4 131 19 151 0 31 20 223 3 MATRIX 1 34 2 32 0 15 3 50 4 MATRIX 0 5 0 4 0 1 0 5 GVM (Gangs Violence Matrix) 0.1% 5.0% 0.1% 1.5% 0.0% 12.6% 0.1% 2.0% HKC (Habitual Knife Carriers) 0.1% 2.9% 0.1% 1.3% 0.6% 11.7% 0.1% 1.4% PRO (Priority Robbery Offender) 0.2% 5.9% 0.1% 0.9% 0.6% 4.3% 0.1% 1.5% PFO (Priority Firearm Offenders) 0.0% 0.5% 0.0% 0.4% 0.0% 5.7% 0.0% 0.4% Robbery MSV Homicide • 25% of all offences recorded against the violence • 74% of all offences were related to MSV in the • 1% of all offences were recorded as Homicide. cohort were related to robbery. This compares to 63% violence cohort, but represented only 33% of the • More likely to be captured on a matrix and more likely of those matched to a matrix – the biggest increase offences on the matrix. to be groups of suspects. when focusing on the matrix cohort compared to all the • Less likely to be captured within a matrix. • 33% of Homicide suspects had committed other other cohorts. • Had the highest amount of repeat victims at 4%. violent offences. • 12% of all robbery offenders from the violence cohort • Had the highest number of offences with 30,006 in the • 89% homicide offences had more than 1 suspect or were matched to a matrix. cohort but also the lowest proportion of offence on a victim involved. • Repeat offending more prevalent (18% of all matrix with 4.5%. offenders), with more offences per reoffender than other groups. Source: Violent Cohort CRIS data extract. *Offence Crossover – this is the % of victims or suspects that appear for each crime type that also appear in the extract for other crime types. ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Disproportionality is observed across violence – especially for homicide • Relative to population, there were 19.3 Black homicide suspects relative to 100,000 population in London (based on mid 2021 estimates). • There were also 6.4 Black victims of homicide per 100,000 population. • Other ethnic groups also accounted for higher rates of victimisation or representation as suspects. For example, there were 3.6 suspects of homicides of Other ethnicity; 8.2 victims of serious violence per 1,000 population and 3.4 victims of weapon enabled robbery per 1,000 population. Per 1000 population Per 100,000 population Weapon Enabled Robbery Most Serious Violence Homicides Rate per population Victims Suspects Victims Suspects Victims Suspects Black 1.0 2.4 4.8 5.6 6.4 19.3 Asian 0.1 0.3 0.4 1.5 0.1 2.0 Other 3.4 0.3 8.2 1.1 2.4 3.6 White 0.9 0.3 2.7 1.8 0.8 1.8 • When exploring the levels of disproportionality within the cohort we see Black individuals were almost 8 times more likely than White individuals to be victims of homicide. • Black individuals were also almost 11 times more likely to feature as a suspect in a homicide investigation than White individuals. • In terms of robberies, Black individuals were almost 8 times more likely to feature as a suspect. Victim Suspect Black Asian Other Black Asian Other Weapon Enabled Robbery x1.1 x0.2 x3.9 x7.6 x1 x1 MSV x1.8 x0.1 x3 x3.2 x0.9 x0.6 Homicide x7.7 x0.2 x2.8 x10.7 x1.1 x2 Source: Violent Cohort CRIS data extract. _I _ I _I I _ ■ ■ ■ ■ ■ 1111.. IIL ■ ■ ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ ONDON Focussing on those individuals on any matrix • 97% of people on a matrix were male. 30% of the total violence cohort and 40% of Black suspects more likely to be on each of matrix • This compares to 69% of people in our suspects were identified as Black, 38% of the total violence cohort and 72% of suspects. cohort and 33% suspects identified as White. HKC PFO PRO GVM Cohort Cohort, 69% Suspect, 72% PNC, 83% Matrix, 97% 66% of the violence cohort matched to a matrix identified as Black, 14% were identified as White. W/E Robbery Splitting by crime type we see variance disappear for suspects of MSV, whilst Black individuals are 3.5x more likely to be suspected of homicide and 2.6x MSV more likely to be suspected of w/e robbery. Disproportionality increases and varies across cohorts identified within different matrices i.e. Black homicide suspects 21x more likely to be identified in Homicide the PRO cohort, than White suspects. 4.1X 3.6X 5.0X 5.3X 4.3X 3.4X 4.0X 4.0X 21.0X 9.0X 3.9X 3.9X 2.6X 1.1X 3.5X Offences by Cohort 74% 71% 62% 63% 36% 33% 25% 27% 4% 1% 2% 2% Cohort Suspect PNCID Matrix MSV Homicide Robbery • Matrix nominals were more likely to • 85% of offences from suspects on the suspected of robbery offences. matrix were repeat suspects, compared • 59% of offences that had a victim or to 21% of those not on any matrices. suspect on a matrix had more than 1 suspect. 50% Age group 40% 30% 20% 10% 0% 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45+ Cohort Suspect PNC Matrix • 38% of the matrix cohort was aged between 15­ • Individuals were more likely to be on PFO (47%), 19 this compares to 26% of the entire violence HKC (39%) or GVM (42%) if aged between 20-24. cohort. • 15-19 year olds matched to a matrix represented • 42% of the individuals that are on 3 or more 57% of the PRO cohort. matrices were aged 15-19. Source: Violent Cohort CRIS data extract. Using a regression model we are better able to understand the variations observed within the data >> - I I I I ..................................................................................................................... I- - I Suspect being Black Xl.6 .__ ____________________________________ QNDON OFFICE FOR POLICING AND CRIME Regression analysis shows key predictors for Matrix inclusion Regression analysis was used to identify which variables had the strongest association with being on any matrix. The diagram below shows factors above the line that make an individual over twice as likely to feature on a matrix, whilst those below are significant but contribute less. Being the suspect of multiple offences Matrix Being younger Having an average harm index score* Being the suspect of robbery above the standard deviation Being male Involvement in group offending More than 2x more likely Less than 2x more likely Being suspected of murder Suspect being Black Factors linked with being on a matrix Likelihood of inclusion Suspects aged 18 20 compared to over 35 year olds X13 Repeat suspects; +2 offences +5 offences +10 offences X2 X3 X4 Harm index score* above the mean X3 Being a robbery suspect X3 Being Male X2 Offending in groups X2 Being a suspect of murder X1.9 Suspect being Black X1.6 • The largest impact - besides being aged 18-20 - comes from whether an individual was suspected of more than one offence, with the likelihood of inclusion increasing in correlation with the number of suspected offences. • If an individual recorded an average harm score higher than a standard deviation, they were 3x more likely to be included in a matrix. Whilst those who were one of multiple suspects linked to an offence were 2x more likely. • When the individual was suspected of robbery inclusion was 3x likely, whilst suspects of murder were 1.9x more likely. • Individuals who were male were 2x more likely whilst those suspects identified as Black were 1.6x more likely. *Harm index score refers to the Cambridge harm index(CHI) where each offence was matched to a score on the CHI that relates to the severity of the crime – an average harm score was given to each individual within the cohort. luctwnond Upon J • • HOUnllo,,o,. sm, h&Ful • • Upon Th.-mH. E Sou hw ric • • • • • • C1oydon • 11. II. 11. I._ I._ I._ ■ ■ ■ • lsl,ngton • W himFor~ • E I • ,ng,on • • Ri<hmond VPon Tha e c • Hounsk!w MAYOR OF LONDON OFFICE FOR POLICING AND CRIME boroughs Repeat robbers are the suspects in 60% of robbery offences 1 in 4 offences within the cohort related to weapon enabled robbery - the highest rate of repeat suspects (22% of the total, average of 3.5 offences per suspect). 79% (5,879) of weapon enabled robbery suspects did not deviate from this offence type, this compares to 63% (1,033) for repeat robbery suspects. In our violence cohort there were; Weapon enabled robbery – age groupings • 7,396 suspects of weapon enabled robbery, suspected of committing 11,438 offences. • 12% of these suspects appeared on a matrix – this compares to 4% across the entire cohort. 63% 59% 62% • 37% (3,694) of all repeat w/e robbery offences involved more than one suspect. 48% 37%37% 35% 29% • Repeat robbers were suspected in 60% of all robbery offences within the cohort. 24% 17% 16% 16% 15% 12% 11% • Repeat w/e robbery offenders were more likely to be aged 15-19 compared to non-repeats. 5% 4% 3% • Black suspects made up the highest proportion of each cohort (matrix 44%, total cohort 41% Total Cohort Matrix PNC Total Cohort Matrix PNC and PNC 41%) this was the same for both singular and repeat offenders. Single offence Repeat offender • Repeat suspects were predominately male (94%), similar to single suspects. 15-19 20-24 25-29 The top 10 prolific robbery suspects were active across many Top 10 suspects The top 10 most prolific offenders were all known to the police (PNCID) • 8/10 of suspects were mobile offending across multiple boroughs. • 4% of the top 10’s offences deviated from w/e robbery, compared to 18% of offences for all repeat robbery suspects. • 2 suspects committed offences across 7 different boroughs. • 9/10 of the top 10 were Black with 1 being white; 9 were Male. • Ages ranged between 16 and 21. • 6/10 were on at least 1 matrix. • The 10 most prolific robbers were suspected of committing 209 offences (between 20- 26 offences each). • 26% of all their offences happened in Lewisham. Source: Violent Cohort CRIS data extract. [J I _._ - - - ... c::::::J c::::J - - [J I 0 OD I I - - -- 0 I I - - - - - -0 I I . . . . . . . . . ■ ■ ■ ■ ■ DO O 0 ■ ■- ■ I I I I I-• A I ---- . ~--• ■ 0 ■ ------- -: "'" -l:,. -- ~-­ . ___ ____. ■ ■ . . . . . MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Case study – a countdown to homicide 634 homicide suspects in our violence cohort. 153 individuals were identified on a matrix. 19 high risk flags were attributed to homicide suspects within the cohort – a total of 265 individuals were identified as high risk across all 4 matrices. A case study timeline leading up to the murder of a youth - captured within our cohort - illustrating the complexities of violent offending. • Both individuals are suspects in a robbery nine months before the murder. • The suspect has been suspected of a serious wounding offence two months prior to the murder. • The victim has not been involved in violent offending excluding robbery. • Both of these individuals had a PNC. • Neither of these individuals flagged up on any of the matrix tools. • Nearly all of their crime involved a collective of people. Southwark Suspect in Serious Wounding Victim in Serious Wounding Crossover Suspect in Robbery of Personal Property Suspect in Murder Victim in Murder Southwark Westminster Aug 2019 Aug 2019 Oct 2019 Homicide Suspect Apr 2019 Nov 2019 Oct 2019 Jan 2020 May 2020 May 2020 Jun 2020 Feb 2020 Mar 2020 Lambeth 15 mins between being a victim and offender of SV Homicide Victim Southwark Source: Violent Cohort CRIS data extract. MAYOR OF LONDON OFFICE FOR POLICING AND CRIME behind the violence 3.Violence in context In depth case coding to understand the motivations and attrition – )f { ' ' \ ':: -( ~" ~ -v ! I r ) I- MAYOROFL OFFICE FOR ONDON POLICING AND CRIME Deep dive coding of the more violent cases to learn more … The research adopted the same 'deep dive coding' of CRIS police data as used within the MOPAC Rape Review; The Lighthouse Evaluation and the Deep Dive into Domestic Abuse. In Total: 282 cases of SYV offences taken from police systems (CRIS) were coded across 95 variables in 5 areas. Cases were randomly generated from Jan 2019 Feb 2021. Jan 19 – Feb 21 95 variables in 5 main areas: Victim/Survivor Perpetrator Offence Procedural Case Outcomes 282 cases randomly selected from CRIS Methodological shortcomings • Not able to explore trends over • Police system (CRIS) is not time given the coding. designed as a research tool. • Recent trends/changes are not In depth case coding is time intensive, so a spread of cases were selected from three • Data recorded & held by the included in the sample. boroughs - Tower Hamlets (95), Southwark (103), Croydon (84). This is not a CPS, the courts, & victim representative sample, the selection of these boroughs were determined by considering the support services is not included. crime counts for GBH & GBH with intent, along with a measure of public health deprivation • Unrecorded cases are missing for all London boroughs developed by the Violence Reduction Unit. from this methodology. A pilot phase examined 40 cases to test the effectiveness of the coding framework at capturing the variables and to narrow the sample to contain the salient characteristics. • • ' ' • • ' ' • • II II II __ • • • - • • • • ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME The demographics of the selected coded cases • There were a total of 337 victims, with 44 of cases having multiple victims. 40% suspects had a PNCID. • Ages of victims ranged from 16 to 70 years, half being aged 15-20 (n166, 51%) and the most common ages being 16 & 17. • There were a total of 489 suspects, with 117 cases having multiple suspects. • Suspect age ranged from 10-50 years, but over half (n230, 56%) were aged between 15-20. The most frequent age was 18 years (n58, 14%). • Where known for victims (n156): 58% (n90) were in full time education, 22% (n35) employment, and 20% (n31) not in full time education or employment. • For suspects, where known (n162): 60% (n98) were in full time education, 28% (n46) not in either full time education/employment, and 11% (n18) in employment. • 20% suspects had prior offence history. 13% suspects suspected of being in a gang. Victim/Suspect Relationship Weapon Use Friend Acquaintance Stranger Knife/Bladed object Improvised weapon 216 (64%) 121 (36%) 188 181 Blunt Instrument No weapon Male Victims Female Victims 140 123 353 (72%) 136 (28%) 34 Male Suspects Female Suspects 39 12 Victim/Suspect Relationship Ethnicity (victim & suspect) Incident Location Weapon Used Victims Suspects 156 105 110 92 80 76 102 In the 84 At Private 31 At Education 16 Street/Pavement residences establishment 6 White Black Asian Other Not recorded Victimlsusp protective/possessive over friend/family Domestic disagreement/DA Playground scuffle Alcohol/NTE Gangs/explottation Drugs Robbery/properiy damage Bullying/peer pressure Premedttated Snttching Fraudn11oney other Suspect retaliating Self defence claim Online provocation MH crisis/suspect distressed Misunderstanding/accidental 0 20 40 Cluster groups ■ Group violence Single suspect 60 80 100 120 Count MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Typologies and Motivations of serious youth violence A statistical technique known as Cluster Analysis was used to identify a typology of violence - two broad groupings were the best fit for the overall data, that of ‘single suspect’ (60%) and ‘group of suspects’ (40%). capture was Suspects being a previous perpetrator and victim being a previous perpetrator were found to be significant drivers of the victim suffering moderate-severe injuries, as was number of suspects (e.g., 3 perpetrators compared to just 1, made moderate -severe injury 6 times more likely). Significant predictors of there being multiple suspects in the incident were that the victim was a repeat victim (8 times more likely), and that police had suspicions of the suspect being associated with gangs (12 times more likely). Cluster 1 'single' (60% of the 282) Cluster 2 'group' (40% of the 282) Multiple perps involved No 85% Yes 82% Weapon involved No weapon 69% Knife 53% Victim/suspect relationship Acquaintance 57% Strangers 62% Offence location Private dwelling 40% Outside 63% Witness present Associates 51% Bystanders 56% Violence duration Momentary 64% Prolonged 56% Injury levels 69% none/minor 67% moderate/severe Motive Immediate retaliation (34%) Unclear (49%) Gang involvement 4 victims and 1 suspect were known or suspected gang members. 21 suspects and 16 victims were either known or suspected gang members Victim age Median age 17 Median age 19 Victim sex 58% male, 42% female 76% male, 24% female Victim ethnicity White is most common ethnicity for victims (38%) Black is most common ethnicity for victims (40%) Suspect age Median age 18 Median age 19 Suspect sex 76% male, 24% female 93% male and 7% female Suspect Ethnicity Black is most common ethnic group for suspect (39%) Black is most common ethnic group for suspects (36%) Offence characteristics Demographics MPS Data limited on motives Prior CJS involvement, investigative stages and victim attrition Victims withdrew in half (50%, n141) of all cases - the most frequent withdrawal reason being that the victim became disengaged or was not willing to engage from the start. In 38 cases, the victim withdrew within a day of the case being reported. The time to withdrawal ranged from one day to 654 days. The median number of days to victim withdrawal was 12. Regression analysis showed that victims were significantly less likely to withdraw when: - if public bystanders were present (victims were 3.5x more likely to not withdraw); - if there were associates of victim and/or suspect present (victims were 2.7x more likely not to withdraw). Regression analysis also showed that a significant predictor of the police NFA-ing a case was when they express doubt over victim credibility* (6x more likely to NFA in this case). Witnesses being present and the offence being recorded or photographed were significant predictors of the case being submitted to the CPS - if bystanders were present the case was 38x more likely to go to the CPS, and if recorded then 5x more likely. Injury level and weapon usage were also predictors of submission to CPS - if injury levels were moderate to high a case was twice as likely to be submitted, and if a weapon was used they were 4x more likely. Attrition stage Number and proportion of cases Suspect identified 247 (88%) Suspect arrested 133 (47%) Suspect interviewed under caution +3 79 (28%) Victim withdraws support 141 (50%) Police NFA 73 (26%) Submitted to the CPS 70 (25%) CPS charge 52 (18%) Conviction 12 recorded (Unclear) ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME - Low number of convictions – 12 recorded – possibly more (data accuracy). - Sometimes other case resolutions were recorded, such as community resolution and cautions (2%, n7), but this was not thought to be consistently captured. * Within the deep dive coding, police expressing doubts over victim credibility relates to officers documenting perceived contradictions in terms of victim behaviour and or account based on evidential factors. ■ ■ ■ ■ ■ ■ - ■- ■ ■ ■ ■ ■ MAYOR OF LONDON OFFICE FOR POLICING AND CRIME Glossary Police or Criminal Justice terms Most Similar Force (MSF): a group of forces determined by the Home Office to be comparable based on social demographic and economic characteristics. The similar group to the MPS includes Greater Manchester, West Midlands and West Yorkshire police force areas. Released Under Investigation (RUI): a term that relates to an individual that is suspected of a criminal offence being released whilst the police investigation proceeds. Bail: a set of restrictions imposed on a suspect to ensure they comply with a police investigation or before a court hearing or trial. Suspect: an individual suspected of a crime but not charged. Accused: an individual charged or cautioned for a crime. No further Action: a recorded outcome of an offence whereby no further action is taken. Gang: a gang is defined as defined as a "…relatively durable, predominantly street-based group of young people who see themselves (and are seen by others) as a discernible group, and engage in a range of criminal activity and violence. Gang (gang involvement): gang involvement is any offence where the suspect or victim is involved in gangs, or where intelligence indicates that the intended target of the offence was a known gang member. Group Violence: any violence offence involving a suspect group of three or more individuals. First Time Entrants: term to describe an individual who has not previously been charged or cautioned – e.g., no previous experience of the criminal justice process. Crime Reporting Information System (CRIS): the MPS system for recording reported crimes and log investigations. Police National Computer (PNC): the principal police database used by the police containing text information relevant to policing. Gangs Violence Matrix (GVM): an intelligence tool the MPS use to identify and risk-assess gang members across London who are involved in gang violence. Priority Robbery Offenders (PRO): a priority robbery offender is any individual that has come up at least three times as a Suspect on CRIS for robbery personal property in the last two years. Habitual Knife Carriers (HKC): any individual that has come up at least twice as a suspect on CRIS in the last two years or once on CRIS and once on intelligence in the last six months for Possession of Offensive Weapon / Knife / Bladed Article or a Knife Injury excluding Domestic Abuse offence. At least one of these offences has to have been committed in the last twelve months unless they have been in custody or have more than three offences. Priority Firearm Offenders (PFO): any individual that has been named twice as a suspect on CRIS for Possession of a Lethal Barrelled Firearm or a Lethal Barrelled Gun Crime Discharge offence in the last 5 years or have featured as a suspect and there is recent intelligence in the last 6 months relating to firearms (this includes intelligence around having access to firearms, supplying, importing or discharging a firearm). Licence: a set of conditions imposed that if broken can lead to prison recall. Community Order: a non custodial sentence which requires an offender to perform community service, curfew, undergo treatment (or such) instead of going to prison. Cambridge harm index(CHI): a tool that seeks to measure the seriousness of different crime types as opposed to the volume. Technical and other useful words Regression: a statistical technique to predict or visualise the relationship between different features/variables. Linear and logistic are both types of regression whose selection depends upon the data (i.e., binary or continuous). Regression Coefficient: a regression coefficient is a value used within regression to estimate the influence of the predictor variable. Cultural sites: published as part of the Greater London Authorities Cultural Infrastructure Plan, documenting places where culture is experienced or produced (this covers sites such as galleries, cinemas, libraries, theatres, music venues, parks, legal street art walls. Cultural Infrastructure Map - London Datastore Odds ratios: a statistical measure of the level of association between variables, representing the odds that an outcome will occur. Case coding / deep dive coding: a research technique where specific police records/data are systematically coded up to allow for analysis. National Data Quality Improvement Service (NDQIS): a tool to assist forces with improving data quality and comparability for knife crime. Currently used by 36 of the 43 police forces in England and Wales. Rolling 12 month trend: where the most recent 12 months are viewed together as a single period. Normalised figure: a technique that refers to the scaling down of the data set so that new data falls between 0 and 1 to enable comparisons. Useful offence categories Violence Against the Person (VAP): a crime grouping that includes a range of offences from minor offences such as harassment and common assault, to serious offences such as murder, actual bodily harm and grievous bodily harm. Violence without injury: offences of assault without injury (common assault, harassment etc) where the person is assaulted, obstructed, controlled – including when the level of injury is none, minor, or where other property such as clothing is damaged. Violence with injury: offences of assault with injury (murder, attempted murder, serious wounding, assault with injury etc.) where the person is assaulted and received injuries. The MPS include homicide within this classification. Weapon Enabled Robbery: theft with the use of force or the threat of force, where a weapon is seen, intimated, or used. Most Serious Violence (MSV): similar to violence with injury and relates to the most serious categories of violence (murder, attempted murder, serious wounding etc.). Serious Youth Violence: victims of most serious violence aged between 1 and 24 years of age at the time of the offence. Knife crime with injury victims not-DA (1-24): victims of selected crime types where a feature code of ‘knife used to injure’ aged between 1 and 24 years of age at the time of the offence. Gun crime: selected crime types where a feature code identifying weapon usage has been added to the report - firearm fired, firearm used to injure (as a blunt instrument), firearm used as a threat or firearm intimated. Knife crime offences: selected crime types where a feature code identifying weapon usage has been used - knife used to injure, knife seen, knife intimated. Does not include possession. Weapon-enabled offending: combined gun and knife crime as above. London Gang Exit: Final Impact Evaluation Report Dr Tom Davies & Dr Paul Dawson MOPAC Evidence and Insight CONTENTS Executive Summary .................................................................................................................... 2 1. Introduction ........................................................................................................................... 4 1.1 Background ....................................................................................................................... 4 1.2 London Gang Exit Overview ............................................................................................. 6 1.3 LGE Delivery Model .......................................................................................................... 7 2. Evaluation Methodology ........................................................................................................ 9 2.1 Evaluation Sample Selection ............................................................................................ 9 2.2 Understanding Impact: Offending & Victimisation Analysis .......................................... 10 2.3 Process Learning ........................................................................................................ 11 3. Programme Data .................................................................................................................. 12 3.1 The Overall LGE Cohort .................................................................................................. 12 3.2 The LGE Evaluation Cohort ............................................................................................. 12 3.3 Context and Concerns .................................................................................................... 13 3.4 Mutually Identified Intervention Pathways ................................................................... 14 Case Study 1: Richard’s Story ............................................................................................... 16 Case Study 2: Shaun’s Story ................................................................................................. 18 3.5 LGE Intervention Data .................................................................................................... 20 4. Exploring the Impact of LGE ................................................................................................. 24 4.1 Offending over lifetime of the LGE cohort ..................................................................... 24 4.2 Proven Offending - before versus after ......................................................................... 25 4.3 Impact of Offending against a comparison group ......................................................... 27 4.4 Exploring an impact on Victimisation ............................................................................. 28 4.5 Other Assessments of Impact ........................................................................................ 30 5. Discussion ............................................................................................................................. 33 Appendix 1: Offending & Victimisation Sample Methodology ................................................ 36 Appendix 2: LGE Core Outcomes ............................................................................................. 37 1 Appendix 4: Offending & Victimisation across sub-groups ..................................................... 39 Appendix 5: Eligibility Criteria .................................................................................................. 40 References ............................................................................................................................... 41 Acknowledgments MOPAC would like to express their gratitude to the Safer London team, and particularly Lisa for her efforts around data provision and queries. We would also like to thank Dr Toby Davies for his advice and input regarding the ‘shuffle’ methodology, as well as Hasnayn at King’s College London. 2 Executive Summary London Gang Exit (LGE) is a holistic support service to help individuals involved in gangs or at risk of gang violence or exploitation. Originally jointly commissioned by the Mayor’s Office for Policing And Crime (MOPAC) and the London Community Rehabilitation Company (CRC)1, the intervention went live in February 2016. MOPAC took over sole funding of the project from October 2017 and LGE is currently funded to March 2022. The programme is being delivered by Safer London with key delivery partners Catch-22 and the Department of Work and Pensions (DWP). The pan London service is designed to complement and enhance existing local services, matching need to bespoke support across mental health, employment, family support, housing advocacy and specialist support for girls and women. Through this programme of support, the key overarching aims for LGE are: • Reduced / ceased harm from gangs; • Reduced / ceased involvement in gangs; and • Reduced seriousness and frequency of offending.2 The Evidence and Insight (E&I) Unit are the Mayor’s Office for Policing and Crime’s (MOPAC) in-house social research and analytical team were commissioned to evaluate the service. Over the course of the programme, E&I have produced two interim evaluation reports focusing on the process of implementation as well as three indicative analyses of offending and victimization to inform Payment By Results (PBR) awards over the life of the programme. Results Summary of performance insights  Since launch, the programme has received over 2000 referral enquires and over 900 children and young people have been referred onto the programme, with 270 having completed the full six-month intervention by the end of FY2020/21.  Referrals have been received from all London boroughs, across statutory agencies and including self-referrals. Referrals generally align to boroughs with well documented gang and serious youth violence issues.  The LGE clients were predominantly young, male and black. The average age of individuals completing the programme is 19; 70% are male and 89% BAME. Senior managers highlighted a range of structural inequalities and referral bias that may contribute to these demographics.  The group present many vulnerabilities. Large proportions had needs relating to personal safety (77%), two thirds were assessed as having ‘gang exit’ needs, nearly 1 Withdrew as a partner in late 2020. 2 Please see the Appendix 2 for a full list of the outcomes. The current key outcomes refer to reducing/ceasing harm from and involvement in ‘serious youth violence’ rather than ‘gangs’, considering victimisation, severity and frequency of offending. This is reflective of the subjective nature of gang assignation. 3 half were assessed as having housing needs, with just over a third having ETE (Education, Training and Employment), health and wellbeing or offending needs.  Most clients presented multiple needs; the average number across all valid cases was 3.4 (n=770). Just over a fifth of cases listed 5 or more needs (22.5%, n= 171).  84% (n=542) had at least one sanction at any point in their criminal history. This group were responsible for a total of 7020 sanctions at an average of 10.8 sanctions each (ranging from 1 to 62).  At least 37% of the in-scope LGE cohort have been on the MPS Gangs Violence Matrix (GVM) at some point (n=239).  The intervention on average lasted just under 6 months. Impact on Offending & Victimisation  Impact was undertaken through a variety of research methods (broadly aligning to programme aims); proven reoffending before and after the programme3, police data in victimisation, as well as staff and client assessments.  When examining proven offending, results demonstrate a non-significant reduction in overall offending. However, there is a significant reduction in the proportion of violent offenders in the 24 months after start date compared to before. To illustrate; - 43.3% of clients offended in the 12 months prior to LGE versus 40.5% in the 12 months subsequent to LGE (not a significant difference); - 61.1% of clients offended in the 24 months prior to LGE versus 59.1% in the 24 months subsequent (not a significant difference); - 37.4% committed a violent offence in the 24 months prior to intervention start, compared to 28.7% in the 24 months after the start date. This is significant; and - Deeper analysis identified the specific reduction in violence occurring within year two, in particular around the 18-month mark.  Exploring victimisation in the 12 months pre and post intervention produces similar findings to previous reports, with statistically significant decreases across all, violent and serious number of incidents of victimisation. For example, for all LGE cases 38.6% were victimised in the 12 months prior compared to 26.8% in the 12 months subsequent (at a reduced rate from 0.73 to 0.48). These are significant differences.  Across measures of case worker risk assessment and mutually assessed distance travelled findings were positive on the progress made - particularly against key outcomes such as gang involvement - while valuable, this should be caveated by data quality issues, the subjective nature of the reporting tool and absence of a control.  This final report - which focuses on the impact of LGE - should be read in conjunction with previous published reports and brings the evaluation to a close. 3 Analysis considered all, violent and serious proven offending. 4 1. Introduction 1.1 Background Since 2014 England and Wales has seen a substantial and well documented rise in serious violence – be this knife crime, knife violence with injury or homicide.4 A key issue for London over recent years has been continuing concerns over the prevalence and severity of serious violence involving children and young people, and the role that gangs play in this. This issue was a key focus of the 2017-21 Police and Crime Plan, which set out strategies to address the problem of gang related violence, offending and exploitation. The Mayor of London’s Knife Crime Strategy also addressed the wider but overlapping problem of knife-related violence in the capital. Whilst recognising the continuing need for enforcement, both approaches also emphasise prevention and intervention, highlighting the importance of joined-up, diversionary approaches in helping vulnerable children and young people out of serious offending and violent victimisation. It is in this context that London Gang Exit was devised. In the five years since LGE started, a stronger emphasis on early intervention and prevention has been apparent, as can be seen in the launch of the Mayor of London’s Violence Reduction Unit (VRU) in 2019, forwarding a holistic, public health orientated approach. While serious violence involving children and young people remains a key concern, there have been some recent positive signs that this violence has begun to stabilise in London – particularly the more severe type of offending. Table 1 below demonstrates recent decreases seen in knife crime with injury and gun crime offences. Although the COVID-19 pandemic has clearly impacted on all serious violence, decreases in serious gun and knife violence were also evident before the onset of lockdowns in FY2020/21. 4 It should be noted that improvements in police crime reporting practices contributed to some of the increase post 2014. 5 Table 1: Violence and weapon enabled crime by year 1.1.2 Gang Violence and Exploitation Despite these recent positive indications on violence more generally, gangs continue to have a significant impact on the landscape of violence in London, with 55% of Londoners believing that gangs are a problem in their local area.5 Gangs range from organised criminal networks involved in Class A drugs supply and firearms, through to street-based gangs involved in violence and personal robbery6. Difficulties in accurately measuring this have been widely acknowledged,7 and it is important to recognise the subjective nature of gang terminology, but with these caveats in mind it is useful to present what proportions of certain offences are believed to be related to gangs. Seemingly, the proportion of ‘gang involvement’ would appear to have reduced across many violence measures (i.e., homicides, knife injury under 25, lethal barrel discharge) – see Table 2. Nonetheless, gang-related violence still accounts for a significant proportion of the most serious violence in London; over half of shootings and nearly one quarter of homicides are believed to be linked to gangs. 5 MOPAC Public Attitudes Survey (PAS) 55% think gangs are a major/minor problem in their local area (FY 20/21 – Q3/Q4; MPS level response). Please note that this figure is not directly comparable to previous years due to changes in survey methodology. 6 These estimates are dictated by a number of factors, particularly the definition of ‘gangs’ applied, thus it should be noted that the figures can vary across sources. Particularly there are gangs who are involved in less serious crimes, whose activity will not be reflected by these figures. 7 See, for example Update on Gangs Matrix Review (February 2021) Pre COVID 2018/19 2019/20 2020/21 % change (2018/19 vs 2019/20) % change (2018/19 vs 2020/21) % change (2019/20 vs 2020/21) Violence Against the Person 209,121 219,029 224,559 5% 7% 3% Violence with Injury 77,145 73,484 67,646 -5% -12% -8% Homicide 139 144 127 4% -9% -12% Youth Homicide 29 30 33 3% 14% 10% Serious Youth Violence 7,800 7,651 5,866 -2% -25% -23% Non-Domestic Abuse VWI 53,137 49,578 44,732 -7% -16% -10% Violence withouth Injury 131,837 145,401 156,786 10% 19% 8% Possession of Weapons 7,647 6,966 5,869 -9% -23% -16% Knife Crime Offences 14,850 15,604 10,366 5% -30% -34% Knife Crime With Injury 4,289 3,834 3,104 -11% -28% -19% Knife Crime With Injury victims under 25 (non DA) 1,772 1,547 1,170 -13% -34% -24% Knife Crime with Injury Personal Robbery 598 533 414 -11% -31% -22% Knife Possession 5,813 4,973 4,534 -14% -22% -9% Gun Crime Offences 2,391 2,073 1,580 -13% -34% -24% Gun Crime Lethal Discharge 393 281 268 -28% -32% -5% Gun Crime Personal Robbery 597 446 296 -25% -50% -34% Violence Against Person Weapons Post COVID 6 Table 2: Gang related violence by year 1.2 London Gang Exit Overview London Gang Exit (LGE) is a holistic support service to help individuals involved in gangs or at risk of gang violence or exploitation. Originally jointly commissioned by the Mayor’s Office for Policing And Crime (MOPAC) and the London Community Rehabilitation Company (CRC), the intervention went live in February 2016 and was initially funded until October 2017. MOPAC took over sole funding of the project after this date and LGE is currently funded to March 2022. LGE also received contributory funding from the Mayor’s Young Londoners Fund between 2018-21. The programme is being delivered by Safer London with key delivery partners Catch-22, Department of Work and Pensions (DWP)8 and the community CRC9. The pan London service is designed to complement and enhance existing local services, matching need to bespoke support across mental health, employment, family support, housing advocacy and specialist support for girls and women. The programme works with children and young people between the ages of 15-24 years on a voluntary basis. The key overarching aims for LGE, measured through a combination of police recorded data and a practitioner/client ‘distance travelled’ assessment tool are: • Reduced / ceased harm from gangs; • Reduced / ceased involvement in gangs; and • Reduced seriousness and frequency of offending.10 8 DWP contributed to programme funding 2016-17, providing a secondee from 2018 onwards. 9 Withdrew as a partner in late 2020. 10 Please see the Appendix 2 for a full list of the outcomes. The current key outcomes refer to reducing/ceasing harm from and involvement in ‘serious youth violence’ rather than ‘gangs’, considering victimisation, severity and frequency of offending. This is reflective of the subjective nature of gang assignation. 7 1.3 LGE Delivery Model LGE has seen a number of changes to its delivery model, as would be expected for a programme in its sixth year. This section outlines the delivery model in place currently, before briefly highlighting key changes. It should be noted that changes to programme delivery can complicate the evaluation in terms of understanding impact and issues pertaining to programme integrity. To be eligible for referral to LGE, the young person (male or female) is required:  To be aged between 15 and 2411;  To be involved in or at risk of becoming involved in serious youth violence OR to be at significant risk of harm from or exploitation by those involved in serious youth violence; and  To show some motivation to want to move away from violence and a willingness to engage in the service. Referrals can come from any agency, voluntary and community sector organisations, family or children and young people themselves. Incoming referral forms are reviewed by the Referral and Assessment Team,12 and through dialogue with the young person to inform risk and motivational assessments. Once accepted on the programme, the young person will have a detailed assessment to determine the core and specialist strands most suited to their individual needs. Referrals can be made for specific, specialist parts of the LGE service (e.g., family support, mental health, housing) or for more generic support/casework:  Core Services – providing one to one support / mentoring; cores services include targeted support for girls and young women and those with substantial health needs.  Specialist Services – providing one to one specialist support across the following areas: o Education Training & Employment (ETE); o Emotional Wellbeing; o Family & Community; and o Housing. The intervention is designed to last a minimum of six months (this has fluctuated over the course of the programme due mainly to the variance in client need), with individuals regularly updating their perceptions of progress within core and specialist strands with their 11 Individuals outside this range may be supported in exceptional situations where this is considered necessary, assessed on a case by case basis. 12 Initially referrals were processed by CRC, which caused issues with communication. Therefore, a decision was made to move the referral system into LGE’s offices. 8 caseworker via the distance travelled tool. Onward referrals can also be made if the young person is deemed to have specific needs to those covered by the LGE services. Referral pathways were changed and strengthened over the course of the programme13, whilst the exact nature of the support packages offered to clients also varied over time as the programme adapted to structural/organisational changes, staffing churn at all levels, funding and resource fluctuations and the impact of COVID-19. In recent years the overarching strategy and ethos of Safer London has shifted, positioning itself as a conduit to the community, and offering a centralised, ‘single front door’ policy for referrals to its range of youth support services including LGE. The COVID-19 pandemic has necessitated significant changes to service delivery. Whilst service delivery has now reverted to face-to-face interaction, in line with Government advice Safer London moved all client work engagement online for significant periods of 2020 and 2021.14 Safer London have used three different case management systems over the course of the programme to date.15 This has caused a number of issues with data extraction and quality, with the first year’s programme data being particularly affected in terms of needs and vulnerabilities and ‘distance travelled’ outcomes data. These issues meant that data requests for the final evaluation had to be revised to mitigate impact on Safer London resources. For example, it was not possible to retrieve activities and outcomes on separate intervention strands or conduct thorough analysis of how well interventions were matched to needs. Whilst the programme delivery and case management system changes described above are not unexpected for a multi-year programme, it should be noted that they can present risks to programme integrity and complicate the evaluation of impact. This is particularly true of the changes to the eligibility criteria which directly affects the client base being worked with. Shifting several times over the course of the intervention, based primarily on varying degrees of evidence of gang involvement required. The current criteria encourage referrals from a wider base including lower risk individuals involved in serious youth violence. 13 For example, LGE ran a Trauma Unit referral pathway between 2017 and 2019, whilst more recently Safer London’s parallel Rescue & Response service has received LGE referrals for individuals involved in County Lines 14 Feedback from practitioner interviews/monitoring meetings suggested this was sometimes beneficial to engagement. 15 Excel based (2016- March 2017; Lamplight (April 2017 – August 2020); Apricot (September 2020 to date). 9 2. Evaluation Methodology This section sets out the methodology used within the Evaluation. Evidence and Insight (E&I) are the dedicated analysis and research function within the Mayor’s Office for Policing And Crime (MOPAC). The team is an amalgamation of a long-established analysis function within MOPAC and a team of social researchers that moved across from the MPS in 2014. E&I were commissioned to undertake a multi-year mixed methods evaluation of the LGE programme. So far this has resulted in a suite of published reports delivering a range of learning on the set up, delivery, process and indicative impact – however this final report focusing on impact of LGE against its stated outcome completes the evaluation. The objectives of the E&I 2021 final LGE impact evaluation are to:  Outline the key performance data captured across the duration of the LGE programme;  Assess LGE’s impact on the client’s proven offending, focusing on violence and serious offences; and  Assess LGE’s impact on police recorded victimisation, focusing on violent and serious offences. 2.1 Evaluation Sample Selection 2.1.2 Programme Data The LGE cohort in scope for evaluation consists of all accepted referrals between programme start in February 2016 and March 31st 2020 (end of year 4). The cut off to allow sufficient time for offending and victimisation data capture post intervention. In total there are 782 cases in scope, featuring 756 individuals (26 individuals were referred on the programme on two separate occasions). The evaluation cohort represents 80% (n=782/983) of all accepted referrals over course of the programme to date (31/03/2021), and 97% of completed cases (n=263/270).16 2.1.3 Impact Analysis A rigorous matching process was carried out on the 756 individuals in the evaluation cohort to identify records on both the Police National Computer (PNC)17 and MPS Crime Reports 16 Current case status was updated for the evaluation cohort until February 2021.Not all completed cases are included in the impact analysis. 17 The Police National Computer (PNC) is a system that stores and shares criminal records information across the UK. 10 (CRIS) for victimisation. 18 Please see Appendix 1: Offending & Victimisation Sample Methodology for a detailed explanation of this process. After matching, 70% of individuals were eligible for 12 months pre/post proven offending analysis (n=524) and 79% of individuals were eligible for 12 months pre/post police recorded victimisation analysis (n=594). 2.1.4 Gangs Violence Matrix The LGE cohort was also matched to the MPS Gangs Violence Matrix (GVM) to ascertain which individuals had featured on the GVM at any point. Changing data collection methods and missing demographic or PNC-ID data mean that accurate matching was not possible across the entire LGE cohort. Nonetheless, analysis indicates that as would be expected, a notable proportion have appeared on the GVM. At least 37% of the in-scope LGE cohort have been on the MPS Gangs Violence Matrix (GVM) at some point (n=239). At least 32% (n=205) were on the GVM before their referral to LGE, with a sizeable proportion of these (73%, n=148) remaining on the GVM post LGE intervention. At least 32 individuals were placed on the GVM for the first time after their LGE intervention (it should be noted that these additions were not related to LGE referral). 2.2 Understanding Impact: Offending & Victimisation Analysis To explore impact, the report draws on a number of data sources. Overall metrics on offending are drawn from the Police National Computer (PNC). These relate to proven reoffending; “A proven reoffence is defined as any offence committed in a one-year follow-up period that leads to a court conviction, caution, reprimand or warning in the one-year follow- up or within a further six-month waiting period to allow the offence to be proven in court”. This metric is the most robust metric for police recorded offending – although still presents the caveats of not capturing all crime (i.e., unreported crime). Victimisation analysis draws on data from police crime reports (CRIS), meaning similar caveats should be applied in terms of unreported crime. The MPS Gangs Violence Matrix (GVM) is also used to understand risk of violence and gang involvement. The basic analysis compares offending and victimisation before and after the LGE intervention, for a period of 12 and 24 months. This is a useful assessment and the 24-month timeframe is able to provide a longer follow on time than many evaluations. However, such a methodology is not able to robustly assess impact as it does not compare results against anything and is subject to a wide array of limitations (i.e., seasonality etc). 18 Crime Reporting Information System (CRIS) - an electronic management system for the recording and processing of crime in the MPS. 11 In order to understand impact on offending and victimisation robustly the development of a counterfactual or ‘control’ group through which to compare outcomes is essential (i.e., what would have happened in the absence of the treatment). There are numerous methods for creating the counterfactual – however, the design and implementation of LGE presented many challenges in terms of deciphering impact. For example, although there was a criteria for entry to the programme, this changed over time and was subjective – resulting in selection bias that could not be modelled, making many statistical techniques (i.e., Propensity Score Matching) infeasible. Similarly, other complex methods such as Regression Discontinuity Design (RDD) was not possible as there was no single selection score determining acceptance on LGE. In the absence of a suitable existing methodology, a novel quasi-experimental approach was developed inspired by 'the Knox Test', a technique commonly used in the spatio-temporal analysis of crime. This approach was developed in collaboration with University College London’s (UCL) Centre for Global Cities Policing and originally utilised to assess the impact on offending of the MPS Gangs Violence Matrix (GVM) (MOPAC 2018). 2.3 Process Learning As indicated above, the process of implementation is not covered in this final LGE impact evaluation report. Previous interim reports have covered process learning in detail, with Quarterly Monitoring meetings also acting as a forum where process barriers and benefits are explored. 12 3. Programme Data Performance data is a key ingredient in understanding the basic delivery of the LGE – that is - what has been delivered, when, and to whom. It enables the evaluation to consider the delivery of the programme – as well as factors such as cohort demographics, needs, harm and vulnerability and information on soft outcomes measures such as self and practitioner assessed ‘distance travelled’. This section of the report provides a brief overview of performance data to date (up to March 31st, 2021), before focusing in more detail on the data relating specifically to the impact evaluation cohort (up to March 31st, 2020). 3.1 The Overall LGE Cohort In terms of an overview of the programme to date; up to 31st March 2021, LGE received 2229 referral enquires, of which 983 were accepted referrals. This has resulted in 703 opened cases, 270 of which had been successfully completed. Appendix 3: Performance provides a detailed breakdown of throughput by year. The following section focuses on the LGE evaluation cohort, that is the 782 cases featuring 756 unique individuals accepted onto the scheme between January 2016 and March 2020. 3.2 The LGE Evaluation Cohort The LGE cohort is predominantly young, male and black (see Table 3 below). Overall nearly 90% of referrals were male and this was consistent across the period. Female referrals slightly increased, from 8.3% in year 1 to 10.6% in year 4, with a peak of 15% in year 3 but were always the minority of cases. The mean age at referral across the period of analysis was 19.5, with Year 4 seeing the oldest mean (20.3 years). The youngest age at referral was 12, and the oldest 32. The median age for referral was 19. The largest demographic group across the overall period was Black (63.5% of clients) which remained broadly stable. Outside of this, the proportion of overall referrals who identified as white19 was 13% (n=92) (also remained stable) and there was a notable increase in the number of referrals identifying as mixed ethnicity over the evaluation period, increasing from 6.6% (n=14) in year one to 18.8% (n=33) of referrals in Year 4. This was offset by smaller decreases in the number of Black and Asian referrals. Looking at the four-year period overall, age and ethnicity are broadly comparable with that of the MPS Gangs Violence Matrix (MOPAC 2021); the LGE cohort however features a greater proportion of females. It is important to acknowledge the likelihood that cohort demographics are influenced by a variety of external factors, and senior managers at Safer London highlighted young black- males they worked with as particularly affected by issues concerning referral bias and 19 Safer London used the 16+1 self-classification for ethnicity. A full breakdown of SCE for the evaluation cohort is included in the Appendices. 13 adultification, driven by wider inequalities and social injustice (Lammy 2017; Treloar & Begum 2021). Table 3: LGE client demographics 3.3 Context and Concerns A fundamental objective of the LGE programme was to provide a bespoke, tailored response to children and young people’s individual needs. The programme data has enabled this to be explored to a degree - unfortunately some of the data captured across the course of the programme was lacking, due mainly to changing recording processes across the various case management systems used. The available data still allows insight into the levels of needs and vulnerability within the cohort, albeit in a somewhat inconsistent manner. Table 4 below presents an overview of the various strands of vulnerabilities, behaviours and concerns data, including information from referral assessments, and ongoing case management. Year 1 Year 2 Year 3 Year 4 Overall Male 91.7% 90.2% 85.5% 89.8% 89.4% Female 8.3% 9.8% 14.5% 10.2% 10.6% Average 19.6 19.0 19.2 20.3 19.5 % Under 18 34.7% 34.8% 35.8% 19.8% 31.1% % 18 to 24 56.5% 60.4% 58.7% 68.0% 60.8% % 25 plus 8.8% 4.9% 5.6% 12.2% 8.1% White 13.7% 12.2% 13.0% 13.1% 13.0% BAME 86.3% 87.8% 87.0% 86.9% 87.0% Mixed 6.6% 13.5% 13.0% 18.8% 12.6% Black 65.6% 64.1% 64.8% 59.1% 63.5% Asian 10.4% 7.7% 4.3% 5.1% 7.1% Other 3.8% 2.6% 4.9% 4.0% 3.8% Individuals 216 164 179 197 756 Gender Age Ethnicity 14 Table 4: Combined Vulnerabilities, Behaviours & Concerns Within each of the broad categories variation exists. With the exception of ‘context on referral’, assessments are dynamic and change over the course of the programme as individual’s relationship with their caseworker develops or new information comes to light. 3.4 Mutually Identified Intervention Pathways The LGE intervention is focused on building trusted relationships between children and young people and their support workers to encourage mutual identification of needs and subsequent intervention pathways on a one to one basis.20 Figure 1 illustrates the proportion of the evaluation cohort identifying needs matched to the eight core and specialist intervention strands. Unsurprisingly, large proportions had needs relating to personal safety (77%), whilst just over two thirds were assessed as having ‘gang exit’ needs. Nearly half were assessed as having housing needs, with just over a third having ETE, Health & Wellbeing or Offending needs. Most clients presented multiple needs; the average across all valid cases was 3.4 (n=770). Just over a fifth of cases listed 5 or more needs (22.5%, n= 171).21 20 In order for a need to be recorded and progressed as an intervention pathway the young person must themselves identify the area as a need. 21 Within the one-to-one needs assessment, clients also define their primary need in terms of LGE support. Illustrative of the vulnerability of the cohort, ‘Safety’ was the predominant theme, accounting for 57% (n=440) of all primary needs. ‘Gang Exit’ made up just over one fifth of primary needs (22.5%, n=174), whilst housing accounted for 10.5% (n=81). All other needs were less than 4% of the total. n Cohort % Context On Referral Known to Social Services 229 469 50% Disability 181 519 35% Benefits 167 528 33% Current/Former Looked After Child (LAC) 133 756 18% Homeless at referral 109 616 18% Caring Responsibilities 83 519 15% Designated Child In Need (CiN) 62 756 8% Current Child Protection Plan (CPP) 47 756 6% Crime & Environmental Gang association 610 756 80% Explicitly linked to named gang 150 756 20% Weapon Carrying 191 540 35% Victimisation 270 756 36% Violent Victimisation 135 756 18% Sexual violence or exploitation 50 756 7% Family Safeguarding Concerns 260 540 48% Risk of Homelessness 168 540 31% Experienced Domestic Violence 77 756 10% Health & Wellbeing Substance Abuse 252 756 33% Mental Health Concerns 116 540 21% Employment & Education Unemployed/NEET 361 756 48% SEN or other concerns 97 756 13% 15 Figure 1 –Intervention Strand The distribution of need did not remain constant over time; the most recent period saw a more even distribution between Safety (33.6%), Gang Exit (29.4%) and Housing (22.7%), indicating a change in the type of referrals accepted over the duration of the programme. The number of clients with multiple overall needs increased over the course of the programme, with clients being assessed as having a greater number of needs in the most recent year 4 (4.1 needs, 39% with 5 or more, n= 83). Direct comparisons to the plethora of contextual information presented in the previous section are not possible given their dynamic and crosscutting nature. For example, whilst 20% of clients were officially homeless (no fixed abode) at the point of referral, complex housing situations including hostels, refuges and unsafe or crowded accommodation mean that a far greater proportion identified a need in this area. Similarly, ‘Gang Exit’ suggests a focus on involvement within a gang which one third of the cohort did not identify with, being affected by serious youth or group orientated violence in different ways. Such nuances are powerfully demonstrated in the two cases studies below. Overall, we see a disadvantaged, vulnerable group with many individuals presenting multiple and complex needs. Although gaps in the data does not allow for more granular analysis matching intervention to need it is clear these broadly align, it is also clear that the cohort is made up of a variety of (sometimes intersecting) groups; from those more deeply entrenched in gang related criminality, to those being exploited by gangs to those on the peripheries but ‘at risk’. 16 Case Study 1: Richard’s Story “I was sold a lot of dreams, with no receipts to take them back” Richard experienced a lot young. From being expelled from mainstream education, a victim of extreme violence and imprisoned young – growing up in London wasn’t what it should be in a city renowned for endless opportunities. When life is hard, you can give up. But Richard came out of the other side determined and ready to fight for his future. “We grew accepting the block, the lifestyle and everything that came with it as normal, but it wasn’t” Growing up all the kids had on the estate was a park with one swing and a broken seesaw. The youth clubs were shut all the time. We had no one teach us about life, or take us on trips. There were no programs, or anything interesting to keep kids off the street. The police always thought we were up to something, even when we weren’t. The first time I got stopped was when I was riding my bike from the shop. A police man started running and yelling ‘I see three males’, I was by myself and didn’t think nothing of it. Then I was dragged of my bike and asked a lot of questions which sounded more like accusations. I found out I was a minority and I wasn’t sure how to feel about it. We feel as if we were judged left, right and centre. If you’re black or Asian you’re considered a minority who will not make it to 30. 17 My older brother was in a gang so I was exposed to a lot young. By primary school I knew about girls, sex and drugs. I was interested by the lifestyle and the older I got, the more intrigued I became. Until eventually l decided to join the gang. I wanted to feel power and respect. A gang is what we were referred as, but I saw them as family – as my brothers and sisters. I felt no one cared about me or helped me as much as they did. What we went through, we all went through. I didn’t feel I could speak to anyone but my brothers, they were all I trusted and all I knew. But family, loyalty and love were all illusion, they never really existed. They just made me think that. It started off being fun and mischievous, but turned violent and evil, which seems to be a cycle never ending. In the end I was willing to take a life for an area that doesn’t belong to me and for people who will continue their lives as normal, even when I’m gone. “I lost years of my life that I would never get back. I went in a boy and came out a man. I knew I never wanted to waste any more of my time not contributing to my future.” Once the verdict was read my bubble burst. I went from dreaming to reality. I realised I wasn’t invincible, I had been judging all my life and now it was my time to be judged. Too many people are sent to overcrowded prison units and not given the help and support they need. Some people complete their sentences without doing any work or making any progression. There’s times they leave the same person as when they came in. Without knowing any better or being any different and end up back to where they started. More time should be taken to find out what young people want, rather than what they need. I grew older and became wiser, saw things for what they were rather than what I would like it to be. It was all lies, illusions and time wasted – but I could not waste anymore. I understood myself better, what I wanted and where I was going. Reading became my mind’s best weapon. Emotions become something I could feel and learn to control. “Even though we hadn’t been through the exact same things, it was similar and we could trade stories for days” I never got along with probation, my younger eyes saw them as the enemy and not showing enough help and support. I was put in touch with Safer London. To be honest at the beginning before my first meeting with my Support Worker, I was hesitant. I wasn’t sure what I was getting myself into. But within minutes of meeting for the first time we connected instantly. He was easy to communicate and get along with. He taught me to grow, share and that I should never give up. We made a plan together and within a year we completed what we needed. I met Safer London’s housing officer, he worked with me and helped me a lot. Even though I was a man I never knew everything a man should, but he helped me learn. By the time I finished working with him I knew about bills, council and housing. We went through hours of council meetings and applications. He helped me into my hostel flat and we successfully got a grant, which I doubt I would have received without his help. 18 Case Study 2: Shaun’s Story “I knew I wanted to make a change in myself” Growing up in London for me was tough. It’s a place where there’s a lot of things happening and it’s hard for us as young people to find a different route in life. I’ve always been very ambitious, but no one ever really understood me or how ambitious I really was. Growing up there was some good parts and some bad parts. Family life was pretty poor. I didn’t really grow up with my parents. I grew up with my auntie and it was hard for me to talk to her about certain situations. She wouldn’t really understand what was going on, how much it was affecting me or what I was going through. There was a lot of different gangs coming into the area. They were selling drugs to young kids and telling them to go and make money. There were lots of people getting stabbed and sliced. The main issue in the area was that everyone was under a lot of peer pressure. “From a young age I got involved with a lot of wrong people and I associated myself with the wrong crowd.” I saw them as brothers from another mother. I thought they would support me and help me with anything I needed. Then things went sideways. I was staying out late and things just started building from there. My main issues were gang affiliations and County Lines. A lot of people were saying to me ‘you can make a lot of money doing this’. I had no one, I was getting moved from place to place. I thought that this was the best way to make money. My situation got very bad. I did a lot of wrong to a lot of people. Certain things I’m not proud of doing. My safety was out of control. I knew that I was going to have a lot of people after me, but I didn’t really care. “I knew I wanted to make a change in myself.” I almost lost my life three times. Someone stabbed me, someone tried to shoot me and I got beaten up. I know that once you die, you die. But I wasn’t ready to lose my life in that moment. I knew I wanted to make a change in myself. At one point I was living on the streets and I found myself in a hostel. It was whilst I was living in the hostel that I was referred to Safer London. Through them I received support for my mental health and my gang affiliations. “It was different from the other services I worked with because I didn’t get judged. There was never anything that I did that was right or wrong.” 19 What made it easier to get out of the lifestyle was having one person I knew who I could trust and rely on. There was a lot of consistency. I knew I always had someone to talk to when I needed them. Working with their mental health worker was a hard experience for me. I didn’t know what to expect. I didn’t know the type of angles they would take or how it would work. At the time my mental health was pretty bad. I wouldn’t engage with workers that I was meant to be engaging with. I would put off meetings with my social worker because I wasn’t feeling myself. The support was useful and beneficial because now I can engage with a lot more people. I feel more comfortable in myself now. Working with Safer London helped me come to terms with things. I thought I was right about a lot of things, but they made me see it from a different angle. The most helpful part was knowing what can occur that can make you re-offend. What behaviours that I or anyone else needs to make sure that it doesn’t happen again. “I’m now involved in Safer London’s Young Leader work. It’s been interesting because I’ve been able to hear a lot of different people’s stories and what they’ve been through. I know we can all come together and make a difference.” My life right now is a bit hectic! I’m taking care of my daughter, trying to look after myself and find work, and trying to maintain a healthy household. Without Safer London my life would most definitely have been different. I wouldn’t have learnt to have the self-control I need to make sure I don’t re-offend again. The encouragement I received gave me a lot of insight on what I could do to help myself. It helped me change my mindset and helped me really to become a better person. “It’s really made me feel like I’m myself now.” 20 3.5 LGE Intervention Data The report now turns to LGE intervention data – that is, referral pathways by organisation and borough, throughput and intervention length. 3.5.1 Referral Agencies Local Authority referrals made up over a fifth of total referrals (see figure 2). The source of referrals changed over the course of the programme. Earlier referrals were much more likely to be from the police, probation and CRC. As the programme became more widely known through continued engagement and promotion, the range of referring agencies increased. The work of Catch-22 is evidence by the proportion of secure estate referrals. Although outside the scope of this dataset, recent increases in VCS and self-referrals have been anecdotally attributed to the impact of COVID 19 and changing ways of working for Safer London. Figure 2: Referral Agency over the entire duration In terms of geographical distribution, there was a wide range of boroughs to refer into LGE. Referrals generally align to boroughs with well documented gang and serious youth violence issues. To illustrate, Lambeth accounts for 12.4% of all referrals over the period of analysis (n=96), nearly double the next highest location (Greenwich; 7.2%, n=56). Referrals were received from all boroughs, although 6 boroughs generated under 10 referrals accounts. The top ten locations for referrals account for 57.3% of all referrals (n=444). See figure 3. As with referral agencies, the geographical distribution of referrals changed over time. This was due to a number of factors including relationships built through Safer London’s precursor to LGE, ‘Safe and Secure’. As LGE expanded to include all boroughs, mutual agreements with Local Authorities saw the geographical scope of referrals widen, including to boroughs with less documented gangs and youth violence concerns. The latter point is to be expected given the remit of LGE to address gaps in service provision. 21 Figure 3: Referrals by Borough Due to the impact of COVID, enquires, referrals and open case load have fluctuated in line with national and local lockdowns, resulting in a backlog. Over the last 18 months Safer London have reported more young people presenting complex needs, with a greater proportion of individuals assessed as high ‘risk to self’ categorised by low self-esteem, thoughts of self-harm, anxiety and depression. While this observation is outside the scope of the evaluation data, an increase in the proportion of referrals with above average needs can be seen in Year 4 (pre COVID), with 60% indicating high needs compared to an overall average of around 40%. The pandemic has further created issues with following up clients following case closures. 3.5.2 Case Completions and Closures Table 5 presents case completions across the period of analysis; from the 782 referrals in scope:  34% were recorded as successfully completed (n=263).  42% (n=215) of cases were closed ahead of the initial meeting with the client or otherwise before being formally opened and agreement signed, usually due to client non-engagement.  37% (n=293) cases were closed without completion during the intervention period. o Of these, the majority relate to non or disengagement (n=189, 65%), whilst a further 22% (n=63) were closed due to the client being remanded in or returning to custody. 11 were closed because the service need was not met, and 10 cases were closed when the client moved out of London.  11 cases were open and ongoing. 22 Table 5: Case Closures and Completions 3.5.3 Intervention Length Looking only at cases completed, open or closed after opening (n=541), 60% (n=283) were opened within 21 days (the target timespan), rising to 89% opened within two months of referral (n=480). The median average days between referral and case opening for completed cases is 20 days. See figure 4. In terms of intervention length, there was considerably more variation. The average intervention length across the entire cohort was 176 days (just under 6 months). While a minimum 6-month intervention period was specified, process learning informed decisions to be flexible with an emphasis on client need rather than fixed targets. The graph below displays intervention length by Case Status. Figure 4: Intervention Length by Case Status 54% of successfully completed cases had an intervention length of between six and twelve months (n=141), rising to 73% when including interventions lasting over 12 months. Whilst n % n % n % n % n % Intervention completed 80 37% 60 36% 63 34% 60 28% 263 34% Closed - No Case / Pre Intitial Appoint 46 21% 45 27% 67 36% 57 27% 215 27% Closed - During Intervention 91 42% 64 38% 55 30% 83 39% 293 37% Closed - Non/Disengagement 62 29% 38 22% 31 17% 58 27% 189 24% Closed - Custody 16 7% 15 9% 15 8% 17 8% 63 8% Closed - Breached Service Agreement 3 1% 1 1% 1 1% 0% 5 1% Closed - Moved out of London 5 2% 2 1% 2 1% 1 0% 10 1% Closed - Service not net need 0% 7 4% 2 1% 2 1% 11 1% Closed - Unknown 5 2% 1 1% 4 2% 5 2% 15 2% Open 0% 0% 0% 11 5% 11 1% Grand Total 217 100% 169 100% 185 100% 211 100% 782 100% Year 1 Year 2 Year 3 Year 4 Total 23 cases closed due to the client going into custody typically lasted less than 9mths (85%, n=53), nearly half of these were between 6 and 9 months (n=24). The proportion of closed – ‘other’ cases is similar, with the majority of cases ending slightly earlier, within 6 months of opening (67%, n=150). This section has set out the available LGE performance data across the total duration of the programme. Data indicates that LGE received 2229 referral enquires, of which 983 were accepted referrals, from across London – albeit there was borough variation although these generally align to boroughs with documented challenges in this area. The LGE clients were predominantly young, male and black as well as well as presenting with a range of vulnerabilities. On average the intervention lasted just under 6 months and only 34% of cases were marked as successfully complete. 24 4. Exploring the Impact of LGE This section focuses upon examining the impact of the LGE programme – both on offending and victimisation that broadly aligns with the aims of the service in reducing harm, seriousness and frequency of offending, and reduced involvement in gangs. This is undertaken through a variety of methods; proven reoffending before and after the programme (i.e., offending), police data on victimisation (i.e., harm reduction), as well as staff and client assessments across a range of wider outcomes (including involvement in gangs). See appendix 2 for more detail. Before examining impact, the report first establishes the overall level of offending of the LGE group of individuals. 4.1 Offending over lifetime of the LGE cohort In terms of offending – almost all (91%, n=588/649) of the eligible LGE cohort had a Police National Computer record. 84% (n=542) had a last one sanction22 at any point in their criminal history. This group were responsible for a total of 7020 sanctions at an average of 10.8 sanctions each (ranging from 1 to 62). Within this group 27% (n=149) had received between 1 and 5 sanctions, whilst at the higher end of prevalence 20% (n=111) had over 20 sanctions each. Drugs offences make up the largest proportion of sanctions; 28% (n=1981). Of these, 50% (n=995) were minor cannabis possession offences, whilst one third were drugs supply offences (n=725, 37%). Theft and Kindred offences (including robbery) made up 18% (n=1276). At the other end of the spectrum, only seven Sexual Offence sanctions were identified (0.1% of all sanctions). In terms of the proportion of individuals sanctioned across offence types:  70% (n=456) sanctioned at least once for violence including robbery;  66% (n=427) sanctioned at least once for violent offence (Offences Against the Person, Weapons);  54% (n=350) had a least one sanction for a serious offence23;  38% (n=245) had at least one drugs supply sanction; and  25% (n=166) were convicted of at least one Public Order offence. Indicative of the wide-ranging offending behaviours presented by the cohort, it should be noted that seven individuals were convicted of murder during the period of analysis. Of these, 22 The majority are guilty disposals at court (n=6792). The remainder; Cautions (n=126); Reprimand (n=58); Warnings (n=44). 23 Serious offences include homicide, serious violence, firearms possession, robbery, serious sexual offences and drug trafficking. 25 six were convicted prior to beginning their LGE intervention (i.e., recorded as ‘Closed – custody’ or ‘Closed - no case’). One individual was convicted of murder 20 months after completion of the LGE programme. Those sanctioned were reasonably diverse in their offending, with 59% (n=321) of those sanctioned spanning 4 or more crime types. Overall, offending history is reasonably comparable to that of individuals on the Gangs Violence Matrix24, when taking into account the larger proportion of individuals with no offending records but below adult offending cohorts such as the London Criminal Justice Partnership’s Diamond reoffending pilot25 (20.7 previous sanctions) or Gripping the Offender26 (36.2 previous sanctions), as would be expected given these later programmes are aimed at older prolific offenders. 4.2 Proven Offending - before versus after Calculating Proven offending is the most robust measure of offending available to the evaluation (see methodology). When comparing proven offending in the 12 months before and after the LGE intervention start date, there are some small reductions in proportions to offend (i.e., 43.3% vs 40.5% comparing the 12 months prior to subsequent) – however, these differences are not significant in either proportion or the average number of offences per individual. Looking at 24 months before and after the intervention start date presents a similar picture in terms of comparing average offences per individual. However, a significant decrease in the proportion of violent offenders is apparent; 37.4% committed a violent offence in the 24 months prior to intervention start, compared to 28.7% in the 24 months after the start date. See table 6. 24 https://www.london.gov.uk/sites/default/files/gangs_matrix_review_-_final.pdf 25 https://www.london.gov.uk/sites/default/files/slp_reducing_reoffending_board_-_may_2011_-_info_item_- _diamond_year2_final_120411.pdf 26 https://www.london.gov.uk/sites/default/files/gripping_the_offender_interim_report_feb_2017.pdf 26 Table 6: Proven Offending Breaking the 12 months before and after period by completed and closed cases, the offending is more prevalent and more serious across closed cases. Whilst there are no significant decreases in proven offending for both completed or overall cases, serious offending shows a significant increase in closed (e.g., non-completed) cases, although the sample size is small. Table 7: Proven Offending - Completed or Closed 12mths Pre 12mths Post 24mths Pre 24mths Post All (PNC) Cohort 524 524 342 342 Incidents 684 682 929 913 Individuals 227 212 209 202 Proportion 43.3% 40.5% 61.1% 59.1% Cohort Avg 1.31 1.30 2.72 2.67 Violence (PNC) Incidents 149 170 201 204 Individuals 106 95 128 98 Proportion 20.2% 18.1% 37.4% 28.7%* Cohort Avg 0.28 0.32 0.59 0.60 Serious (PNC) Incidents 82 122 132 159 Individuals 45 53 68 66 Proportion 8.6% 10.1% 19.9% 19.3% Cohort Avg 0.16 0.23 0.39 0.46 Proven Offending Bold indicates significant difference: *McNemar **Wilcoxon Signed Rank 12mths Pre 12mths Post 12mths Pre 12mths Post 12mths Pre 12mths Post All (PNC) Cohort 192 192 332 332 524 524 Incidents 208 198 476 484 684 682 Individuals 74 66 153 146 227 212 Proportion 38.5% 34.4% 46.1% 44.0% 43.3% 40.5% Cohort Avg 1.08 1.03 1.43 1.46 1.31 1.30 Violence (PNC) Incidents 50 52 99 118 149 170 Individuals 30 28 76 67 106 95 Proportion 15.6% 14.6% 22.9% 20.2% 20.2% 18.1% Cohort Avg 0.26 0.27 0.30 0.36 0.28 0.32 Serious (PNC) Incidents 32 30 50 92 82 122 Individuals 17 12 28 41 45 53 Proportion 8.9% 6.3% 8.4% 12.3% 8.6% 10.1% Cohort Avg 0.17 0.16 0.15 0.28** 0.16 0.23 Bold indicates significant difference: *McNemar **Wilcoxon Signed Rank Complete Closed Overall 27 Analysis was also able to explore sub-groups (i.e., risk, needs, age, intervention length etc.) (see Appendix 4 for full breakdown) related to offending. Results found little significant differences between groups. Linked to the significant reduction in the proportion of individuals committing at least one violent offence in the 24 months after LGE start date outlined above, sub group analysis demonstrates a significant reduction in proven violent offending across the 24 months before and after period for individuals with above average needs was identified, which may point to potential longer-term benefits. Closed cases show a significant decrease in proportion committing violent offending; this is likely to be driven by a complex array of factors including individuals moving out of London or being incarcerated. A significant decrease in the proportion of LGE committing violent offences was also identified for those individuals aged under 20 at the point of referral, suggestive of the benefits of early intervention. Conversely, a significant increase in proven serious offending was identified for those individuals aged over 20 at the time of referral at 12 months before and after. While low base sizes mean in both cases results should be interpreted with caution, findings chime with Safer London practitioner experiences in that complex cases demand a longer intervention period. Secondly, earlier intervention in terms of age may be beneficial. 4.3 Digging Deeper: Impact of Violent Offending against a comparison group Before versus after, as outlined earlier, is one of the weakest methods in exploring impact and is not recommended in drawing conclusions. To be able to have more confidence, those receiving the treatment need to be compared to something – such as a similar group on the same measure. Focusing on violent offending where previous analysis identified a significant decrease in the proportion of offenders in the 24 months after the LGE intervention, the report now turns to this aspect through the use of the ‘shuffle’ approach developed by MOPAC/UCL and previously used in the MOPAC Review of the Gangs Violence Matrix (MOPAC 2018) to unpick this aspect more. The findings from this approach can be seen in Figure 5 – violent sanctions below, which presents the proportion of the entire eligible cohort sanctioned by month. The blue ‘observed’ line represents overall offending by month. The green line represents the estimated offending of the counterfactual, that is those individuals that are comparable but did not receive any intervention, with the band around it representing the upper and lower limits of normal variation. That is – to reach a significant difference the observed line has to exceed the green band to be significant. The monthly data enables a richer examination of offending behaviours. For example, as would be expected, spikes in the proportion of offending can be seen in the weeks prior to case opening. This is consistent with other interventions where increased offending is likely to be a factor in referral (e.g., Matrix Review) in bringing them to notice. Examining offending, month by month subsequent to the start, goes some way to illustrate our findings from earlier. One can see little difference within the first 12 months of the 28 observed group (i.e., LGE) with both peaks and dips, and the main reduction in offending appears to occur during the second year, with the lowest point reached at around the 18- month mark (before some final fluctuation). 4.3.2 A deeper focus: Violent Sanctions Figure 5: Violent Sanctions By Month 12 months before / after LGE - For the observed (LGE) group, an average of 2.02% of individuals per month received at least one sanction for a violent offence across the 12 months before entry to the LGE programme. In the 12 months after entry the figure was 1.89%. - For the shuffled (control) group, an average of 1.75% of individuals per month received at least one sanction for a violent offence across the 12 months before entry to the LGE programme. In the 12 months subsequent to LGE the figure was 1.71%. - These differences are not statistically significant. 24 months before / after LGE - For the observed (LGE) group, an average of 2.05% of individuals per month received at least one sanction for a violent offence across the 24 months before entry to the LGE programme. However, in the 24 months after entry the figure was 1.57%. - For the shuffled (control) group, an average of 1.76% of individuals per month received at least one sanction for a violent offence across the 24 months before entry to the LGE programme. In the 12 months subsequent to LGE the figure was 1.68%. - As is demonstrated on the above graph with the movement of the blue ‘observed’ line outside the light green band of normal variation, the decrease in violent offending is statistically significant when compared to the control. 29 4.4 Exploring an impact on Victimisation The report now turns to the impact upon victimisation of the LGE programme. Previous published reports suggested that the LGE intervention did have an impact on victimisation, although findings were limited to only including successfully completed cases. Whilst a useful indicator, the final report strengthens the approach and examines victimisation, with a larger sample, including all the eligible cohort regardless of completion, thus proving a more robust assessment. Exploring victimisation for 12 months pre and post intervention start date produces similar findings to previous reports, with statistically significant decreases across all, violent and serious number of incidents of victimisation). Table 8: Police Recorded Victimisation Comparing victimisation rates by year the individual was referred also tells a similar story - all years report at least one significant decline (in either proportion or cohort average). In terms of violent and serious victimisation, reductions are again observed across all years, but in these cases, it is years 1 and 2 that see significant reductions (violent victimisation significant drop in years 1 and 2, serious victimisation significant for year 1 only). 12mths Pre 12mths Post 12mths Pre 12mths Post 12mths Pre 12mths Post All Cohort 212 212 382 382 594 594 Incidents 165 91 270 193 435 284 Individuals 85 56 144 103 229 159 Proportion 40.1% 26.4%* 37.7% 26.9%* 38.6% 26.8%* Cohort Avg 0.78 0.43** 0.71 0.51** 0.73 0.48** Violence Incidents 103 49 144 87 247 136 Individuals 67 39 99 60 166 99 Proportion 31.6% 18.4%* 25.9% 15.7%* 27.9% 16.7%* Cohort Avg 0.49 0.23** 0.38 0.23** 0.42 0.23** Serious Incidents 45 18 75 43 120 61 Individuals 38 16 63 37 101 53 Proportion 17.9% 7.5%* 16.5% 9.7%* 17.0% 8.9%* Cohort Avg 0.21 0.08** 0.20 0.11** 0.20 0.10** Closed Overall Bold indicates significant difference: *McNemar **Wilcoxon Signed Rank Victimisation (CRIS) Complete 30 Table 9: Police Recorded Victimisation By Year As with offending, analysis explored differences between sub-groups for victimisation (see Appendix 4). Victimisation saw significant reductions across all sub-groups, and regardless of whether the case was completed or not. This suggests that the eligibility criteria and nature of the referral process may have some baring on these findings. 4.5 Other Assessments of Impact Outside of the offending and victimisation data, the evaluation sought to bring together wider measures of impact, such as caseworker assessed risk and mutually assessed ‘distance travelled’ against a range of intermediate outcomes (see Appendix). The report now turns to these different measures of success. 4.5.1 Case worker Assessed Risk Assigned caseworkers were required to assess the risk of those referred onto LGE at various timepoints within the intervention. Risk was considered across three separate areas; Risk To Others; Risk From Others and Risk To Self. This perceived risk was measured on a three-point scale from Low to High, and based on a range of available information from both client interaction and input from statutory agencies. Client risk assessments were revised across the duration of the programme. Assessments were down to the discretion of the caseworker and informed by client meetings, group supervision and discussion team meetings. Figure 5 demonstrates the movement in risk rating comparing assessment at the time of referral to assessment at competition of the programme. Risk ratings were captured for the majority of completed cases (range n=224-253; 85-96%). When comparing the risk at referral 12mths Pre 12mths Post 12mths Pre 12mths Post 12mths Pre 12mths Post 12mths Pre 12mths Post 12mths Pre 12mths Post All Cohort 185 185 141 141 172 172 96 96 594 594 Incidents 126 82 113 66 125 88 71 48 435 284 Individuals 69 49 61 38 62 44 37 28 229 159 Proportion 37.3% 26.5%* 43.3% 27.0%* 36.0% 25.6%* 38.5% 29.2% 38.6% 26.8%* Cohort Avg 0.68 0.44** 0.80 0.47** 0.73 0.51 0.74 0.50** 0.73 0.48** Violence Incidents 78 38 68 30 67 47 34 21 247 136 Individuals 53 23 47 26 44 32 22 18 166 99 Proportion 28.6% 12.4%* 33.3% 18.4%* 25.6% 18.6% 22.9% 18.8% 27.9% 16.7%* Cohort Avg 0.42 0.21** 0.48 0.21** 0.39 0.27 0.35 0.22 0.42 0.23** Serious Incidents 38 15 30 16 36 20 16 10 120 61 Individuals 34 13 24 14 29 18 14 8 101 53 Proportion 18.4% 7.0%* 17.0% 9.9% 16.9% 10.5% 14.6% 8.3% 17.0% 8.9%* Cohort Avg 0.21 0.08** 0.21 0.11 0.21 0.12 0.17 0.10 0.20 0.10** Victimisation (CRIS) Year One Bold indicates significant difference: *McNemar **Wilcoxon Signed Rank Year Two Year Three Year Four Overall 31 to risk at competition, statistical tests indicate a significantly lowered risk.27 These positive finding across all risk areas speaks to the intended outcomes of the programme, and tallies with victimisation impact findings above (‘risk from others’). Although valuable, the subjective nature of the reporting tool and absence of a control somewhat limit the confidence in the result. Figure 5: Practitioner Assessed Risk (Completed Cases Only) 4.5.2 Mutually Assessed Distance Travelled The LGE programme uses a Distance Travelled Assessment Tool to measure the progress made by individuals over the course of their intervention against set outcomes28. Both the client and support worker complete the assessment using a 5-point grading scale for each element at set points including the initial appointment, six weeks later once the support worker has developed a relationship with the young person, and again at 3 months and 6 months. For the evaluation, data was only available 27 A Wilcoxon Signed Rank Test indicated a statistically significant reduction in ‘Risk To Others’ when comparing referral assessment to final assessment (completed cases only), z= -9.381, p .000, with a large effect size (r = 0.5). The median score at ‘Start’ was 2 and the median score at ‘End’ was 1. A Wilcoxon Signed Rank Test indicated a statistically significant reduction in ‘Risk From Others’ when comparing referral assessment to final assessment (completed cases only), z= -9.593, p .000, with a large effect size (r = 0.5). The median score at ‘Start’ was 3 and the median score at ‘End’ was 2. 28 Outcomes assessed include: gang activity, harm from gangs, offending behaviour, housing, health & well-being, relationships, family and education, employment & training. See Appendix X While there was some difference between the aggregated individual level outcomes reported above and those outlined in Safer London’s Quarterly Performance Data, the <10 percentage point differences can be explained by issues with data extraction discussed earlier in this report; it was not possible to retrospectively obtain all individual distance travelled data. It should also be noted that the Family Factors outcome generated particular problems with data capture at the individual level and support across case management systems. 32 for the first and last distance travelled score, however positive progress was noted across all distance travelled outcomes. This is another valuable finding - particularly against key outcomes such as gang involvement and personal safety. There were however some issues with data quality (i.e., only 50% of Family dynamics had completed data) and so these results carry the same limitations as the Caseworker risk data above. Figure 6: Distance Travelled (Completed cases only) In assessing the impact of the LGE programme the research draws from a variety of data. Results demonstrate a non-significant reduction in overall offending but a significant reduction in the proportion of violent offenders in the 24 months after start date. Exploring victimisation for 12 months pre and post intervention start date reports statistically significant decreases across all, violent and serious number of incidents of victimisation. Staff and clients were positive in terms of assessed risk and distance travelled, with statistically significant progress identified against key ‘soft’ outcomes, albeit with caveats on data quality applied. 33 5. Discussion The current report presents final results from a long-term evaluation of the London Gang Exit programme. Over the years this has resulted in a suite of published reports exploring process learning, payment by results findings and interim findings on impact conducted by the MOPAC Evidence and Insight Team. This final report has had a specific focus upon impact using more robust methods and brings the evaluation journey to an end. A reminder of the need for the LGE programme Continuing concerns as to the prevalence and severity of serious violence involving children and young people - and the role that gangs play in this - has been a key issue for London over recent years. The 2017-21 Police and Crime Plan set out strategies to address the problem of gang related violence, offending and exploitation. Recent policy has emphasised prevention and intervention, highlighting the importance of joined-up, diversionary approaches in helping vulnerable young people out of serious offending and violent victimisation. It is in this context that London Gang Exit was devised. London Gang Exit (LGE) is a holistic support service to help individuals involved in gangs or at risk of gang violence or exploitation. Originally jointly commissioned by the Mayor’s Office for Policing And Crime (MOPAC) and the London Community Rehabilitation Company (CRC), the intervention went live in February 2016 and was initially funded until October 2017. MOPAC took over sole funding of the project after this date and LGE is currently funded to March 2022. LGE also received contributory funding from the Mayor’s Young Londoners Fund between 2018-21. The programme is being delivered by Safer London with key delivery partners Catch-22, Department of Work and Pensions (DWP)29 and the community CRC30. The pan London service is designed to complement and enhance existing local services, matching need to bespoke support across mental health, employment, family support, housing advocacy and specialist support for girls and women. The programme works with children and young people between the ages of 15-24 years on statutory orders or on a voluntary basis. The key overarching aims for LGE, measured through a combination of police recorded data and a practitioner/client ‘distance travelled’ assessment tool are: • Reduced / ceased harm from gangs; • Reduced / ceased involvement in gangs; and • Reduced seriousness and frequency of offending. 29 DWP contributed to programme funding 2016-17, providing a secondee from 2018 onwards 30 Withdrew as a partner in late 2020. 34 The basics of LGE delivery In the five years of delivery up until March 2021, LGE had received over 2200 enquiries around referrals and accepted nearly 1000 individuals onto the programme. The evaluation explored impact amongst those referred onto LGE in the first four years, from January 2016 until March 2020; 782 cases featuring 756 unique individuals. The LGE cohort is predominantly young, male and black. Whilst female representation increased in later years it remained always the clear minority. Around two thirds of clients were black; with the proportion remaining relatively constant. The youngest age at referral was 12, and the oldest 32. The median age for referral was 19. LGE programme data clearly demonstrates the multiple and crosscutting needs of the LGE cohort, spanning a range of contexts. Unsurprisingly, personal safety was identified as the most prevalent need. Over one fifth of the cohort highlighted 5 or more needs. Although gaps in the data does not allow for more granular analysis matching intervention to need it is clear these broadly align, it is also clear that the cohort is made up of a variety of (sometimes intersecting) groups; from those more deeply entrenched in gang related criminality, to those being exploited by gangs to those on the peripheries but ‘at risk’. Across the 782 accepted referrals, just over a third of cases (n=263) were successfully completed as of June 2021. 42% (n=215) were closed before being officially. Another third (n=293) were closed without completion during the intervention period. Whilst the completion rate from accepted referrals may appear low, it is not entirely unexpected considering the often-chaotic lifestyles and complex needs of the cohort. On average the intervention lasted just under 6 months. The impact of the LGE programme As outlined, the final report sought to have a specific focus examining the impact of the LGE programme triangulating across a range of data (i.e., victimisation, staff reports and proven offending) that broadly align to the aims of the service (i.e., reduced harm, reduced gang involvement and reduced seriousness and frequency of offending). In terms of offending, there were some small reductions (i.e., proportions to reoffend) but these did not reach significance. However, there was a significant decline in the proportion of violent offences conducted in the 24 months subsequent to the programme. Results also suggested those with above average needs and those aged under 20 had improved results on violent offending. The use of the ‘shuffle’ comparison group was able to pinpoint this decline at midway through the second year. These results raise more of an understanding of the nature and timescales of the impact and working with this group. In exploring victimisation for 12 months pre and post intervention start date reports statistically significant decreases across all, violent and serious number of incidents of victimisation. 35 As we saw the LGE group were young, presented both an offending background and multiple needs, and wider research has documented the link between age and offending and the wider maturation factors that can influence offending (Farrington 1992; Farrington et al 2014;). However, it is within this cohort that LGE reported a significant reduction in the proportion of violent offenders as well as a reduction in the proportion and rate of victimisation. This is a clear positive and should set a foundation for future work. Limitations As with all reports of this nature, it is important to note limitations. These stemmed primarily from data access and quality issues due to use of three different case management systems over the course of the programme. Whilst every effort was made to retrieve missing data by the Safer London team this was not always possible. In terms of analysis, impact analysis was limited in scope due to small sample sizes, meaning robust analysis exploring outcomes by specific sub-groups or over longer time periods was not possible. It should also be noted that we were unable to apply a control group methodology to victimisation analysis. Final thoughts This report marks the conclusion of an in-depth multiple year evaluation which has generated a suite of learning and insight. Importantly, it contributes to a relatively sparse UK evidence base in terms of what works in reducing gang related offending, as well as the wider evidence base on violence reduction. There is no doubt that substantial learning has been gleaned over entire period, with the suite of reports (and the ongoing programme development to generate improvements) highlighting the benefits of supporting largescale programmes with concurrent evaluations, a practice which should ideally become more frequent allowing a richer evidence base to evolve. 36 Appendix 1: Offending & Victimisation Sample Methodology All 756 individuals were submitted to a matching process against police crime reports to obtain victimisation data. Individuals were also matched to records from the Police National Computer (PNC) to enable analysis of proven offending.31 All individuals receiving all or part of the intervention on two separate occasions were excluded (n=23), as well as cases without intervention start dates (n=4). Across the police report and PNC cohorts, cases were excluded from analysis for a variety of reasons, including low matching confidence (n=5). 588 individuals were matched to PNC records and 661 were matched to police crime reports. Where individuals were not matched to crime report or PNC records, further analysis including triangulation between programme and police data was conducted to ascertain whether individuals should be included with ‘nil’ values (‘no trace’). For example, 30 non-PNC matched individuals were excluded from PNC analysis because police report data indicated they should have a valid PNC-ID. Similarly, 14 individuals were excluded from police report analysis because programme and/or PNC data indicated they were well known to police. Table 10: Police data cohorts for impact analysis Table 10 above breaks down the process. For PNC, this resulted in 649 individuals being in scope for impact analysis, including 588 matched PNC records and 63 individuals identified as ‘no trace’. The date of extraction for PNC data was 15/01/2021 with a cut of date for PNC analysis of 15/07/2020, allowing 6 months for information relating to be sanctions to be processed from PNC. Of the 649 in scope for inclusion in impact analysis, 81% (n=524) were eligible for 12 months pre/post impact analysis (i.e. they had at least 12 months of valid offending data either side of referral date) and 53% (n=342) eligible for 24 months pre/post. For victimisation, police crime reports data were obtained on all matched individuals (n=661) for the period 01/02/2015 to 10/10/2020 inclusive, with another 18 identified as valid ‘no traces’ on CRIS using the triangulated approach described above. This resulted in 679 individuals being in scope for victimisation impact analysis. Of these, 87% (n=594) were eligible for 12 months pre/post impact analysis and 36% (n=241) for 24 months pre/post. 31 All matching was undertaken by the Metropolitan Police Service using validated methodologies. Only matches with >95% confidence were included, and further QC/triangulation by E&I on receipt of data. Crime Reports (n =) PNC (n =) LGE Cohort 756 756 Matched 661 588 Included overall 679 649 Inc 12 months pre/post 594 524 Inc 24 months pre/post 241 342 37 Appendix 2: LGE Core Outcomes Of the young people starting on the programme involved in gangs there is a reduction in involvement or ceased involvement at the end of the programme Distance Travelled Assessment Safer London Victimization. Reduction in reports of victimization comparing 6 months before the programme and 6 months from the start of the intervention. Victim of gang related crime PNC Risk from others. Reduction in practitioner assessed harm from others between the start and end of the LGE intervention. Case worker assessed Risk Safer London Risk to Self. Reduction in practitioner assessed harm to self between the start and end of the LGE intervention. Case worker assessed Risk Safer London Outcome 2: Reducing / ceasing involvement in gangs Of the young people starting on the programme experiencing gang related harm, there is a reduction in harm at the end of the programme Distance Travelled Assessment Safer London Risk to Others. Reduction in practitioner assessed harm to self between the start and end of the LGE intervention. Case worker assessed Risk Safer London Frequency and volume of all offending. Comparing 12 months before the programme and 12 months after the start of the intervention with control group. Proven offending PNC Frequency and volume of violent offending. Comparing 12 months before the programme and 12 months after the start of the intervention with control group. Proven offending PNC Frequency and volume of serious offending. Comparing 12 months before the programme and 12 months after the start of the intervention with control group. Proven offending PNC Of those identified as needing housing and money management support, there is an increase in the number either accessing and/or increasing their ability to access housing by the end of the programme Distance Travelled Assessment Safer London Improving health and well-being: Of those with an emotional or physical health need, there is an increase in the number reporting an improvement by the end of the programme Distance Travelled Assessment Safer London Improved relationships: Of those needing support for improved relationships, there is an increase in the number reporting an improvement at the end of the programme Distance Travelled Assessment Safer London Improved family dynamics: Of those needing family support due to family conflict or risk, there are improved family dynamics and safety factors by the end of the programme Distance Travelled Assessment Safer London Improved work-related skills, knowledge and employment: Of the number requiring support, there is an increase in the numbers accessing and sustaining engagement in education, training and /or employment (ETE) Distance Travelled Assessment Safer London Data / Evidence Source Outcome 4: Improved access to pathways of support Outcome 3: Reduction in seriousness and frequency of offending Outcome 1: Reduction in harm caused by gang-related activity Outcome Measurement 38 Appendix 3: Performance Throughput Year 1*: FY2016/17 Year 2: FY2017/18 Year 3: FY2018/19 Year 4: FY2019/20 Year 5: FY2020/21 TOTAL Enquiries Number of enquiries that have come through to referral team. Could be general or referral specific (with a YP in mind). 586 271 345 533 494 2229 Referrals Accepted Number of enquiries that become referrals. Where YP meets the LGE referral criteria. 211 154 188 214 216 983 Disengaged Cases Those disengaged after being accepted as a referral but pre case opening or those that disengage during the intervention 69 121 41 4 5 240 Opened Cases Those that have attended their initial appointment and signed our agreement to commence work in this quarter. 103 116 116 160 208 703 Completed Cases (total to date since Feb 16) Those that have finished/completed the intervention that was planned. This will normally take on average 6 months. 21 74 51 70 54 270 No of completed cases that received core & three or more specialist intervention areas (strands) N/A 72% 47% 29% 37% Open & active cases Q4 Year 1 Q4 Year 2 Q4 Year 3 Q4 Year 4 Q4 Year 5 Open Cases [Snapshot - those being worked with in quarter] 103 66 108 141 202 620 Number of open cases engaged with both core & specialist intervention areas (strands) 77% 100% 77% 75% 59% % that are under 18 28% 32% 23% 20% 52% % that are 18 to 24 55% 62% 69% 67% 44% % that are 25 and over 17% 6% 7% 13% 4% % Open cases that are on a current statutory order (all) N/A 59% 23% 47% 22% Closed Cases Year 1* Year 2 Year 3 Year 4 Year 5 Total number of closed cases (closed after work has started) 28 89 39 114 110 380 Number of pre-initial closures(Closed before work started 12 47 27 5 15 106 Number of cases closed (not completed) after commencing work with the service (opening) 50 74 19 78 54 275 Outcomes [Distance Travelled, Completed Cases to Date] Q4 Year 1 Q4 Year 2 Q4 Year 3 Q4 Year 4 Q4 Year 5 Reducing / ceasing involvement in gangs 79% 85% 85% 84% 85% 85% Reduction in harm caused by gang related activity 80% 83% 84% 84% 84% 84% Offending behaviour 75% 74% 79% 78% 78% 78% Of those needing support for improved relationships, there is an increase in the number reporting an improvement at the end of the programme 77% 89% 85% 83% 82% 82% Increase in number accessing / increasing their ability to access housing N/A 80% 77% 79% 79% 79% Of those with an emotional or physical health need, there is an increase in the number reporting an improvement by the end of the programme N/A 79% 83% 83% 82% 82% Of those requiring support, there is an increase in the numbers accessing and sustaining engagement in education, training and /or employment N/A 80% 81% 74% 73% 73% Of those needing family support due to family conflict or risk, there are improved family dynamics and safety factors by the end of the programme N/A 72% 73% 69% 69% 69% LGE Performance Data: 1st February 2016 - 31st March 2021 *Year one data includes February and March 2016. ** It should be noted that figures presented here do not match exactly with those presented in the impact analysis due to differences in data collection/extraction between datasets. 39 Appendix 4: Offending & Victimisation across sub-groups Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Cohort 257 257 257 257 257 257 139 139 139 139 139 139 292 292 292 292 292 292 Incidents 313 334 68 102 38 62 357 339 89 91 47 61 223 155 128 81 58 37 Individuals 112 114 53 54 22 27 86 86 57 43 27 27 118 86 86 59 48 33 Proportion Off/Vic 43.6% 44.4% 20.6% 21.0% 8.6% 10.5% 61.9% 61.9% 41.0% 30.9% 19.4% 19.4% 40.4% 29.5%* 29.5% 20.2%* 16.4% 11.3% Mean 1.22 1.30 0.26 0.40 0.15 0.24 2.57 2.44 0.64 0.65 0.34 0.44 0.76 0.53** 0.44 0.28** 0.20 0.13** Cohort 248 248 248 248 248 248 192 192 192 192 192 192 282 282 282 282 282 282 Incidents 342 331 73 63 41 59 534 542 105 104 80 95 199 122 112 54 58 23 Individuals 105 89 47 37 22 25 115 107 66 49 39 37 106 69 76 39 50 19 Proportion Off/Vic 42.3% 35.9% 19.0% 14.9% 8.9% 10.1% 59.9% 55.7% 34.4% 25.5% 20.3% 19.3% 37.6% 24.5%* 27.0% 13.8%* 17.7% 6.7%* Mean 1.38 1.33 0.29 0.25 0.17 0.24 2.78 2.82 0.55 0.54 0.42 0.49 0.71 0.43** 0.40 0.19** 0.21 0.08** Cohort 190 190 190 190 190 190 128 128 128 128 128 128 222 222 222 222 222 222 Incidents 307 267 61 68 30 35 450 340 89 69 53 52 148 96 77 50 38 18 Individuals 88 84 41 36 17 16 83 80 55 34 27 24 88 61 57 35 33 17 Proportion Off/Vic 46.3% 44.2% 21.6% 18.9% 8.9% 8.4% 64.8% 62.5% 43.0% 26.6%* 21.1% 18.8% 39.6% 27.5%* 25.7% 15.8%* 14.9% 7.7%* Mean 1.62 1.41 0.32 0.36 0.16 0.18 3.52 2.66 0.70 0.54** 0.41 0.41 0.67 0.43** 0.35 0.23** 0.17 0.08** Cohort 326 326 326 326 326 326 207 207 207 207 207 207 364 364 364 364 364 364 Incidents 367 410 86 100 49 86 456 554 109 128 74 104 284 186 169 85 81 43 Individuals 136 125 64 58 27 36 121 117 71 60 39 40 139 96 108 63 67 36 Proportion Off/Vic 41.7% 38.3% 19.6% 17.8% 8.3% 11.0% 58.5% 56.5% 34.3% 29.0% 18.8% 19.3% 38.2% 26.4%* 29.7% 17.3%* 18.4% 9.9%* Mean 1.13 1.26 0.26 0.31 0.15 0.26 2.20 2.68 0.53 0.62 0.36 0.50 0.78 0.51** 0.46 0.23** 0.22 0.12** Cohort 284 284 284 284 284 284 188 188 188 188 188 188 325 325 325 325 325 325 Incidents 378 395 102 115 64 73 529 542 136 139 84 91 290 183 167 93 72 45 Individuals 121 115 69 62 32 32 116 112 81 61 43 35 143 97 106 64 58 40 Proportion Off/Vic 42.6% 40.5% 24.3% 21.8% 11.3% 11.3% 61.7% 59.6% 43.1% 32.4% 22.9% 18.6% 44.0% 29.8%* 32.6% 19.7% 17.8% 12.3% Mean 1.33 1.39 0.36 0.40 0.23 0.26 2.81 2.88 0.72 0.74 0.45 0.48 0.89 0.56** 0.51 0.29** 0.22 0.14** Cohort 240 240 240 240 240 240 154 154 154 154 154 154 269 269 269 269 269 269 Incidents 306 287 47 55 18 49 400 371 65 65 48 68 145 101 80 43 48 16 Individuals 106 97 37 33 13 21 93 90 47 37 25 31 86 62 60 35 43 13 Proportion Off/Vic 44.2% 40.4% 15.4% 13.8% 5.4% 8.8% 60.4% 58.4% 30.5% 24.0% 16.2% 20.1% 32.0% 23%* 22.3% 13%* 16.0% 4.8%* Mean 1.28 1.20 0.20 0.23 0.08 0.20** 2.60 2.41 0.42 0.42 0.31 0.44 0.54 0.38** 0.30 0.16** 0.18 0.06** Cohort 222 222 222 222 222 222 143 143 143 143 143 143 247 247 247 247 247 247 Incidents 279 284 57 78 32 43 424 399 85 92 58 73 203 122 120 66 50 25 Individuals 95 89 38 41 22 20 92 88 54 45 32 29 104 76 76 50 43 23 Proportion Off/Vic 42.8% 40.1% 17.1% 18.5% 9.9% 9.0% 64.3% 61.5% 37.8% 31.5% 22.4% 20.3% 42.1% 30.8%* 30.8% 20.2%* 17.4% 9.3%* Mean 1.26 1.28 0.26 0.35 0.14 0.19 2.97 2.79 0.59 0.64 0.41 0.51 0.82 0.49** 0.49 0.27** 0.20 0.10** Cohort 298 298 298 298 298 298 195 195 195 195 195 195 343 343 343 343 343 343 Incidents 393 388 89 91 48 79 489 492 110 109 69 84 226 162 121 70 68 36 Individuals 129 120 66 53 22 33 114 111 72 51 34 36 124 83 89 49 57 30 Proportion Off/Vic 43.3% 40.3% 22.1% 17.8% 7.4% 11.1% 58.5% 56.9% 36.9% 26.2%* 17.4% 18.5% 36.2% 24.2%* 25.9% 14.3%* 16.6% 8.7%* Mean 1.32 1.30 0.30 0.31 0.16 0.27 2.51 2.52 0.56 0.56 0.35 0.43 0.66 0.47** 0.35 0.20** 0.20 0.10** Cohort 196 196 196 196 196 196 143 143 143 143 143 143 214 214 214 214 214 214 Incidents 373 326 75 92 44 66 518 491 117 134 73 94 131 84 87 37 52 18 Individuals 116 103 55 46 24 30 110 106 70 56 37 42 87 56 66 31 47 17 Proportion Off/Vic 59.2% 52.6% 28.1% 23.5% 12.2% 15.3% 76.9% 74.1% 49.0% 39.2% 25.9% 29.4% 40.7% 26.2%* 30.8% 14.5%* 22.0% 7.9%* Mean 1.90 1.66 0.38 0.47 0.22 0.34 3.62 3.43 0.82 0.94 0.51 0.66 0.61 0.39** 0.41 0.17** 0.24 0.08** Cohort 328 328 328 328 328 328 199 199 199 199 199 199 361 361 361 361 361 361 Incidents 311 356 74 78 38 56 411 422 84 70 59 65 285 173 148 86 63 37 Individuals 111 109 51 49 21 23 99 96 58 42 31 24 134 95 95 63 50 33 Proportion Off/Vic 33.8% 33.2% 15.5% 14.9% 6.4% 7.0% 49.7% 48.2% 29.1% 21.1% 15.6% 12.1% 37.1% 26.3%* 26.3% 17.5%* 13.9% 9.1% Mean 0.95 1.09 0.23 0.24 0.12 0.17 2.07 2.12 0.42 0.35 0.30 0.33 0.79 0.48** 0.41 0.24** 0.17 0.10** Cohort 192 192 192 192 192 192 127 127 127 127 127 127 212 212 212 212 212 212 Incidents 208 198 50 52 32 30 298 282 73 75 53 50 165 91 103 49 45 18 Individuals 74 66 30 28 17 12 76 69 41 33 23 20 85 56 67 39 38 16 Proportion Off/Vic 38.5% 34.4% 15.6% 14.6% 8.9% 6.3% 59.8% 54.3% 32.3% 26.0% 18.1% 15.7% 40.1% 26.4%* 31.6% 18.4%* 17.9% 7.5%* Mean 1.08 1.03 0.26 0.27 0.17 0.16 2.35 2.22 0.57 0.59 0.42 0.39 0.78 0.43** 0.49 0.23** 0.21 0.08** Cohort 332 332 332 332 332 332 215 215 215 215 215 215 382 382 382 382 382 382 Incidents 476 484 99 118 50 92 631 631 128 129 79 109 270 193 144 87 75 43 Individuals 153 146 76 67 28 41 133 133 87 65 45 46 144 103 99 60 63 37 Proportion Off/Vic 46.1% 44.0% 22.9% 20.2% 8.4% 12.3% 61.9% 61.9% 40.5% 30.2%* 20.9% 21.4% 37.7% 27.0%* 25.9% 15.7%* 16.5% 9.7%* Mean 1.43 1.46 0.30 0.36 0.15 0.28** 2.93 2.93 0.60 0.60 0.37 0.51 0.71 0.51** 0.38 0.23** 0.20 0.11** Bold denotes signifcant difference between pre and post periods. * Wilcoxon Signed Rank Test was used for frequency of offending, and ** McNemar Test used for proportion offending. Not on Matrix Completed Cases Closed Cases Below Avg Needs Intervention > 6mths Intervention < 6mths On Matrix Under 20 Over 20 PNC 12mths ALL PNC 12Mths VIOLENT High Risk Low Risk Above Avg Needs Victim VIOLENT 12mths Vicitim SERIOUS 12mths PNC 12mths SERIOUS PNC 24Mths ALL PNC 24Mths VIOLENT PNC 24mths SERIOUS Victim 12mths ALL 40 Appendix 5: Eligibility Criteria LGE takes self referrals, referrals from families/carers/friends, statutory/non statutory organisations. We encourage you to call to discuss any potential referral before completing a referral form so that we can be sure young people are eligible for the service. If LGE is not the right service we will do our best to sign post to an alternative provision. To be eligible for referral, the young person (male or female) should be:  aged between 16-24 associated with or involved in gangs (if they fall outside the eligible age range they will be considered on a case by case basis)  at significant risk of harm from gang activity, (such as violence or exploitation), a risk to themselves, or pose a risk of harm to others motivated to end their gang involvement (if applicable) and are willing to work with the LGE service. The young person will not be eligible:  if they are not yet motivated to end their gang involvement (if applicable) and not willing to work with the LGE service  if there is no evidence of gang association or risks from gangs  if they are already receiving extensive support from the borough that they are residing in, or from other statutory organisations, or if the services they require are available locally to them (although if services are oversubscribed or not suitable then they will be considered). 41 References Dawson, Stanko, Higgins and Rehman (2011). An evaluation of the Diamond Initiative: year two findings. London: Metropolitan Police Service; London Criminal Justice Partnership. Farrington, D. (1992) Criminal Career Research in the United Kingdom. British Journal of Criminology. 32(4). 521-536. Farrington, D., Ttofi, M., Crago, R. & Coid, J. (2014) Prevalence, frequency, onset, desistance and criminal career duration in self-reports compared with official records. Criminal Behaviour and Mental Health. 24: 241–253. Lammy, D (2017) The Lammy Review: An Independent Review into the Treatment of, and Outcomes for, Black, Asian and Minority Ethnic Individuals in the Criminal Justice System. London: Lammy Review. MOPAC (2018) Review of the Metropolitan Police Service Gangs Matrix. London: Mayor’s Office for Policing And Crime. Available from: https://www.london.gov.uk/sites/default/files/gangs_matrix_review_-_final.pdf MOPAC (2021) Review of the Metropolitan Police Service Gangs Violence Matrix – A one- year update. London: Mayor’s Office for Policing And Crime. Available from: https://www.london.gov.uk/sites/default/files/gvm_update_-_ei_review_- _jan_2021_final_for_publication.pdf 1 The London Rape Review 2021 An examination of cases from 2017 to 2019 with a focus on victim technology Daniela Wunsch, Dr Tom Davies and Barry Charleton MOPAC Evidence & Insight December 2021 2 Acknowledgements MOPAC would like to express their sincere gratitude to Rachael Parker, Emanus Prospere, Michael Keenan, Kirstie McLaren and Jose Gascon Simorte for coding all the cases in the sample. Many thanks also to Dr Paul Dawson for his valuable advice throughout the research, as well as Dr Julia Yesberg for her insights around the predictive analysis. MOPAC would also like to thank the Victims’ Commissioner for London, Claire Waxman and DCS Helen Lyons, the Met’s senior responsible officer for rape, for their support and guidance during this review. Finally, many thanks to the survivors and the police officers who gave their time to share their views and perspectives and provided feedback to the research team. 3 Contents Executive Summary .............................................................................................................. 5 1. Introduction ....................................................................................................................... 8 1.1 Background ................................................................................................................. 8 1.2 The current review ....................................................................................................... 9 2. Methodology ................................................................................................................... 10 2.1 Case coding .............................................................................................................. 10 2.2 Survivor and officer feedback .................................................................................... 12 3. Results ............................................................................................................................ 14 3.1 Attrition of rape allegations ........................................................................................ 14 3.2 The profile of reported rape in London ....................................................................... 16 3.3 Technology Evidence ................................................................................................ 20 3.4 Predictors of attrition .................................................................................................. 22 3.5 Predictors of police requests for victim technology .................................................... 26 3.6 The perspectives and experiences of survivors and officers ...................................... 27 4. Discussion ...................................................................................................................... 34 4.1 Summary of key findings ........................................................................................... 34 4.2 Future Research ........................................................................................................ 37 4.3 Conclusion ................................................................................................................. 38 Appendices ......................................................................................................................... 39 References ......................................................................................................................... 50 4 At a Glance.… The London Rape Review 2021 Victim Technology Evidence 23% of cases involved the police requesting a technology device from the victim. There were no significant differences between the three time periods studied. A 50/50 split was found for victims’ consent to the request. The current review found no evidence of police blanket requests for victim technology. Whilst technology variables had some influence on case outcome - increasing the likelihood of CPS submission - this was not as strong as other - procedural - variables. Markers of case progression, such as completion of a Video Recorded Interview, request for third party material, or seizure of suspect technology, significantly increased the likelihood of a police victim technology request. Victim/survivor feedback suggests frequent concerns about technology requests, relating to the practicalities of not having access, but also perceived intrusiveness and concerns of material being used to discredit them. Officers tended to view technology evidence as an opportunity but raised concerns around the practical implication of processing technology material in the wider context of high caseloads. An examination of 450 rape allegations Random sample; taken from three time periods: 1) October-December 2017; 2) January-March 2018; and 3) October-December 2019. Coded across 140 variables pertaining to characteristics of victim/survivors, suspects, offence circumstances, case progression and outcomes, as well as technology evidence. This is a follow up to the MOPAC London Rape Review (2019). Using the same quantitative case coding approach, the aim was to gain an updated picture of the characteristics of reported rape in London, with an in- depth focus on examining requests for victim technology. The picture of reported rape in London has remained largely unchanged since the 2019 Rape Review Next steps / recommendations o Review support provisions for victims and alternatives to criminal justice outcomes – balanced with maintaining public protection as key aim. o Improvements to the approach taken to engage with victims including understanding the needs of high harm crime victims. o Improvements needed to the practical and emotional support officers receive: -Guidance on how to systematically approach technology evidence. -Training on victim needs and close perception gaps. - Consider reflective practice /clinical supervision. o Future research to include a victim survey of High Harm offences, and to include information drawn from partner agencies including the CPS, ISVAs and support agencies.  65% of cases end in victim withdrawal.  A further 25% end in police ‘No Further Action’.  Only 6% of cases are committed by a complete stranger.  35% are committed by a current/former partner, 29% by an acquaintance/friend.  Most offences take place in a private setting (59%).  The strongest predictors of victim withdrawal and Police ‘No Further Action’ were procedural characteristics. 5 MOPAC Rape Review 2021 Executive Summary About the MOPAC Rape Review 2021 The 2021 London Rape Review is a follow up to the MOPAC London Rape Review (2019). Using the same case coding methodology, it gives an updated picture of the profile of reported rape in London. Additionally, the current research provides evidence around a much-debated topic: that of technology evidence and the use of victim technology in rape cases. The current research analysed 450 allegations of rape - a random sample of cases reported to the Metropolitan Police Service (MPS) during one of three time periods: 1) October to December 2017; 2) January to March 2018; and 3) October to December 2019. These periods were chosen to allow comparisons over time, particularly in relation to the National Disclosure Improvement Plan1. In addition, feedback from survivors and police officers was captured to provide further context. The current landscape of reported rape in London - Comparison with previous reviews Consistent with previous reviews, the 2021 London Rape Review finds steep attrition of cases. Two thirds of cases ended in victim withdrawal and a further one quarter ended in a police ‘no further action’ (NFA) decision. Fewer than 1 in 10 cases had been submitted to the CPS at the time of writing, although, given the recent sample, there were still a number of cases awaiting a final outcome. Similarly, victim/survivor, suspect and offence profiles have remained broadly consistent. Those who report rape in London continue to be pre-dominantly female and young, with almost one third under the age of 18 at the time of the offence. Suspects continue to be men with a slightly older age profile compared to victim/survivors. When compared to their overall proportions in the London population, there continues to be an over-representation of both Black victim/survivors and Black suspects. Finally, and again consistent with previous research, in the majority of cases victim/survivor and perpetrator are known to one another and there is considerable overlap with Domestic Abuse. Victim withdrawal has consistently remained the largest attrition point, and the current research finds evidence that withdrawals frequently take place soon after reporting and are often linked to victims not or no longer wishing to pursue a criminal justice outcome. Predictors of Attrition The modelling replicated previous analytics to explore which factors predicted victim withdrawal as well as police NFA. Similar to the 2019 Rape Review it is procedural characteristics that contribute most to these outcomes, compared to victim, suspect or offence characteristics. Of note is particularly the role of Video Recorded Interview (VRI) completion in making victim withdrawal significantly less likely, suggesting that its function as a barrier has continued. Variables relating to an assessment of the victim’s account and victim credibility also impact on outcome, with inconsistencies in the victim’s account making police NFA 12 times more likely, whilst markers of case progression (request for third party material, suspect arrest, Early Investigative Advice) make police NFA less likely. 6 The role of victim technology evidence A police request to take a technology device from the victim occurred in just under one quarter (23%; n=102) of all cases. There were no statistically significant differences in this across the three time periods examined. In all but 13 cases, the request followed an explicit reference to the possible existence of technology evidence. As such, the current review does not find evidence of police blanket requests for victim technology, nor of a spike in such requests after R v Allan. Following a request, half of victims agreed to handover their device, whilst half declined. The proportion of victim consent drops slightly in the time immediately following R v Allan. Whilst this could be related to publicity around the case and media attention, it should be noted that the change was not statistically significant. The 50 / 50 split in terms of victim consent to the request indicates that more work needs to be undertaken in the engagement of victims with respect to technology evidence. Reference to the existence of technology evidence, as well as request for and obtaining of victim technology were all found to be linked to a significantly increased likelihood of CPS submission. However, when included in the overall predictive models for victim withdrawal / Police NFA - they did not remain significant, suggesting that technology variables have some influence on case outcome, but not as strongly as other - procedural - variables. Procedural characteristics were also strongest at predicting the likelihood of a victim technology request - mainly those that can be considered markers of case progression: VRI completion, request for third party material, accessing other digital material and seizure of suspect technology all significantly increased the likelihood of a police victim technology request. Survivor and officer voices The current review incorporated feedback from survivors and police officers. This found that whilst officers tended to talk about technology as an opportunity to find corroborating evidence, victim/survivor concerns about technology requests were frequent and related to the practicalities of not having access to their phones, but also perceived intrusiveness. Survivors frequently expressed feeling that the overall objective was to check their account of events - making them feel like a suspect. Whilst officers described a considered and reassuring approach, this view was not shared by all the survivors who responded to the survey. The small sample sizes are a limitation, but nevertheless point to the importance of effective victim engagement, with expectation management, fair treatment, reassurance, regular information provision and empathy as critical factors. Current routine measures that capture feedback from victims of crime and their satisfaction with the police do not include victims of high harm crimes, such as rape, and there is therefore a gap in capturing and learning from the victims of these offences to generate insights and shape practice. Officers talked about carrying high caseloads and the challenges of progressing an investigation whilst having to wait on others - with delays putting pressure on their relationships with victims. There is a need to look at what improvements can be made to the practical and emotional support officers currently receive. In relation to technology evidence this may need to take the form of more guidance (via training and / or toolkits) on how to systematically approach phone searches. It also points to the need for further input and training to help officers better meet victim needs and close perception gaps, but also additional support measures, such as reflective practice or clinical supervision. Finally, there was a sense that whilst police requests for victim technology were more likely to follow an explicit reference to the possible existence of such evidence, the Crown Prosecution Service (CPS) were seen to be frequently asking for more speculative phone searches. Indeed, officers felt that, whilst police focused on reasons why a case should go to court, the CPS predominantly focused on the 7 reasons for why it should not. Some of this sentiment is reflected in the survivor feedback: most had not seen their cases progress to trial and there was a strong sense that cases were dropped despite strong evidence. The current research is not able to add to this the perspective of the CPS and those working within it - something that is needed to be able to give a more comprehensive picture. The findings certainly point to the importance of effective partnership working and mutual understanding between these two agencies. Further research The current review identifies three key areas for further research:  Firstly, to gain a more complete picture, there is a clear need to track cases beyond police systems and incorporate data as well as feedback from partners, including and particularly the CPS. This should include further detail in relation to outcomes, what constitutes ‘good’ case file quality and Police-CPS correspondence and decision-making - including that in relation to victim and suspect technology evidence, what material is disclosed, and how disclosed material is being used in court / during trial.  Secondly, in addition to an over-representation of both Black victims and suspects - a finding that has been consistent across different reviews - the current review identified a relationship between BAME suspects and a significantly increased likelihood of victim technology requests. The complexities of these relationships demand exploration beyond the scope of this study but sits comfortably within the framework of the Mayoral Action Plan into transparency, accountability and trust in policing, published November 20191, specifically established to address issues of disproportionality and the use of police powers for Black Londoners. Future research and reviews on this nature should continue to monitor and understand these disproportionalities.  Finally, given the consistent finding that the majority of victim/survivors who come to the police to report their rape subsequently withdraw - often within days of the report - there is scope for further research to explore withdrawal reasons in more depth, possibly alongside research to better understand decision-making to report and victim needs and expectations at this point in order to inform support provisions and explore alternatives to criminal justice outcomes. Conclusion The current review finds much unchanged in terms of the profile of reported rape and the picture of attrition in London, as well as the key factors affecting case outcomes. Additionally, it is able to contribute evidence-based insights to the debate around victim technology evidence and disclosure requests. It is hoped the findings from this review can encourage further discussions about how experiences and outcomes for victims/survivors can be improved and how the different criminal justice agencies and wider partners can come together to facilitate such improvements. 1 https://www.london.gov.uk/sites/default/files/action_plan_- _transparency_accountability_and_trust_in_policing.pdf 8 1. Introduction This report is a follow up to the MOPAC London Rape Review (2019). Using the same innovative quantitative case coding approach, the aim of this review is to gain an updated picture of the characteristics of reported rape in London, with an in-depth focus on examining the prevalence of requests for victim technology, whether there have been any changes in requests over recent years and the impact this can have on investigation, case progression and outcome, but also on victim-survivors. In addition, the report incorporates the perspectives of survivors who decided to report to the Metropolitan Police Service (MPS), and MPS officers working within rape investigation. 1.1 Background 1.1.1 The MOPAC London Rape Review 2019 The MOPAC London Rape Review (2019)2 examined all rape allegations made to the MPS during April 2016 with view to better understanding the profile of victims, perpetrators and case characteristics, as well as attrition and case outcomes. The 2019 review updated the evidence base towards this crime. Of the 501 allegations examined, in 58% of cases the victim/survivor withdrew the allegation, in a further 29% the police decided to take no further action, only 60 were submitted to the CPS, 36 were charged, 23 proceeded to trial and 14 ended in either a guilty plea or verdict - an overall conviction rate in the sample of 3%. The review was able to offer a range of insights in relation to the demographics and needs of the victim-survivors who report rape in London, the profile of suspects, timeliness and evidential challenges, as well as key attrition points and predictors of case outcomes. In doing so, it also challenged some often-held misconceptions around rape, evidencing that in the majority of cases, victim and offender were known to each other, with almost three in five offences taking place in a private or domestic setting and 28 per cent of all allegations relating to domestic abuse. Whilst the 2019 MOPAC London Rape Review provided the most comprehensive picture of reported rape in London at that time, it was not able to collate sufficient and timely data in relation to the use of technology evidence in cases and its impact on victims and investigations. 1.1.2 The use of technology evidence in cases of rape In 2017 two separate reviews – a joint review by Her Majesty’s Inspectorate for Crime Fire and Rescue Services (HMICFRS) and Her Majesty’s Crown Prosecution Service Inspectorate (HMCPSI); and the Mouncher Investigation Report by Richard Horwell QC - identified several shortcomings in police disclosure processes. Both reports cited the ‘explosion’ of personal communication and social media technology as a new source of potentially relevant material which was not necessarily being thoroughly investigated, nor unused material being 2 https://www.london.gov.uk/sites/default/files/london_rape_review_final_report_31.7.19.pdf 9 disclosed. The combined reports resulted in 26 recommendations for Police and CPS to take on, published in January 2018 as a National Disclosure Improvement Plan3. Just before the publication of this Improvement Plan, the issue hit the media when in December 2017, the rape case brought against Liam Allan collapsed in court due to disclosure issues. Text messages exchanged between the complainant and Allan fell under the category of unused relevant material, which had not been disclosed to the defence and, the defence argued, would undermine the prosecution case. The case received considerable media and political attention, publicly highlighted a failure in disclosure processes and brought to the forefront discussions around the prevalence, usefulness and ethics of using victim technology and social media evidence in rape cases. Whilst some argued that technology enabled new opportunities for evidence (e.g., Angiolini, 2015), victim groups in particular raised concerns over the practical implications and risks associated with the seizure and use of such evidence, arguing that disclosure requests were often disproportionate, providing another pathway for victim scrutiny. Following the 2017 reviews, the Liam Allan case and the National Disclosure Improvement Plan there was a great deal of focus on police disclosure procedures, particularly in rape investigations. This included concerns of a possible knee-jerk reaction on behalf of the police from early 2018 onwards, with an increase in automatic blanket requests for rape victims’ mobile phones. In summary, the issue has received considerable attention and has been much-debated; however, to date, no research has been conducted to contribute empirical evidence. 1.2 The current review The key objectives of the current research are to:  Provide an updated picture of reported rape in London - including key attrition points and characteristics of victim/survivors, perpetrators and offences, as well as subsequent actions and responses;  Make observations about any changes since 2016 - the time period examined in the original MOPAC London Rape Review (2019);  Examine the frequency with which requests for technology evidence are made and the extent to which this may or may not differ over time (across three specific time periods in 2017, 2018 and 2019);  Explore the impact of requests for technology evidence have on investigations, case progression and outcome, and on victim/survivors and officers; and  Identify the key factors influencing attrition, as well as those influencing requests for victim technology. 3 https://www.cps.gov.uk/publication/national-disclosure-improvement-plan 10 2. Methodology The current review largely employed a repeat of the methodology used for the 2019 Rape Review. In addition to the case coding, an online survey with survivors and a series of semi- structured interviews with police officers were conducted. 2.1 Case coding Case coding broadly followed the same methodology as that used in the MOPAC London Rape Review (2019): a sample of allegations of rape made to the MPS were taken from the MPS Crime recording and Information System (CRIS) and coded for key characteristics in relation to victims, suspects / accused, offences, case progression and outcome. Key methodological changes in the 2021 report will be outlined in the sections below and pertain to:  The use of a random sample of cases taken from within three distinct time periods (rather than using a full sample of all allegations made during the month of April, as was done for the 2019 review); to explore any differences across times; and  The addition of a suite of coding variables specifically pertaining to technology evidence. 2.1.1 Sample All crime records with an initial classification of rape reported to the MPS during three distinct time periods were extracted from CRIS.4 The time periods were: 1. October 2017 and December 2017 (i.e., prior to the national disclosure improvement plan); 2. January 2018 and March 2018 (i.e., following the national disclosure improvement plan); and 3. October 2019 – December 2019 (i.e., more recent). A random sample of 1505 cases was drawn from each of these time periods, resulting in three subsamples and a total sample of 450 cases. Cases were excluded and replaced if: the initial classification was attempted rape; the classification later changed to a non-sexual violence offence; the case was transferred out of the MPS to another police force; or the record was identified as a duplicate of another. Cases that had been no crimed or given a classification of ‘rape not confirmed’ or ‘rape contradictory evidence’ were also excluded. Whilst this means that direct comparisons between the current and the 2019 review are not always possible, it increased the likelihood of the final sample 4 Crime records were extracted using Full Business Objects Client. 5 8% of 1,867 offences during period one (where N100 current classifications and rape transferred current classifications had been removed from initial CRIS download); 9% of 1,660 offences during period 2 (same criteria); 7.5% of 1,999 offences during period 3 (same criteria) 11 containing a sufficient number of cases that have progressed enough to ascertain whether technology evidence was an issue or not. 2.1.2 Case coding and data analysis A team of researchers from MOPAC Evidence & Insight and MPS staff read each individual CRIS report for the 450 sampled cases and coded them on a total of 140 variables in the coding framework (see Appendix A for an overview)6. The coding framework was divided into six categories: Victim/survivor characteristics; Suspect characteristics; Offence circumstances; Procedural characteristics (the police response and investigation); Use of technology evidence; and Outcomes.7 All researchers received training on the coding framework and the coding procedures to be followed. Throughout the coding phase of the research, regular meetings between the researchers took place to discuss and address any concerns or uncertainties with cases and / or variables. Spot checks on randomly selected cases were conducted whereby researchers double-coded each other’s cases to determine consistency in scoring. The resultant dataset was analysed using descriptive statistics to explore relationships between case characteristics and enable comparisons to the previous review8. Additional in- depth analysis was conducted on the ‘Use of technology evidence’ variables. Replicating the analytical approach used in the London Rape Review (2019), a series of logistic regressions were conducted to explore whether case characteristics predicted case outcomes, as well as requests for victim technology. Each category of case characteristic was examined separately (i.e., victim/survivor, suspect, offence, and procedural (including technology) characteristics) before being combined into an overall model. The two outcomes of interest were victim withdrawal and police no further action, capturing both victim and police decision making. Univariate logistic regressions were first conducted on each case characteristic and the two outcomes.9 Significant univariate predictors were retained for inclusion in a multi-variate model for each of the four categories of case characteristics.10 The characteristics that remained significant in each multi-variate model were then retained for inclusion in the final overall model. The same approach was taken to conduct the analysis of predictors of police technology requests. All analyses were conducted in SPSS version 24. 6 All researchers were security vetted and followed strict data management and protection protocols. 7 Most of these variables were coded dichotomously (1=yes, 0=no), whereas others were categorical (e.g. ethnicity, offence location, relationship between victim and suspect). 8 Descriptive statistics were calculated on the full sample of 450 cases. Predictive analysis excluded ongoing cases (n=18) and believed false allegations (n=15). For Police No Further Action (NFA), victim withdrawals were also excluded (n=288). 9 Univariate logistic regression explores the relationship between one independent variable (i.e., one case characteristic) and one dependent variable (i.e., one outcome). 10 Multi-variate logistic regression explores the relationship between two or more independent variables and one dependent variable. 12 2.1.3 Limitations of the methodology The research has a number of limitations. As with the London Rape Review (2019), case information as coded solely from CRIS meaning there were areas of the journey from offence to trial outcome that were either not captured or were recorded in an inconsistent manner. Although the current coding framework captured a significant number of variables, data recorded and held by the CPS, the courts, and victim support services may provide a more complete picture than data available exclusively on CRIS. With regards to the CRIS data, the system was designed as an investigation record, not a comprehensive data capture tool. Therefore, the quality and detail of the data itself is variable. Although there are specific fields available on the database for many of the variables coded as part of this research, completing them is not always mandatory. Furthermore, most of the information coded is derived from the Details of Investigation: a long, free text log, completed by any number of different police officers and staff throughout the duration of the investigation. There may be details or pieces of information that are of interest which are not recorded during the completion of this log, and it is not possible to verify the accuracy and comprehensiveness of the information recorded within. Furthermore, many of the variables were coded only for their presence which means we cannot determine whether the absence of a variable is due to omission in data recording or it not being a factor in the case. As such this limits the consistency and reliability of what was coded. A further limitation, which is shared with all research that uses police or official data, is that the cases coded and analysed in this research are only those that were reported to the police. Given the high rate of under-reporting of sexual violence, these cases represent a small proportion of the rapes that may have actually taken place. Finally, random sampling was selected to identify the cases under study – simply, this means that every eligible case had an equal chance of being selected for the research. This method of sampling seeks to generate an unbiased representation of the whole population. Whilst, the sample size was sufficient to enable meaningful analysis and is reflective of what was feasible in the time and with the resources available (in terms of the in-depth case coding), the sample represents 8% of total allegations overall - and therefore caution should be applied with regards to any generalisations. 2.2 Survivor and officer feedback In order to explore the experiences and perspectives of survivors who reported the rape they had experienced to the police, an online survey was disseminated via the Victims’ Commissioner for London, Claire Waxman, and key victim support organisations in London. The survey asked a mixture of closed and open-ended questions to ascertain the experiences and views of survivors of being asked to provide technology evidence (usually phones) and the impact this had on them and their cases (please see Appendix B for the survey questions). The survey was in the field throughout October 2020. A total of 57 responses were received. Additionally, five in depth semi-structured interviews with MPS officers working within rape investigations were carried out to explore the experiences and perspectives of officers. This included both Sexual Offences Investigation Trained Officers (SOITs) and Officer in Charge 13 (OIC) roles. Invites were sent to all SOITS and DCs overseeing rape investigations in the MPS - 11 officers agreed to take part and for the final research five were interviewed. Responses were recorded, transcribed and analysed using thematic analysis to draw out key themes (see Appendix C for the interview schedule). 14 3. Results Before presenting findings in relation to police requests for victim technology devices, the first section of analytics seeks to compare the new time periods with the original MOPAC London Rape Review (2019). 3.1 Attrition of rape allegations In line with the 2019 Rape Review, the current research finds a high attrition of cases – across the sample as a whole and across the three time periods. Table 1 below shows case outcomes or current status of the full sample (overall), with a comparison of the three separate time periods studied (Period 1, Period 2, Period 3), as well as the findings from the 2019 Rape Review (2019 RR)11. Table 1 3.1.1 Victim withdrawal Victim withdrawal remains the most common form of attrition, accounting for almost two thirds of attrition overall (65%; n292)12. This remains consistent across the three time periods studied and is comparable to the 2019 Rape Review. 64% (n=264) of female victims withdrew, while 55% (n=28) of male victims withdrew. The tendency for victim withdrawal to take place early in the process - noted by the 2019 Rape Review - appears to have become more pronounced in the current sample: of those 11 Outcome information is not presented in an attrition funnel - such as the one used in the 2019 Rape Review, because the high number of still ongoing cases in the current sample at the time of analysis, make a like for like comparison of samples not possible. 12 When referring to withdrawal this includes victims that did not wish to pursue a police allegation from the outset, and those that never engaged or ceased to engage with the police investigation. 2019 RR Overall Period 1 Oct-Dec 17 Period 2 Jan-Mar 18 Period 3 Oct-Dec 19 450 150 150 150 V Withdrawal 58% 65% 64% 64% 66% Police NFA 29% 25% 29% 26% 21% Submitted to CPS 14% 7% 8% 11% 3% Ongoing - 5% 1% 3% 11% CPS Charge 9% 2% 3% 3% 1% CPS NFA - 5% 5% 7% 2% Reaches Trial 6% 1% 1% 2% 1% Awaiting Trial - 1% 1% 1% 0% Convicted 3% 1% 1% 1% 1% Acquitted - 0.2% 0% 1% 0% Case progress and outcomes 15 who withdraw, over one third (38%; n=112) do so within 7 days of reporting, and almost two thirds (64%; n=186) do so within 30 days13. In comparison, in the 2019 Rape Review, of those who withdrew, only 18% had done so within 30 days, 48% within 90 days and 77% within 180 days. The 2019 Rape Review noted that withdrawal reasons tended to be complex and interrelated, and that often multiple issues influenced the withdrawal decision. The current review coded withdrawal reasons across 14 categories (see table 2). For one in five cases that ended in victim withdrawal (22%; n=65) no specific withdrawal reason could be established, because the victim never engaged or ceased to engage with the police (i.e., respond to phone calls, messages and a seven-day letter asking if they wanted to continue). The most commonly identified withdrawal reasons were not wanting to or not having intended to report to the police in the first place (n=52) or wanting to put what had happened behind them and move on (n=52). External stress factors were identified as a reason in 32 cases (see table 2 below for a detailed overview of withdrawal reasons). Table 2 For those who withdrew within 7 days, within 30 days, and longer than 30 days from reporting, the top three reasons noted in the table above remain consistent with one exception. The bulk of withdrawals whereby the reason was ‘not wanting or intending to report rape’ (49 of the 52 instances) occurred within 30 days of reporting, and 29 of these within seven days – which is much to be expected. 13 17 of the 112 victims who withdrew within 7 days had attended a Haven; 33 of the 186 who withdrew within 30 days had attended a Haven. Reason for withdrawal No. Unable to establish 65 Did not intend to report rape 52 Wants to forget event and move on 52 External stress factors 32 Not wanting perpetrators life negatively affected 26 To be made safe / threat removed 21 Does not perceive assault as rape or crime 21 Length/stress/pressure of investigation 20 Fear of repercussions 14 Reporting out of civic duty / have it recorded 11 Concerns over courts/trial 11 Does not see point in proceeding 8 Concerns over VRI 5 Does not want to hand over requested tech 5 Unhappy with police procedure 2 Coded withdrawal reason 16 The victim disengaging or their withdrawal reason otherwise being unable to be determined is the most common withdrawal during various time-frames – be it within 7 days, within 30 days or over 30 days 3.1.2 Police NFA and CPS Submission After victim withdrawal, Police no further action (NFA) was the second most common attrition point, accounting for one quarter (25%, n114) of cases overall. The victim demographics in cases that resulted in police NFA were similar to those of the overall sample.14 While the proportion of cases ending in police NFA is similar to that observed in the 2019 Rape Review (29%), those submitted to the CPS for consideration to charge is lower: 6% (n=28) (compared to 14% in the previous review). This is likely to relate to the fact that 23 cases were still ongoing at the time of analysis, with most of these pertaining to the most recent time period examined (i.e., October-December 2019; n=19). 3.2 The profile of reported rape in London As with case outcome, the profile of victim, offender and offence characteristics remains broadly similar to that reported in the 2019 Rape Review. 3.2.1 Victim/survivor characteristics Consistent across the three time periods, the general demographic profile of victims matches the 2019 Rape Review (see table 3 below). Those who report rape in London are still predominantly female and young (with an overrepresentation of those aged under 18). The small increase in the proportion of Asian victims - as observed in the 2019 Review - appears to have continued. A similar, but slightly lower proportion of victims in the overall sample, 37% (n=166), presented with mental health issues when compared to the 2019 Rape Review (41%). Table 3 14 55% white ethnicity, 31% black ethnicity; 84% female, 16% male. 2019 RR Overall Period 1 Oct-Dec 17 Period 2 Jan-Mar 18 Period 3 Oct-Dec 19 Gender Female 89% 89% 89% 88% 89% Ethnicity White 66% 62% 62% 62% 62% Black 24% 25% 25% 25% 23% Asian 7% 10% 9% 9% 11% Age at offence Under 18 31% 30% 34% 32% 25% 19 to 29 36% 36% 35% 39% 35% 30 to 49 27% 28% 29% 25% 30% 50 plus 6% 5% 3% 4% 9% Proportions calculated with unknown values excluded 17 3.2.2 Suspect profile Again, with regard to perpetrators, demographics are similar to those reported on in the 2019 rape review and are reasonably consistent across the three time periods studied (see table 4). Table 4 Proportions calculated with unknown values excluded Just over one third of those named as suspects (35%; n=159) were noted to have a history of offending – a significantly higher proportion15 than that noted in the 2019 Rape Review (29%). This included 40 with a history of either sexual offending or domestic abuse but no other crime types; 63 with a history of other offences (but not sexual or domestic); and 56 with an offending history across these different types. Similar to 2019 Rape Review findings that suspects in domestic abuse cases were more likely to have a prior offending history, the current research found that in almost half of the cases (47%, n=65) where the relationship was intimate, the perpetrator was known for previous domestic abuse and/or sexual offending, compared to just 10% (n=31) of cases involving any other kind of relationship between the parties. 3.2.3 Types of rape being reported As per the 2019 Rape Review, and across the three time periods, in the majority of cases, victim and perpetrator were known to each other prior to the offence taking place - most commonly as either current or former intimate partners (35%; n=139), followed by acquaintance/friend (29%; n=116). In only 6% of cases the perpetrator was a total stranger.16 15 Statistically significant at p < 0.05. 16 To clarify: the Met Police use the terminology Stranger 1 to refer to a perpetrator who is a complete stranger (for example attacks a victim they do not know in a park or alley) and Stranger 2 to refer to a perpetrator that the victim did not personally know before the events that lead up to the offence (for example someone they have met in a bar, or on a train, on via a dating app). 2019 RR Overall Period 1 Oct-Dec 17 Period 2 Jan-Mar 18 Period 3 Oct-Dec 19 Gender Male 99% 100% 100% 100% 100% Ethnicity White 43% 43% 45% 44% 41% Black 35% 36% 35% 36% 39% Asian 19% 15% 17% 15% 13% Age at offence Under 18 16% 18% 18% 21% 14% 19 to 29 40% 33% 31% 35% 32% 30 to 49 38% 38% 39% 33% 43% 50 plus 6% 11% 11% 11% 11% 18 Table 5 In findings unchanged from the 2019 Rape Review, three in five offences (59%, n=264) took place in a private setting (victim’s home, 19%, n=86; perpetrator’s home, 21%, n=94; or a shared home 19%, n=84). This was mostly consistent across the three time periods, with the exception of offences taking place at the perpetrator’s home address which accounted for 26% (n=39) of cases during period two as opposed to 19% (n=28) and 18% (n=27) during periods one and three. As reported in the 2019 Rape Review, beyond the intrinsically violent nature of the offence in itself, the majority of cases did not involve serious violence. The victim suffered a physical injury in 17% (n=75) of cases - significantly17 lower than the proportions reported in the 2019 Rape Review (23%). The injury was recorded as ‘minor’ in 56 of those cases and ‘major’ in only 4. The victim resisted in 40% (n=180) of cases - the same proportions as those noted in the 2019 Rape Review - and were more likely to be verbal (38%, n=172), than physical (12%, n=54) (with there being an overlap of both in some cases). 3.2.4 Reporting Similar to the 2019 review, two in five cases (39%; n=175) are reported to the police within a week of the offence occurring, a quarter (24%, n=106) within 24 hours. Almost one third of cases (31%, n=139) are reported a year or more after the offence was committed (see table 5). On the whole the three time periods are consistent, with the exception of a considerably higher proportion of cases reported between a month and six months after the offence during period 1 compared to the other two time periods. 17 Statistically significant at p < 0.05. 2019 RR Overall Period 1 Oct-Dec 17 Period 2 Jan-Mar 18 Period 3 Oct-Dec 19 Base 450 150 150 150 Stranger 1 7% 6% 7% 5% 6% Stranger 2 11% 15% 17% 11% 17% Intimate 35% 35% 35% 33% 36% Familial 12% 13% 10% 17% 13% Acquaintance / friend 34% 29% 31% 32% 24% Professional / carer 2% 2% 1% 1% 4% Relationship between victim and perpetrator 19 Table 6 3.2.5 Video Recorded Interview Completing a video recorded interview (VRI) is a necessary part of the investigative process if a case is to progress to the CPS and court (although the victim does have the option of completing a written MG11 statement). The 2019 Rape Review found that the VRI often represented a key barrier for victims. And indeed, the proportion of cases where a VRI was completed have remained unchanged between the 2019 and the current review (34% and 35%, n=157, respectively). Almost half of VRIs completed are done so within a week of reporting (47%, n=74), and almost three quarters are done within a month of the report (73%, n=114). In 36% of all cases, the victim declined to take part in the VRI and in a further 8% the VRI was arranged, but not completed. In almost all of these cases (182 out 196, 93%) the victim subsequently withdrew from involvement with the investigation (13 or the remaining cases ultimately resulted in police NFA while one is still ongoing). A considerable proportion of victim withdrawals took place before a VRI could be completed (n=238, 85%). In these cases, withdrawal reason was more commonly noted to be related to the victim not wanting to report or not having wanted to report in the first place, not or no longer perceiving what happened as rape or a crime, or not wanting the perpetrator’s life negatively affected).18 3.2.6 Suspect identification and arrest A suspect is named or subsequently identified in 62% (n=281) of cases, a figure that remains consistently similar across the three time periods. The suspect is subsequently either arrested or interviewed under caution in 42% of cases (figures comparable to those observed from the 2019 Rape Review data where 60% were identified and 39% arrested or interviewed under caution). 18 This difference was statistically significant according to a chi-square test: 2 (1) = 6.01, p=.014. 2019 RR Overall Period 1 Oct-Dec 17 Period 2 Jan-Mar 18 Period 3 Oct-Dec 19 Base 450 150 150 150 24 hours 29% 24% 20% 24% 27% 1 day - 1 week 16% 15% 19% 13% 15% 1 week - 1 month 12% 11% 9% 11% 13% 1 month - 6 months - 13% 21% 9% 8% 6 months - 1 year - 6% 6% 5% 8% 1 year plus 27% 31% 25% 38% 29% Time between offence and report 20 3.2.7 Third Party Material Police requested third party materials, such as medical, educational or social services records in a quarter (25%, n=111) of all cases. While the proportions were similar between Periods 1 and 2 (29%, n=44 and 30%, n=45 respectively), requests dropped significantly19 in the third period (15%, n=22). Across all periods examined in the current research, proportions of cases where third party material is requested remain lower than those found by the 2019 Rape Review (36%). As outlined in table 7 below, victim medical records were the most commonly requested material, followed by social services records and education records. Commonly (66 occasions) there was an overlap with more than one type of record or material being requested. Table 720. Mostly, victims agreed for personal records to be obtained, declining to give consent in just 7% (n=8) of cases. There was a wide range of time between third party materials being requested and received (between 0 – 686 days in the current sample). The mean time difference is 88 days, the median 38.5 days. 3.3 Technology Evidence The research so far has sought to compare results across 2017, 2018 and 2019 to the 2016 data reported on in the 2019 Rape Review. The next section will focus on the evidence in relation to key questions that have remained unanswered by previous research, examining the role of technology, including overall prevalence of requests for victim technology, whether victims subsequently consented - and police obtained the device - or whether they declined, and how this impacted on case progression and outcome. Where possible, comparisons to the 2019 Rape Review will be made21. 19 Statistically significant at p < 0.05. 20 There is no comparison to the 2019 Rape Review as although police requests for third party material was coded for as part of that research, it was not broken down and specified as it has been for the current review. 21 The 2019 Rape Review captured variables in relation to whether reference to the existence of possible technology evidence was made within the CRIS report; whether victim/survivor technology was obtained; whether suspect technology was obtained; and whether obtained technology evidence supported the victim’s, the suspect’s or neither case. n % of TP requests Medical records 74 67% Social Service Records 51 46% Education Records 40 36% Other 42 38% Types of Third Party material requested 21 3.3.1 Victim Technology Requests A police request to take a technology device from the victim occurred in just under one quarter (23%; n=102) of all cases. There were no statistically significant differences in this across the three time periods examined (see table 8). In 48% (n=49) of cases where police made a request for victim technology, the victim consented to this request and police consequently obtained the device. There were two cases where the victim consented to their device being taken, but the case was concluded due to other factors before the technology was actually obtained. In all other cases, where a victim’s device was requested but not obtained by the police, this was as the result of the victim declining. When considering the overall sample, this makes for an exact 50/50 split between those who consent and those who refuse to hand over their device22. Overall, this constitutes 11% of cases in the sample where victim technology is obtained by the police - the same proportion as in the 2019 Rape Review. Whilst proportions of police requests for phones remain consistent across the three time periods examined, proportions of victims consenting to having their device taken appear lower during period 2 (36% compared to 61% in period 1). However, this difference is not statistically significant. Table 8 In most cases where a request for the victim’s device was made, this was preceded by a specific reference to the existence of technology evidence (87%, n=89) 23. There were only 13 cases overall where this had not been the case. Again, this remains consistent across the three time periods. Altogether, reference to the existence of possible technology evidence was made in 37% (n=166) of cases - significantly24 higher than the 27% of cases in the 2019 rape review. 22 The outcomes for cases where the victim refused consent to hand over their device to the police were: victim withdrawal in 31 cases, police NFA in 16 cases, and submission of the case to the CPS in 6 cases (2 of these were charged, and 2 were ongoing at time of the research). 23 This was coded when there was reference in the CRIS to potentially relevant evidence involving communications devices and/or social media in the victim’s account of the crime. This included where messages or calls may have been exchanged between the victim and perpetrator and / or witnesses in reference to the assault and circumstances surrounding it, or relevant social media posts. 24 Statistically significant at p 0.05. Overall Period 1 Oct-Dec 17 Period 2 Jan-Mar 18 Period 3 Oct-Dec 19 Base 450 150 150 150 Request for device (% of all cases) 23% 22% 26% 20% Victim consents to request (number) 49 20 14 15 Victim consents to request (% of requests made) 48% 61% 36% 50% Requests for and obtaining victim's tech device 22 3.3.2 Further observations in relation to technology evidence Suspect Technology A technology device was obtained from the suspect in 18% (n=80) of all cases - a slightly higher proportion than the 13% reported in the 2019 Rape Review. This constitutes 41% of cases where the suspect was spoken to (either arrested and / or interviewed under caution). In two thirds of the cases where a technology device was obtained from the suspect, the police had also made a request for victim technology (65%; n=52). Overall, there were only 27 cases where police obtained both victim and suspect technology - constituting one third (34%) of all the cases where suspect technology was seized and a little over half (55%) of all those cases where the victim consented to hand over their device. Technology evidence and timeliness Data on when victim technology was obtained and subsequently returned could only be gauged from the CRIS reports in 20 cases. It is therefore not possible to provide reliable information on timeliness in relation to victim technology evidence. To illustrate - the mean number of days a device was kept by the police in the 20 cases where relevant data was available was 89 days. However, this contains considerable variation: half of the devices were returned within a month (eight of them within 4 days), whilst in six instances the device was kept for more than 100 days and in one case for more than a year. Evidential material gleaned from victim technology devices Information on whether potentially relevant material had been found on victim technology devices could only be ascertained from the CRIS report in 25 of the 49 cases where the victim agreed to hand over their device. In 11 of them, no relevant material was found, whilst in 14, material taken from the victim’s phone provided evidence that was considered to potentially support the prosecution case (n=9), the defence (n=4) or both (n=1). 3.4 Predictors of attrition The next stage of the report seeks to replicate the methodology from the 2019 Rape Review and conduct analysis to identify the variables that can predict the two main forms of attrition in this sample: victim withdrawal and police decision to take no further action (NFA). The innovative aspect in this report specifically seeks to incorporate the new victim technology variables to see if these relate to the key outcomes of interest. It was not possible to conduct predictive analysis on the likelihood of attrition following CPS submission because of the low numbers of cases at this stage (n=29 submitted to CPS). As before, case characteristics were split into four categories: victim/survivor, suspect, offence, and procedural. The four categories were first examined in separate models before being combined into an overall model.25 This process allowed the relative contribution of each 25 Univariate logistic regressions were first conducted on each case characteristic and the two outcomes. Significant univariate predictors were retained for inclusion in a multi-variate model for each of the four categories of case characteristics. The characteristics that remained significant in each multi-variate model were 23 category to be determined, and variables that remained significant in the overall model can be considered predictive of outcomes after controlling for the characteristics in the other categories. The subsequent sections summarise the key findings of the predictive analysis. Detailed findings in relation to those case characteristics that were significant predictors of victim withdrawal and police NFA can be found in Appendix D. 3.4.1 What factors predict victim withdrawal? The 2019 Rape Review found that procedural characteristics were more important in predicting withdrawal than victim, suspect and offence characteristics. The current research finds a broadly similar picture. When examining each category of case characteristic separately, only one victim/survivor characteristic significantly predicted victim withdrawal: victims/survivors aged under 18 at the time of reporting were significantly less likely (x2) to withdraw compared to victims/survivors aged over 18. It is worth noting that whilst in the 2019 Rape Review, male victims were found to be three times less likely to withdraw compared to female victim/survivors, this is no longer the case in the current sample. In terms of suspect characteristics, cases where the suspect was or had been in an intimate relationship with the victim/survivor were significantly more likely (x2) to end in victim withdrawal. This is different from the 2019 review, though may be related to the impact of reporting via DASH, which will be outlined in the next section. When looking at offence characteristics, unknown offence location was found to be a significant predictor, making victim withdrawal 3 times more likely. This finding appears to relate to cases with substantial evidential difficulties, with significantly more cases with an unknown location seeing a victim withdrawal that took place within the first month of the investigation and where the withdrawal reason related to the victim not having wanted to report in the first place or no longer considered what had happened to them a crime. Furthermore, and in line with the 2019 analysis, victims/survivors who reported their rape in response to the DASH questions asked when police attended a domestic abuse call were significantly more likely (x3) to withdraw than victims/survivors who reported by other means. Similar to the 2019 Rape Review, variables relating to the reference, request or obtaining of victim or suspect technology26 were significantly related to case outcomes on their own but did not remain significant when included with other procedural characteristics in multivariate models. However, cases where police encountered technical difficulties with victim and/or suspect technology evidence were significantly less likely to end in victim withdrawal (x5). then retained for inclusion in the final overall model. Cases flagged as likely false allegations were excluded from the predictive analysis (n=15). Ongoing cases with no outcome were excluded from all predictive analysis (n=18). When predicting police no further action, cases that were withdrawn by the victim were also excluded (n=288). 26 Please note that the 2019 review did not specifically code for whether police made a request for technology, only whether a reference was made and whether technology was obtained. 24 A number of other procedural characteristics were significant predictors of victim withdrawal: cases where police expressed doubts as to victim credibility (3 times less likely). Finally, and in a repeat of findings from the London Rape Review 2019 review, cases where the victim/survivor completed a Video Recorded Interview were significantly (11 times) less likely to end in victim withdrawal, further evidencing the potential barrier to case progression first outlined in the 2019 review. Indeed, most victims who withdrew did so before participating in a VRI, with only 15% (n=44) of those who withdrew completing a VRI, compared to 35% in the sample overall. The finding that when police expressed doubts as to victim credibility, victim withdrawal was significantly less likely may seem counter-intuitive at first but may be explained by the fact that this variable in fact made another outcome more likely: that of police no further action. When looking at the four categories of case characteristics simultaneously to determine the relative contribution of each category to predicting victim withdrawal, as in the London Rape Review 2019, the strongest predictors of victim withdrawal were the procedural characteristics: all three (i.e., police encounter difficulties with technology evidence; police express doubts as to victim credibility; completion of VRI) remained significant in the overall model. Suspect and offence characteristics were no longer significant predictors after taking the procedural characteristics into account, suggesting these variables were less important in predicting victim withdrawal. Another way of determining the relative contribution of the four categories of case characteristics is by looking at model fit. Analysis indicates that the procedural models and final models were better performing than the victim, suspect or offence models. In line with findings from the 2019 London Rape Review, Procedural characteristics were better predictors of outcome. 3.4.2 What factors predict police no further action? When exploring what predicts ‘no further action’ (NFA) on a case, caution should be taken in interpreting the results due to a lower sample size (n=139), and the small number of CPS submissions in particular (n=29). In terms of victim/survivor characteristics, cases where the victim was aged over 30 at the time of the offence were significantly more likely to result in NFA (x4). To some extent this is likely to be reflective of the evidential difficulties associated with intimate partner rapes; victims over 30 were significantly more likely to have such a relationship with the perpetrator.27 Where suspects had either previous offending histories (x3) or mental health issues (x3), cases were significantly less likely to result in NFA. As would be expected given the Full Code test guidelines on evidence, offences which had named witnesses were significantly less likely 27 35% (n=15) of cases where the victim was aged over 30 had an intimate relationship with the suspect, compared to 19% (n=18) if the victim was under 30. This difference was statistically significant according to a chi-square test:2 (1)= 4.27, p.039. 25 to result in NFA (x3). Where the victim had suffered a physical injury as the result of the offence, cases were also significantly less likely to end in an NFA decision, again reflecting evidential influences. Whilst victim injury may be considered another strong evidential factor in progressing a case. As with the 2019 Rape Review, procedural characteristics were generally the strongest predictors of NFA. If the police noted an inconsistent account from the victim/survivor, cases were significantly more likely to end in NFA (x12). Conversely, seeking early advice from the CPS (x12) or requesting 3rd party materials (x13) indicated a stronger evidential case and made cases significantly less likely to end in NFA. Similarly, where a suspect was arrested the case was significantly less likely to end in NFA (x12). The overall model looked at all four case characteristic categories simultaneously, enabling comparison as to the relative contribution of each category in predicting police NFA. This shows that victim and suspect characteristics, as well as offence characteristics are no longer significant after taking procedural characteristics into account. As with victim withdrawal, and in line with previous findings, procedural characteristics remained significant in the overall model, suggesting these variables are most important when predicting NFA compared to variables relating to the suspect or offence28. 3.4.3 The role of technology variables in predicting case outcome The only technology related variable that remained a significant predictor of outcome in the overall model, was whether police struggled with technology evidence - this made victim withdrawal significantly less likely (x5). It may be that this variable acts as a marker of case progression - in other words, cases reach the point where police try and extract technology evidence (and encounter issues). Indeed, the proportion of cases ending in CPS submission was significantly greater where police struggled with technology around evidence compared to those where no difficulties were noted.29 When considered independently reference to, request and obtaining of victim technology were all significant predictors of case outcome30, although analysis suggests this should be 28 It should be noted that two variables were excluded from the case characteristic and overall models due to co-linearity or a high degree of correlation to the dependent variable (NFA outcome) ‘Police doubt victim credibility’ correlated significantly with ‘inconsistent victim account’. On its own, cases where police expressed doubts about the victim’s credibility were significantly more likely to end in NFA (x11). In addition, ‘Previous sexual or domestic abuse victimisation‘ had a strong correlation with the dependent variable. Further exploration demonstrates that these results should be interpreted in the direction of CPS Submissions. The proportion of cases in which the victim has suffered previous victimisation for sexual or domestic abuse is similar for both NFA and Victim Withdrawal outcomes; the variable is not predictive of either NFA or Victim Withdrawals, but rather makes CPS submission less likely. Analysis also highlights some learning around variable creation, with an indication that sexual offences and domestic abuse victimisation should be coded separated, and greater consideration given to the nuanced issues of false or mistaken allegations within the sample. 29 For cases where either suspect or victim technology was seized, 41% of cases where police struggled with technology ended in CPS submission, compared to 15% of cases where no struggle is mentioned. This difference was statistically significant according to a chi square test: 2 (1)= 5.98, p.0145. 30 Univariate analysis found that cases where reference to the existence of technology evidence is made are significantly less likely to end in victim withdrawal (2x); 3x less likely to end in withdrawal where victim technology is requested and 4x less likely where victim technology is obtained. 26 interpreted in terms of a significantly increased likelihood of CPS submission, as opposed to making either NFA or Victim Withdrawal more likely. Further, the seizure of suspect technology (3x) and existence of other digital evidence (2x) also made victim withdrawal less likely. Unsurprisingly, when considered in isolation, victim refusal to hand over technology made victim withdrawal significantly more likely (4x). Further exploration indicates that victim technology requests are best seen as an indicator of the evidential strength of a case, becoming more likely as a case progresses, thus increasing the likelihood of CPS submission. In summary, when included in either the case characteristic level or overall models, neither reference, request or seizure of Victim Technology remain significant predictors of Victim Withdrawal or police NFA, suggesting that technology variables have some influence, but not as much as the other procedural variables that remained significant in the multivariate model. 3.5 Predictors of police requests for victim technology Using the same approach applied to case outcomes, analysis also explored which variables predicted police requests for victim technology (please see Appendix E for a detailed overview of variables that predicted police requests for victim technology). In terms of victim/survivor characteristics, victims aged over 30 at the time of reporting were significantly less likely to be asked for their mobile phones or other technological devices (x2). Similarly, in cases where suspects were aged over 30 at the time of reporting, police were significantly less likely to request victim technology (x2). Where the victim and suspect were in or had been in an intimate relationship, victim technology was also less likely to be requested (x2), compared to cases where the suspect had been a stranger (1 or 2 combined). Linked to this, cases with a Domestic Violence flag were significantly less likely to result in a police request of victim technology (x3). Where the case featured a black or minority ethnic (BAME) suspect, victim technology was more likely to be requested (x3). This finding may be linked to the role played by victim- suspect relationship: BAME suspects were significantly more likely (x3) to be in a stranger 2 relationship with the victim, compared to any other relationship, indicating a likelihood that BAME suspects are more involved in the type of offence which generates victim technology requests. Cases in which named witnesses were identified were significantly more likely to result in police requesting victim technology (3x), as were those where the victim/survivor resisted the attack either physically or verbally (3x). Both are likely related to evidential grounds. A number of procedural characteristics predicted police requests for victim technology. The referral to an ISVA was linked to an increased likelihood of victim technology requests (2x), as was completion of a VRI (2x), seizure of suspect technology (6x), police accessing other digital material (2x) and requesting third-party material (2x). When examining all four case characteristic categories simultaneously, predominantly procedural variables remained - mainly those that can be considered signals of case progression: completion of VRI; suspect technology obtained; other digital material accessed; 27 and third-party material request. An exception to this pattern is the referral to an ISVA, which also remained a significant predictor of police technology request in the overall model, alongside victim resistance and a suspect aged over 30. Suspect ethnicity was no longer a significant predictor of victim technology requests in the overall model. 3.6 The perspectives and experiences of survivors and officers The final section of the report incorporates the views of both victim/survivors as well as officers working on rape investigations to bring additional context to the data discussed so far. In doing so, it focuses on the experiences of 57 survivors who responded to an online survey and those of five MPS officers who shared their views and perceptions in semi- structured interviews. The survivor survey responses were analysed, and content analysis was conducted on both the free text responses to the survivor survey, as well as the officer interviews. This identified a number of key themes, some of which overlap across victim/survivor and officer responses, whilst others illustrate the existence of differences and contrasts between perspectives and experiences. 3.6.1 Survivor-reported experiences with technology evidence The majority of the 57 survey respondents reported that police had made either a formal request to examine their device (n=37) or referred to this as a possibility, though no actual request was later made (n=16). This difference in prevalence, compared to the crime data sample, is likely a reflection of those who responded to the survey being more representative of victims whose cases had progressed (e.g., only a minority had withdrawn from the investigation or court case, whilst 16% (n=9) had seen their case either charged by the CPS, had reached or were awaiting trial, or had seen a conviction. Of the 37 respondents who said that the police had made a formal request, 27 agreed to hand over their device. 3.6.2 Survivor and officer experiences in relation to technology evidence Officer decision-making on victim technology requests: There was consensus among the officers interviewed that police requests for victim technology usually only took place in response to specific reference to the existence of such evidence. In this context, the VRI was seen as a key point in the investigation, providing an opportunity to review and discuss evidence and next investigative steps, including phone evidence where relevant: ‘I will sit down with the victim, normally when we go to do the VRI. We tell them why we want their phone, what we want to download and that we are not going to download everything...’. The officer feedback fits with the findings of the data analysis, which suggests a link to the VRI and other markers of case progression. However, feedback from the survivor survey suggests this is not always the case and technology requests can be made at various stages in the investigative process. Of the 37 survey respondents who said that the police had made a formal request, 16 said that it was later in the process, whilst 19 indicated that it had 28 happened earlier on31 (for 8 when they first reported and for a further 11 when the SOIT officer first contacted them32). Victim concerns about technology evidence: Whilst some officers reported that they rarely encountered victim concerns or resistance, others reported that concerns and push backs were common, with younger victims in particular being more likely to raise concerns about their phone being taken. When concerns were encountered, they tended to be about timeliness and the practicalities of not having access to a phone, but also concerns and questions about what will be looked at and who by. Concerns about the police request for technology were expressed by the majority of survey respondents: 31 perceived it to be intrusive - a view shared by those to whom it was referred to as a possibility, but no request was made (n=13). The majority of both those asked (n=29) and those to whom the possibility was suggested (n=14) also felt that handing over the device would cause them practical difficulties. Whilst police felt they were usually able to reassure victims (by outlining what they would be accessing and explaining that they would not be looking through every detail on the victim’s device), survivor feedback suggests that unease and discomfort frequently remained. Of the 37 survey respondents who said that their device was requested, 27 felt that the police explained the purpose for this to them – 16 of them were satisfied with this explanation, while 11 were not. Key reasons for finding the police explanation unsatisfactory related to either a perceived lack of clarity and/or detail, including in some cases being given only a ‘blanket policy’ as an explanation, or feeling the police explanation strongly implied that the key objective of technology seizure would be to check the victim’s account of events, which was felt to be very unsettling: "I thought it was unreasonable - I was not under investigation/a suspect of a crime". It is also worth noting that whilst both survivors and officers talked about intrusiveness and practicalities as key concerns, officers tended to talk about these issues in the context of the investigation (i.e., officers going through material), whilst survivors’ concerns were more focused on later stages of the criminal justice process (relevance and use of material during trial). In this respect, survivors’ concerns related to potential privacy breaches and fear of who may access the material, as well as questions surrounding how the content would be used during a trial. Relevance of the material was a key concern: in historic cases for instance, where devices did not exist at the time of the event, yet a request had been made. Accessing technology evidence: Officers talked about the practical challenges of dealing with technology evidence. This mainly related to having to rely on others who were trained to undertake phone downloads, the provision and accessibility of which was seen as insufficient and as inconsistent across 31 Two respondents said they could not remember. 32 Typically, the SOIT will be in contact with the victim as soon as possible: the victim will make the report, it will be referred to Sapphire, and a SOIT will call them as soon as they can, while there are, of course, sometimes delays in this, and some victims who disengage with the police very early may not be responsive to contact. 29 boroughs and teams: ‘…it is a postcode lottery. If you have no downloaders, then your request goes into a queuing system…the whole downloading thing is pot luck’. With current processes not as reliable, streamlined or efficient as they could be, subsequent delays to case progression also meant additional pressure on officers to manage the victim relationship whilst waiting for others. This was reflected in the survey responses, where survivors gave mixed feedback on the length of time their phones had been kept: of the 27 respondents who agreed to hand over their phone to the police, the device was returned to the victim within a week in 8 cases and a further 2 had it returned within a month. However, for 8 victims it was up to 6 months, and in 1 case up to a year (the remaining respondents either could not remember (n=2) or said that the police still had their device (n=6). Where there was reliable, consistent access to a trained downloader (or where officers had been able to access and complete the training themselves), it made the process considerably more straightforward, resulting in greatly reduced turnaround times for victims’ phones and the ability to coordinate a phone download with minimal disturbance to the victim. Indeed, a suggestion was made by officers interviewed that training SOITs in downloading and looking through material would be beneficial in terms of timeliness and officer autonomy in progressing an investigation, but also in terms of reassuring victims that the officer going through their phone would be someone they had already built a relationship with. Phone searches and identifying relevant material: Whilst officers acknowledged that recent software improvements had enabled better searching, it was still seen as very time consuming to go through downloads. It was not uncommon to spend a whole shift on one phone download - something that was incompatible with other demands, such as responding to new and / or urgent reports, resulting in delays and backlogs. These delays in reviewing material impacted overall timeliness but also ran the risk of evidence being compromised or lost (e.g., by the time material is accessed there may be technical issues, which then means having to go back to the victim, who may have deleted messages / material). Most officers interviewed felt confident in conducting searches, but nevertheless were very aware of the impossibility of going through everything, whilst also worrying that something important could be missed. This had been exacerbated by the impact of R v Allan and some suggested that this had made officers spend more time on reviewing downloads: ’It’s one of those things that is a necessary evil but otherwise you would end up in a R v Allan situation. There is a certain amount of fear that you may miss something, when I go through a phone I err on the side of caution definitely and spend way too much time because I personally don’t want to be in that position…’. Overall, there did not seem to be a consistent methodology to reviewing downloaded material and officers interviewed described their own individual approaches and taking a case by case approach. Whilst any review of a phone download will inevitably need to be driven by the needs of the individual case, this nevertheless suggest there may be a need for more guidance (and training) on how to systematically approach phone searches. Indeed, one interviewee talked about limited police knowledge and training need regarding better 30 understanding social media, apps and how they are being used; and how to optimise a phone download and search. Only very limited information could be gleaned from survivor feedback regarding what material was gleaned from devices and how proportionate or relevant it was. The majority of respondents answered ‘no’ to the question ‘Did the police tell you what material they retrieved from your device?’ (19 out of 26), whilst 2 of the 7 who said yes disagreed that the material retrieved was relevant. Overall, 15 of 26 survey respondents had mixed views over the material retrieved from their device being used in a court case; 7 were happy for it to be used and 4 not happy for it to used. Differing perspectives when assessing (technology) evidence: When officers talked about technology evidence, they did so in the context of an opportunity to secure corroborating evidence (e.g., victim texting friends about what happened; or using phone location). This viewpoint was felt to be in direct contrast to that of the CPS, who were seen to be frequently asking for more speculative phone searches, often to check whether there were undermining messages. Indeed, there was a perception that the CPS would drop a case if a victim did not give their phone - a message the police feel obliged to pass on to victims: ‘There have been cases where people have been reluctant to give us the phones, then we have to say the CPS will drop the case, which is what they have said to us and is really horrible.’ Some of the contrasting views in relation to technology evidence were seen as part of much wider issues around different perspectives between police and CPS. Those interviewed felt that getting cases through the CPS was a key challenge and illustrated some of those contrasts: while police focused on the evidence they had and on reasons why a case should go to court and be decided by a jury, the CPS was seen as focusing predominantly on the weaknesses and reasons for why a case should not go to court - a situation that, in the eyes of the officers interviewed, has been exacerbating in recent years. The majority of survivors who responded to the survey had not seen their cases progress to trial. There was a strong sense from respondents that cases were dropped despite strong evidence and many were left feeling that the overall objective had been to prevent cases from progressing, possibly for reasons related to limited finance, time or resources available within the criminal justice system. One respondent felt “things were used to drop the case that shouldn’t be“, and another stated that the case was “dropped by the CPS due to lack of evidence despite my rapist filming the attack with in-home CCTV…“. Such instances go to stress the importance of ensuring effective and just victim engagement and treatment throughout the process. 3.6.3 Wider reflections Meeting the needs of the victim and those of the criminal justice system: Officers interviewed felt that wider emotional support and counselling were the most pressing needs for many of the victims they worked with, yet something that was very difficult to access - either due to waiting lists or due to limitations around pre-trial therapy whilst an investigation is ongoing. In this context, there was a lot of appreciation of the wider services 31 available to provide support to victims, particularly with regards to Independent Sexual Violence Advisors (ISVAs): ‘a lot of the general and psychological support is now removed from the police which I think it should be, you should be empathetic and a SOIT officer is really well trained at that but a lot of that has been deferred to other support services.’ Particularly for SOITs, working in a role that required them to support the victim, but also a system they considered to be frequently scrutinising of victims, was felt to be challenging. For some interviewees this had meant taking steps to protect the victim from potential scrutiny and judgement - for example by not mentioning within the CRIS report that the victim had expressed concerns about progressing the investigation because this may be used against them in court. At the same time, officers also talked about disclosing information to the CPS that they felt was not relevant but still considered it something the CPS would want to know: ‘I may put in my material to the CPS that the victim has multiple sexual partners every week, based on phone downloads from Tinder, which I don’t think is undermining necessary but it is something the CPS should know for them to make their decision. It is relevant, I don’t think it helps or undermines the case, it’s just context that’s known.’ For victim/survivors, the disclosure of private, intimate information to the defence, the accused and subsequently the public, was a commonly expressed concern. For some, such a disclosure felt like a form of betrayal (by police and CPS), insofar as enabling the defence, during court proceedings, to twist the narrative and/or call into question the victim’s character and account based on a misrepresentation of the disclosed information. Stark contrasts in officer vs survivor perceptions of relationships and experiences: Looking across the feedback and experiences shared by both officers and survivors, some considerable differences in perceptions can be seen. Officers talked about having gone into their roles with the specific goal of working with and supporting victims. In this sense they spoke about feeling a sense of care and protectiveness, and that they considered it important for victims to have a positive, supportive experience of the police. This is in stark contrast to the negative experiences that were frequently recounted by the survivors who responded to the survey. More than half of the survey respondents (26 out of 41) described experiencing negative interactions with the police. The main concerns included victim blaming attitudes, poor communication and a perceived absence of empathy and support. Repeatedly victims felt their treatment was more akin to that of a suspect. One respondent said there was “no level of understanding and I was made to feel like I was an inconvenience “, whilst another concluded: “my experience of the police has put me off ever engaging with them again". Importantly, not all interactions reported were negative; 5 of the 41 survey respondents documented favourable interactions with the police, focusing predominantly on the police treatment of them as respectful, kind and empathic: "The DC I dealt with was absolutely wonderful…I really trusted him, and he was so kind and compassionate when I needed it the most". 32 Nevertheless - and it is difficult to deconstruct experiences specifically with the police from those with the CJS overall - survey respondents frequently spoke of the adverse effects with regards to their emotional, mental and/or physical health of their experiences following the incident and the process that ensued once they reported: “The whole experience from my disclosure was and remains as traumatizing as the rapes themselves”. The length of the overall process, and delays in particular, were highlighted as particularly problematic. Victim withdrawal and ‘good’ outcomes: Officer experiences reflected the findings of the current and previous rape review case coding: that victim withdrawal was common and often happened very early in the process. Indeed, officers talked about seeing early indications of likely victim withdrawal - and felt this was usually not only unchangeable but a choice that needed to be respected. Echoing in part the victim withdrawal reasons from the data sample - in the experience of the officers interviewed, reporting rape did not necessarily mean a victim wished to pursue a criminal justice outcome. Often, victims reported because they wanted to have what happened noted, they wanted to tell someone, but not take it further: ‘sometimes they are doing it because they feel they should, or they want someone to acknowledge what happened to them and believe them, whilst not necessarily having it go to court….’. Despite this, officers talked about trying in each and every investigation even if they got a sense the victim was likely to withdraw - there was always potential for it to be picked up again in the future. Furthermore, in the context of low conviction rates - something officers felt most victims were very aware of - the experience of reporting can be closure - and this was why a lot of importance was put on police treatment of the victim and the victim experience of the police as sensitive, being taken seriously and as having their case progressed efficiently. Practical challenges as key stressors for officers: Whilst it was acknowledged that the nature of the work, especially when working on particularly difficult cases, could impact officers, any stress experienced was felt to be predominantly to do with practicalities, structural issues and lack of support, not as a direct result of working within rape investigations. All officers interviewed described high workloads and caseloads – and always chasing a backlog. The latter was often a consequence of dependencies on others, both internal (e.g., lab results / toxicology, or phone downloads) and external (e.g., third-party material), but holding all these strands of an investigation, whilst also managing a relationship with the victim, carried considerable emotional labour for officers, SOITs in particular. Overall, those interviewed felt that lack of resources made their roles stressful and that there was a need for support with practical / basic things to enable them to do their job more efficiently. This included having access to cars, USB sticks, evidence bags, interview disks – but also access to people sufficiently trained to support them (e.g., phone downloaders or someone with PNC access to conduct checks). 33 Related to this was a perceived lack of access to training – again, it was mainly basic things that were needed, such as PNC access, or phone downloads. This would enable officers to do things themselves and progress cases rather than being dependent on others. 34 4. Discussion The 2021 London Rape Review revisited an in-depth case coding methodology to analyse data on reported rape in London. Through the coding of 450 crime reports across 140 variables, the resulting data set provides an updated picture and a follow up to the 2019 MOPAC London Rape Review. Additionally, the current research provides evidence around a much-debated topic: that of technology evidence and the use of victim technology (phones) in rape cases. 4.1 Summary of key findings The current landscape of reported rape in London - Comparison with previous reviews Consistent with previous reviews, the 2021 London Rape Review found steep attrition of cases following initial reporting. In the current sample, two thirds of cases ended with the victim withdrawing, whilst a further quarter ended in a police NFA decision. Similarly, victim/survivor, as well as suspect profiles and offence circumstances have remained broadly consistent. Those who report rape in London continue to be pre- dominantly female and young - with almost one third under the age of 18 at the time of the offence. Black victims continue to be over-represented when compared to their overall proportions in the population of London. Suspects in the cases reported continue to be men with a slightly older age profile compared to victim/survivors. Again, black suspects are over- represented when compared to their overall proportions in the London population. Consistent with previous research and the 2019 Rape Review, in the majority of cases victim/survivor and perpetrator are known to one another, and there is a considerable overlap between rape and Domestic Abuse. In terms of the above characteristics of reported rape in London, there was little change over the three time periods under study. Victim withdrawal has consistently remained the largest attrition point (65% of cases withdraw) and the current research finds evidence that withdrawals frequently take place soon after reporting and are often linked to victims not or no longer wishing to pursue a criminal justice outcome. This suggests a need to revisit whether alternatives to criminal justice outcomes need to be considered (see also Angiolini, 2015) and review support provisions for victims, as well as additional steps that could be taken with the aim of public protection. Predictors of attrition The modelling replicated previous analytics into exploring the factors that can predict victim withdrawal as well as police NFA. Here we see many similarities with the 2019 Rape Review and overall it is procedural characteristics that contribute most to the outcomes of interest as opposed to victim, suspect of offence characteristics. Of note is particularly the role of VRI completion in making victim withdrawal significantly less likely, suggesting that its function as a barrier has continued. Variables relating to an assessment of the victim’s account and victim credibility also impact on outcome, with inconsistencies in the victim’s account making police NFA more likely, whilst markers of case progression (i.e., request for third party material, suspect arrest, Early Investigative Advice) make police NFA less likely. 35 The role of victim technology evidence Turning to the use of victim technology - one of the core research aims - the current review found that in just under one quarter of cases a request for victim technology was made by the police - a proportion that remained broadly consistent across the three time periods examined. In all but 13 cases, such a request followed an explicit reference to the possible existence of technology evidence. As such, the current review does not find evidence of police making blanket requests for victim technology, nor of a spike in such requests after R v Allan. Following a request for technology, half of victims agreed to handover their device, whilst half declined. The proportion of victim consent drops slightly in the time immediately following R v Allan. Whilst this could be related to publicity around the case and media attention – it should be noted that the change was not statistically significant. Either way, the 50 / 50 split in terms of victim consent to the request indicates that more work needs to be undertaken in the engagement of victims with respect to technology evidence. When considering overall case progression and outcomes, reference to the existence of technology evidence, as well as request for and obtaining of victim technology were all found to be linked to a significantly increased likelihood of CPS submission. However, when included in the overall predictive model they did not remain significant, suggesting that technology variables have some influence on case outcome, but not as strongly as other - procedural - variables. The current research also sought to understand which cases victim technology was requested within. Again, procedural characteristics were strongest at predicting the likelihood of a victim technology request in the current review - mainly, again, those that can be considered markers of case progression: VRI completion, ISVA referral, request for third party material, accessing other digital material and seizure of suspect technology all significantly increased the likelihood of a police victim technology request. Outside of procedural variables, victim resistance and BAME suspect made it more likely that victim technology was requested, whilst a suspect over 30 made it less likely in the overall model. The finding that cases featuring BAME suspects appear significantly more likely to involve a victim technology request is worthy of further attention. Whilst this aspect does not remain significant in the overall predictive model, and analysis suggests links to the victim-suspect relationship variable, it also sits within a wider context of disproportionality of both black victims and suspects in the overall sample – a consistent finding across previous rape reviews. These findings also speak to the Mayor’s Action Plan to improve trust and confidence which has a focus on the use of police powers and how they impact on Black Londoners. The findings of the current review suggest that the prevalence of Black victims and suspects as well as any potential impact of ethnicity on case progression will be key issues to monitor going forward. The perspectives of survivors and officers The current review, for the first time, incorporated feedback from both survivors and police officers, adding a much-needed voice to the statistics. Victim/survivors’ concerns about technology requests were frequent and related to the practicalities of not having access to 36 their phones, but also the perceived intrusive aspect of this. The latter was a particular concern for victims in the context of worrying about potential privacy breaches and how the content would be used during a trial. In this respect, survivors frequently expressed feeling that the overall objective was to check their account of events - making them feel like a suspect - or, particularly if disclosed to the defence - to discredit them. Officers, on the other hand, tended to talk about technology as an opportunity to find corroborating evidence and were clear they only requested victim technology if there was an explicit reference to the existence of such evidence in a case. A key theme across the victim/ survivor feedback pertained to issues of fair treatment and engagement, with a consideration here of how the police broached the topic on requesting technology. Results indicate something of a perception gap: whilst officers described a considered and reassuring approach, this view was not shared by all the survivors who responded to the survey. They outlined that whilst officers had explained the purpose of the technology request in most cases (albeit not all), many victims were not satisfied with the explanation, wanting more detail. The small sample sizes in the qualitative element here are a limitation, but they nevertheless point to the importance of effective victim engagement, with expectation management, fair treatment, reassurance, regular information provision and empathy as critical factors. There are routine measures in place that capture feedback from many victims of crime and their satisfaction with the police (e.g., via the MOPAC User Satisfaction Survey33). However, this currently does not include victims of high harm crimes, such as rape, and there is therefore a gap in capturing and learning from the victims of these offences to generate insights and shape practice. It is proposed that future research will include a victim survey of High Harm offences. Indeed, the review found some further contrasts in the way survivors and officers reflected on their overall experiences of the criminal justice system. These were particularly prominent in relation to the support given to victims, with officers emphasizing their goal of providing a positive, caring experience, whilst victim/survivors frequently recounted feeling a lack of empathy and support. Whilst issues around effective victim engagement remain relevant here, officers also talked about carrying high caseloads and the challenges of progressing an investigation whilst having to wait on or chase others (e.g., for lab results, phone downloads, or for third-party material) - with resulting delays putting pressure on their relationship with the victim. This suggest that there is a need to look at what improvements could be made to the practical and emotional support officers currently receive. In relation to technology evidence in particular, this may need to take the form of more guidance (via training and / or toolkits) on how to systematically approach phone searches, but also points to the need for further input and training to help officers better meet victim needs and to close perception gaps. This could include joint training with victim support agencies, but also additional measures to support officers, such as offering reflective practice or clinical supervision - models of which have been found to be beneficial in other professions (e.g., nursing, teaching, social work34). 33 https://www.london.gov.uk/what-we-do/mayors-office-policing-and-crime-mopac/data-and- statistics/taking-part-mopacs-surveys 34 See, for example, Cutliffe et al (2018); Farhat (2016); Vandette and Gosselin (2019). 37 Finally, there was a sense that whilst police requests for victim technology were more likely to follow an explicit reference to the possible existence of such evidence, this was in contrast to the CPS, who were seen to be frequently asking for more speculative phone searches. Indeed, officers felt that, whilst police focused on reasons why a case should go to court, the CPS predominantly focused on the reasons for why it should not - a situation that, in the eyes of the officers interviewed, had exacerbated in recent years. Some of this sentiment is reflected in the views of the survivors who responded to the survey: most had not seen their cases progress to trial and there was a strong sense that cases were dropped despite strong evidence. The current research is not able to add to this the perspective of the CPS and those working within it - something that is needed to be able to give a more comprehensive picture. The findings certainly point to the importance of effective partnership working between these two agencies. 4.2 Future Research The current review identifies three key areas for further research: Firstly, the review focuses on the police investigation stage following an allegation of rape. To gain a more complete picture, there is a clear need to track cases beyond police systems and incorporate data as well as feedback from partners, including and particularly the CPS. This should include further detail in relation to outcomes, what constitutes ‘good’ case file quality, police-CPS correspondence and decision-making, including that in relation to victim and suspect technology evidence, what material is disclosed, and how disclosed material is being used in court / during trial. Information from other partners such as ISVAs / support agencies should also be sought, with multi-agency forums being one potential route to such data. Secondly, in addition to an over-representation of both Black victims and suspects - a finding that has been consistent across different reviews - the current review identified a relationship between BAME suspects and a significantly increased likelihood of victim technology requests. The complexities of these relationships demand exploration beyond the scope of this study but sits comfortably within the framework of the Mayoral Action Plan into transparency, accountability and trust in policing, published November 201935, specifically established to address issues of disproportionality and the use of police powers for Black Londoners. Future research and reviews on this nature should continue to monitor and understand these disproportionalities. Finally, given the consistent finding that the majority of victim/survivors who come to the police to report their rape subsequently withdraw - often within days of the report - there is scope for further research to explore withdrawal reasons in more depth, possibly alongside research to better understand decision-making to report and victim needs and expectations at this point in order to inform support provisions and explore alternatives to criminal justice outcomes. 35 https://www.london.gov.uk/sites/default/files/action_plan_- _transparency_accountability_and_trust_in_policing.pdf 38 4.3 Conclusion The current review finds much unchanged in terms of the profile of reported rape and the picture of attrition in London, as well as the key factors affecting case outcomes. Additionally, it is able to contribute evidence-based insights to the debate around victim technology evidence and disclosure requests. It is hoped the findings from this review can encourage further discussions about how experiences and outcomes for victims/survivors can be improved and how the different criminal justice agencies and wider partners can come together to facilitate such improvements. 39 Appendices Appendix A: Variables included in the coding framework 1. Initial Classification 2. Current Classification 3. Victim age at time of reporting 4. Victim age at time of offence 5. Victim Sex 6. Victim Ethnicity 7. Victim requires interpreter? 8. Victim mental health issues 9. Victim learning disability 10. Victim physical disability 11. Victim is a sex worker? 12. Victim is a missing person or homeless? 13. Victim previous victimisation in relation to sexual assault and / or domestic abuse 14. Victim previous victimisation - any other crime 15. Time between offence and report 16. Offence Location 17. Victim had been drinking 18. Victim had been taking drugs 19. Victim believes that they were drugged 20. Evidence to show that offence was drug facilitated 21. Perpetrator had been drinking 22. Perpetrator had been taking drugs 23. DV Flag 24. Offence reported as part of DASH / 124D questions 25. Witnesses 26. Offence Recorded/ Photographed 27. Mix of consensual and non-consensual sex: This occasion 28. Victim verbal resistance 29. Victim physical resistance 30. Weapon used 31. Weapon type 32. Victim Injury Level 33. Multiple Perpetrators 34. Number of Perpetrators 35. Perpetrator age at time of reporting 36. Perpetrator age at time of offence 37. Perpetrator deceased 38. Perpetrator Sex 39. Perpetrator Ethnicity 40. Perpetrator requires interpreter? 41. Victim - Perpetrator relationship 42. Perpetrator mental health issue 43. Perpetrator learning disability 40 44. Perpetrator physical disability 45. Perpetrator previous history of domestic abuse 46. Perpetrator previous history of sexual assault / rape 47. Perpetrator previous history of any other types of offending 48. How reported 49. Report to Police made by 50. Third Party Description 51. Body Worn Video (BWV) 52. Havens Attendance 53. EEK (Early Evidence Kit) administered 54. Officers record that there were no forensic opportunities 55. Early Investigative Advice (EIA) sought from CPS 56. Victim referred to and/or receives ISVA and/or RCC support 57. Video Recorded Interview (VRI) completed? 58. If no, was VRI arranged? 59. Date of VRI, if applicable 60. Time to VRI (days) 61. Police identify perpetrator 62. Perpetrator arrested 63. Perpetrator interviewed under caution + 3 64. Inconsistencies in victim account are noted 65. Victim unsure if offence took place 66. Victim unsure whether they gave consent 67. Perpetrator denies intercourse/ sexual contact 68. Perpetrator claims consent 69. Is this CRIS linked to others? 70. Reference to existence of tech evidence is made (e.g. social media interaction, phone calls, emails, messages exchanged etc.) 71. Police request victim technology 72. If so, which technology is requested: Phone? 73. If so, which technology is requested: Computer / tablet? 74. If so, which technology is requested: Other? 75. Police seize victim technology? 76. If so, which technology is seized: Phone? 77. If so, which technology is seized: Computer / tablet? 78. If so, which technology is seized: Other? 79. Victim refuses / declines to hand over technology 80. If so, which technology is refused: Phone? 81. If so, which technology is refused: Computer / tablet? 82. If so, which technology is refused: Other? 83. Reasons given for victim refusal to hand over tech 84. Victim agrees to handover tech, but expresses concerns / doubts / asks for reassurance 85. If yes, Briefly outline victim concerns. 86. Date technology seized 87. Date technology returned to victim 88. Time between tech seized and returned 89. Time from report to tech seized 90. Was material found within the device(s) that could aid the prosecution case? 41 91. Was material found within the device(s) that could undermine the prosecution case or support the defence case? 92. Was perpetrator technology seized and/or downloaded? 93. Was material found within the device(s) that could aid the prosecution case? 94. Was material found within the device(s) that could undermine the prosecution case or support the defence case? 95. Either victim or perpetrator tech seized (or both) 96. Indication police struggle with Tech materials (capacity etc.) 97. Do police access/acquire digital material other than via a technology seizure or third-party material requests? 98. If yes, was this publicly accessible social media? 99. Please briefly describe: 100. Was digital material found that could aid the prosecution case? 101. Was digital material found that could undermine the prosecution case or support the defence case? 102. Did police request any third-party material? 103. If so, what third party material is requested: victim medical records? 104. If so, what third party material is requested: victim social service records? 105. If so, what third party material is requested: victim school or education records? 106. If so, what third party material is requested: any other victim records? 107. If 'yes' to last question, what were they? (specify below) 108. Did victim refuse consent for access to any of these records? 109. If 'yes' to last question, for what reason? 110. Victim consents to third party material request, but expresses concerns / doubts / asks for reassurance 111. Was material found in any of these records that could aid the prosecution case? 112. Was material found in any of these records that could undermine the prosecution case or support the defence case? 113. Date material requested from third party 114. Date material sent to police by third party 115. Time between third party material requested and received 116. Police references to follow ups / chasing / lack of progress in third party material being shared 117. Date of CPS submission 118. Time to CPS submission 119. Did police receive a CPS action plan? 120. Date of CPS charging decision 121. Time to CPS charging decision (from reported date) 122. Police reference to high caseload or workload 123. Police express doubt / negative views of victim credibility/ reliability 124. Police express doubt / negative views of how CPS / jury / judge may view victim credibility/ reliability 125. Victim withdrawal 126. Withdrawal statement completed 127. Primary withdrawal reason 128. Secondary withdrawal reason 129. Date of withdrawal, if applicable 130. Time to withdrawal (days) 42 131. Police NFA (where victim does not withdraw) 132. Date of police NFA, if applicable 133. Time to police NFA 134. Case submitted to the CPS 135. Case charged by the CPS 136. Case is still ongoing 137. Reaches trial 138. Trial Start Date 139. Trial End date 140. Conviction Appendix B: Survivor Survey Questions Please note the following questions were asked as part of an online survey and were routed depending on answers given by the respondent. 1. In what year did you report the rape that happened to you to the police? 2. Thinking about the police response and how they dealt with you, tell us how these statements reflect your experiences...  I felt the police responded quickly enough  I felt supported by the police  I felt the police believed what I said  I felt the police had my best interests at heart  I felt I could make decisions about what I wanted to do  Speaking to the police was a good decision  I felt safer reporting to the police 3. Electronic devices, digital communications and social media are becoming more prominent in police investigations of sexual offences. To what extent do you support or oppose the use of such technology evidence as part of police investigations or court cases? 4. In the case you reported, did the police make any reference to seeing or examining your phone (or laptop, tablet, or other device)?  Yes, they suggested this may happen at some stage, but no formal request was ever made (routed to question 4b below)  Yes, there was a formal request to take my phone (or laptop, tablet, similar device) for examination (routed to question 5)  No (routed to question 4c below)  Can’t remember (routed to question 12)  4b. Thinking about how you felt when the police suggested that you may be asked to give your phone or other device, to what extent do you agree / disagree with the following statements:  I felt that giving over my phone would be intrusive 43  I worried giving over my phone would cause practical difficulties  I felt pressured to hand over my device  I would have agreed to hand over my device to the police if asked (routed at this stage to question 12)  4c. If the police had asked you to give your phone or other device, to what extent do you agree / disagree with the following statements:  If this had been asked of me, I would have found it intrusive  Giving over my phone would have caused practical difficulties  I would have agreed to hand over my device to the police if asked (routed at this stage to question 12) 5. At what stage in the process did the police request to examine your phone or other device?  When I first reported the offence  When the specialist SOIT officer first contacted me  Later on in the process  I don’t remember 6. Did the police explain why they wanted to examine your phone or other device?  Yes (routed to question 6a)  No (routed to question 7)  Can’t remember (routed to question 7) o 6a Were you satisfied with this explanation  Yes (routed to question 7)  No (routed to question 6b)  6b Could you tell us a little about why you were not satisfied with this explanation o Free text o (then routed to question 7) 7. Thinking about how you felt when the police requested your phone or other device for examination, to what extent do you agree / disagree with the following statements:  I felt that giving over my phone would be intrusive  I worried giving over my phone would cause practical difficulties  I felt pressured to hand over my device  I thought it was a sensible course of action 8. Did you agree to give your phone or other device to the police for examination?  Yes (routed to question 9)  No (routed to question 8a)  8a What were your reasons for not doing so? Free text response 44 9. How long did the police keep your device?  Up to a week  Up to a month  Up to 6 months  Up to a year  More than a year  It has not yet been returned  I can’t remember 10. What was the impact on you of your device being kept?  No impact  Minor impact  Major impact 10a Please tell us more, if you wish, about the impact this had? Free text 11. Did the police tell you what material they had retrieved from the device?  Yes (routed to 11a, 11b, 11c and 11d below)  No (routed to 12)  Don’t know (routed to 12) o 11a. Did you agree that this material was relevant?  Yes, all of it  Yes, some of it  No  I don’t know / I can’t remember o 11b. Were you happy for the material to be used as part of your case?  Yes  Mixed  No o 11c. Was there any material that the police or CPS said they were going to disclose to the defence (the lawyers representing the accused)?  Yes (route to question 11d)  No  Don’t know  (routed at this point to question 12) o 11d. Please use this space, if you wish, to tell us more about how you felt about material being disclosed to the defence? Free text 12. What was the outcome of the case that you reported to the police?  I withdrew from the police investigation or the court case (routed to 12a and 12b below)  The police and / or CPS said that they could take no further action  The case went to court and the accused was convicted  The case went to court, but the accused was not convicted 45 o 12a. Was your decision to withdraw at all influenced by the possibility of being asked for, or actually being asked, for your phone or device?  Yes  No  Rather not say o 12b. Please use this space, if you wish, to tell us more about the reasons for your withdrawal decision.  Free text 13. Is there anything else you would like to tell us about your experience with the police overall (including handing over your device if that is relevant)? Appendix C: Officer Interview Schedule General perceptions of the role 1. Just briefly, can you tell me why you moved into your current role and how you have found it? o How long have you worked as a SOIT /OIC in rape investigation? Officer perspective of conducting rape investigations and of providing victim care 2. What would you say are the main barriers, and enablers, to carrying out your role? o What works well? o How do you deal with challenges? o What are the main partners and how do you work with them? 3. What are the key victim support needs you deal with and how do you identify them? o Do you feel enabled / able to meet those needs? o What is particularly challenging? 4. What makes a ‘good�� case outcome? o What do victims indicate would be the outcome they want? o What do you, personally, feel makes for a good outcome? o How do you deal with victim withdrawal? Working with victim technology 5. What is your experience of working with technology evidence and requesting victim technology? o What do you see as the pros and cons of technology evidence? o (On balance, does it provide more investigative options, or has it made things more difficult?) 46 o What (if anything) has changed since early 2018? (prompt, if necessary: after the National Disclosure Improvement Plan and the Liam Allan case). 6. How do you decide whether to request a phone or other device from a victim for evidential examination? o In your experience, do officers explain to victims at the outset, or early on, that at some stage their phones may be requested – or only when it is established that there may be potential evidence on the phone. o (Any changes since early 2018) 7. How do you communicate this to the victim? o How do you explain the reason for request and how the material may be used? o Have you encountered concerns by victims over technology requests and the use of technology evidence? o How do you deal with these concerns? o How commonly do you encounter such concerns? o Where, if anywhere, do you record the interactions with the victim regarding these matters (CRIS dets, SOIT log, elsewhere?) 8. How easy is it to identify and then to access and download relevant material om a victim’s phone or device? o What are the key challenges? o What helps? o (tech magnet?) o What is the impact on timescales? On the investigation? On the relationship with the victim? o How do you work with the CPS regarding technology material? Has this changed since 2018? 9. Do you have anything else that you would like to say about your experience of working with technology evidence and requesting victims’ devices? Officer welfare and support needs 10. How would you describe the impact this role has had on you? o In terms of stress levels; in terms of emotional wellbeing o What impacts you the most? o What helps? 11. Would you say you have what you need to be able to do your job well? o Have you had the training you feel you need? Any gaps? o Do you feel able to manage your workload / caseload? o Resources / equipment 47 o In terms of ongoing support o Is there more that could be done to support you? Final reflections 12. Is there anything else you would like to add? 48 Appendix D: Case Characteristics predicting victim withdrawal and police NFA Note: model fit was assessed using Nagelkerke pseudo R squared. The closer the figure is to one, the better the model fit. Model fit can only be compared across models with the same outcome, not between different outcomes. Odds Ratios reported for case characteristic level models. Variables in bold remained significant in overall model. Victim/Survivor Characteristics Victim aged under 18 (report) x2 less Victim aged over 30 (offence) x4 more Model Fit Suspect Characteristics Intimate relationship x2 more Any offending history x3 less Mental Heath Issues x6 less Model Fit Offence Characteristics Location unknown x3 more Witnesses x3 less Reported via DASH x3 more Injury sustained x3 less Model Fit Procedural Characteristics Video Recorded Interview completed x11 less Inconsistent victim account x12 more Police encounter technical difficulties with victim/suspect technology x5 less 3rd Party Material requested x13 less Police Doubt Victim Credibility x3 less Early advice from CPS x12 less Suspect Arrested x12 less Model Fit Full Model Fit 0.09 0.12 0.42 0.63 0.42 0.68 Victim Withdrawal Police NFA 0.03 0.12 0.03 0.12 49 Appendix E: Case Characteristics predicting police technology request Police Request Victim Technology Victim/Survivor Characteristics Victim aged over 30 (report) x2 less Model Fit 0.04 Suspect Characteristics Suspect aged over 30 (report) x2 less BAME x2 more Intimate Relationship x2 less Model Fit 0.14 Offence Characteristics Domestic Violence flag x3 less Witnesses x3 more Victim resisted (physical or verbal) x3 more Model Fit 0.2 Procedural Characteristics ISVA/RCC referral made x2 more Video Recorded Interview completed x3 more Suspect technology seized x6 more Other digital material seized or accessed x3 more 3rd Party Material requested x2 more Model Fit 0.45 Full Model Fit 0.5 50 References Angiolini, E. (2015). Report of the independent review into the investigation and prosecution of rape in London. https://www.cps.gov.uk/publication/report-independent-review- investigation-and-prosecution-rape-london-rt-hon-dame-elish Cutliffe, J., Sloan, G. and Bashaw, M. (2018). A systematic review of clinical supervision evaluation studies in nursing. International Journal of Mental Health Nursing. 27(5), 1344- 1363. Farhat, A. (2016). Professional development through clinical supervision. Education. 136(4), 421-436. HMCPSI and HMIC (2017). Making it fair: A joint inspection of the disclosure of unused material in volume crown court cases. https://www.justiceinspectorates.gov.uk/cjji/wp- content/uploads/sites/2/2017/07/CJJI_DSC_thm_July17_rpt.pdf Horwell QC, R (2017) The Mouncher Investigation Report. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_ data/file/629725/mouncher_report_web_accessible_july_2017.pdf MOPAC (2019) The London Rape Review: a review of cases from 2016. https://www.london.gov.uk/sites/default/files/london_rape_review_final_report_31.7.19.p df Vandette, M.P. and Gosselin, J. (2019). Conceptual models of clinical supervision across professions: A scoping review of the professional psychology, social work, nursing, and medicine literature in Canada. Canadian Psychology, Vol 60(4), pp. 302-314. Review of the Metropolitan Police Service Gangs Matrix December 2018 2 This Review, carried out by officers in the Mayor’s Office for Policing and Crime (MOPAC), fulfils the commitment in the Police and Crime Plan 2017-2021 to carry out a Review of the Gangs Matrix of the Metropolitan Police Service (MPS), an intelligence tool to supplement enforcement and diversion action against street-focused violence. This Review is to be seen the context of policies and priorities set out in the Mayor of London’s Police and Crime Plan 2017-2021 and the accompanying Knife Crime Strategy 2017-2021. This approach to reducing violence is further enhanced by the Mayor of London’s announcement to establish a Violence Reduction Unit to develop London’s approach to tackling the long-term causes and solutions to violence. An Integrated Impact Assessment accompanies this Review to assess the range of impacts of the Matrix, not least the equalities and data privacy impacts which are particularly significant. At all stages in carrying out this Review, MOPAC has been mindful of its obligations under the Human Rights Act 1998. The statistical analysis and research was conducted independently by analysts in MOPAC’s Evidence and Insight team. The methodology was reviewed by University College London and further assessed by other academic experts. MOPAC organised the Reference Group representing interested organisations and communities, as well as the engagement activities with communities and young people, some of whom had been on the Matrix. Independent legal advice was provided by Tim Pitt-Payne QC from 11KBW who had access to relevant Metropolitan Police Service officers and MOPAC staff and relevant data and documentation. MOPAC is grateful to those who supported the analysis, agreed to be surveyed and participated in the Reference Group, including the Metropolitan Police Service for freely sharing information; and officers from the Information Commissioner’s Office who worked closely with MOPAC both on this Review and their parallel enforcement activity, culminating in their recent issuing of an enforcement notice and the Metropolitan Police Service’s action plan in response. The support of all participants has been instrumental in enabling us to conduct the most authoritative and far-reaching review of the MPS Gangs Matrix ever undertaken. December 2018 3 Contents 4 - Executive summary 7 - Recommendations 10 - Introduction: context of the Review; definition of “gang” 15 - Terms of Reference 19 - Impact of the Gangs Matrix: - How the Gangs Matrix operates; - Relationship between the Gangs Matrix and criminality; - Analysis of the impact of being on the Gangs Matrix to individuals’ future criminal offending and victimisation; - Practitioner perceptions of impact. 41 - Gangs Matrix Processes: - Legal considerations under human rights and data legislation, with associated recommendations for improvement; - In practice: how the Matrix is used at borough level; - Practitioner perspective: aims and purpose of the Matrix; - Practitioner perspective: addition and removal from the Matrix; - Training on the Gangs Matrix and related areas; - Oversight and governance. 56 - Perceptions and understanding of the Gangs Matrix - London level disproportionality; - Do the Matrix individuals come from the most violent areas? - Why does disproportionality occur? - Practitioner views of disproportionality; - Transparency; - The community perspective; - Views of the Reference Group: summary of meetings. Appendix 1: Legal opinion Appendix 2: Reference Group members Appendix 3: Our Equalities Duty as a public body Appendix 4: Matrix Review methodology Appendix 5: Bibliography 4 Executive Summary This is the Review of an enforcement database operated by the Metropolitan Police Service known as the Gangs Matrix. The Review was a specific commitment in the Mayor’s Police and Crime Plan and also in his 2016 election manifesto. The Review, the most comprehensive ever conducted into the MPS Gangs Matrix, combines analysis of those on the Matrix; interviews with practitioners and those in communities affected by the Matrix; and detailed discussions with a Reference Group of practitioners and subject matter experts. The Review also takes note of the recent investigation by the Information Commissioner’s Office (ICO) on the Matrix. This Review is published at a time when tackling serious violence – which has been increasing across the country since 2014 - continues to be the priority issue in the city for the MPS and for the Mayor. Gangs are a driving force behind some of the most serious violence in London. While gang-flagged violence accounts for a relatively small proportion of overall levels of violent crime in London, it represents a significant percentage of the most serious and harmful offending and victimisation. 57% of gang related stabbings featured a serious or fatal injury, compared to 34% of non-gang-flagged stabbings. The harm of gangs extends further, beyond serious street violence and encompassing other serious issues including violence against women and girls, acquisitive crime and drug supply. Those preyed upon by gangs are amongst the most vulnerable children and young people in our city, often from deprived, crime-affected backgrounds and presenting multiple, complex needs such as mental illness and special educational needs. This Review recommends that there is a comprehensive overhaul of the Matrix Operating Model both to restore trust in the Matrix and also to bring it into line with data protection legislation. This overhaul must be completed by the 31st December 2019. The Review does recognise that a Gangs Matrix is a necessary law enforcement tool for reducing violent crime in London. However, the Review also identifies that the representation of young, black males on the Matrix is disproportionate to their likelihood of criminality and victimisation and recognises that communities in London have deep reservations about how the Matrix operates. The Review found significant issues around public understanding of the Matrix and a lack of transparency on the part of the Metropolitan Police Service (MPS) in communicating the aims and purpose of the Matrix – not only to the public but also to practitioners. Furthermore, an investigation by the Information Commissioner’s Office (ICO) found that, whilst there was a valid purpose for the MPS’ use of the Gangs Matrix, inconsistency in the way it was used led to multiple and serious breaches of data protection laws. 5 The Review also recommends that MOPAC and the MPS engage with the Equality and Human Rights Commission as we work to ensure that the Gangs Matrix operates- and is seen to operate - in a way which upholds the human rights of those included on it, and without any element of discrimination. There are further recommendations aimed at improving the operation of the Gangs Matrix as well as building trust in the Matrix among the public and practitioners. Structure of the report and summary of findings The document begins by setting out the wider context for the Review, including and how and why the concept of a ‘gang’ forms the basis of police prioritisation, and the Terms of Reference for the Review (page 10). The Review finds that over the period of analysis, individuals spent an average of 28.5 months on the Matrix (including approximately a third of that time in custody). It would appear that the Matrix does have a positive impact on reducing levels of offending by and victimisation of the individuals included on it – and that the reductions in these risks are sustained after they have been removed from the Matrix. However, limitations on the data available from partner agencies mean it is not possible to identify the specific reasons for this impact. (page 19). Furthermore, analysis of timeliness of arrests and stop and search indicates that levels of police attention experienced by individuals on the Matrix fall to pre- inclusion levels after they have been removed – potentially allaying some concerns about Matrix inclusion ‘following’ individuals after their removal (page 33). The Review raises questions about individuals on the Matrix with low or zero-harm scores - whether their inclusion on the Matrix can be justified, and whether the emphasis of interventions with these persons should be safeguarding rather than enforcement (page 30). Legal advice sought as part of this Review finds that the Matrix can be used lawfully as an enforcement tool. Through placing people on the Matrix, the MPS aims to identify those at risk of committing acts of violence or being victims of violence, so that appropriate enforcement or diversion action1 can be taken (page 41). However, the Review, in light of the legal advice, also finds that there are a number of actions that the MPS must undertake to be fully compliant with the law, including the production of an Equalities Impact Assessment on the Gangs Matrix. Furthermore, the Review finds a number of issues with the processes, practical application and data handling of the Matrix also identified by the Information Commissioner’s office, with whom MOPAC has worked closely through the Review, and which are the subject of a recent enforcement notice (page 44). The data is clear that gang-related violence is a source of tremendous harm; that it disproportionately affects some groups of Londoners more than others, particularly young 1 Such as mentoring from specialist youth workers or treatment for physical and mental health needs 6 black men; and that these young men are more likely to be the victim of gang-related violence, as they are the perpetrator. The Review goes on to analyse the demographic makeup of those on the Matrix and confirms that there is a disproportionate representation of black and minority ethnic Londoners relative to their offending and victimisation (page 56). The Review also highlights significant concerns about the Matrix expressed by young black men and many others in communities most affected by gang-related offending (page 65). In addition to MOPAC’s oversight of the MPS in this space, we will engage with the Equality and Human Rights Commission as we work to deliver the recommendations of this Review, providing further independent assurance that equality remains at the heart of this work; and the MPS will further assess the equalities impact of the Matrix and produce an Equalities Impact Assessment. In light of these findings, the Review does not propose that the MPS ceases to use the Matrix, but that the MPS must take action to comprehensively overhaul it through implementation of these recommendations and those of the ICO, to ensure that the Matrix is used lawfully and with no unjustifiable disproportionality, with stronger processes and oversight, and with greater transparency. 7 Recommendations This review sets out a series of recommendations to the MPS, all of which must be completed by the 31st December 2019: Recommendation: We recommend that the MPS comprehensively overhaul the Matrix Operating Model and review the Model annually. We further recommend that all officers using the Model receive training on how to apply the guidance and to do so consistently across all boroughs. Both the Operating Model and the training should have a particular focus on ensuring: • that the right people are on the Matrix; • that people are added and removed in a standardised, evidence-based manner; • that they can be removed and that the ‘gang’ label will not ‘follow’ them; • that local Matrices are refreshed regularly so that individuals don’t stay on any longer than necessary; • that the guidance on the use of social media for intelligence purposes is updated; and • that the data protection principles and legislation are fully applied. Recommendation: We recommend a thorough reappraisal of the individuals in the Green category, with a focus on: those that currently score ‘zero-harm’; those that have never had a harm score or have remained in the Green category for their entire time on the Matrix; and those under the age of 18. This reappraisal should begin as soon as possible and be concluded no later than 31st December 2019. This reappraisal should consider whether: the level of risk they present justifies their continued inclusion; their inclusion is consistent with the published purpose of the Matrix; and whether their inclusion is compatible with Article 8 (2) of the Human Rights Act 1998. Where an individual does not meet these criteria, they should be removed from the Matrix. Recommendation: That MOPAC and the MPS engage with the Equality and Human Rights Commission as we deliver the recommendations in this Review and the ongoing work of the Gangs Matrix, supporting the MPS’ work to further assess issues around human rights, disproportionality and produce an Equalities Impact Assessment. Recommendation: We recommend that the MPS strengthen their governance of the Matrix and the officers and partners that use it, creating single points of responsibility on each Borough Command Unit to: • ensure there is no discriminatory practice; 8 • ensure that risks around data breaches are properly assessed and mitigated; • track progress against recommendations published in this report; • review the intelligence processes and flows that create organisational knowledge around gangs; • assess the Matrix in the light of technological advances to future proof it; • oversee the continuation and expansion of regular borough audits; and • work with MOPAC to conduct new analysis exploring issues relating to disproportionality and the Matrix (e.g. micro-level demographics analysis in gang affected areas). Recommendation: That the MPS improves systematic data capture across all aspects of the Matrix process. To include but not limited to: • demographics of Matrix individuals (gender, age and ethnicity); • nature and extent of police activity for those on the Matrix; and • nature and extent of non-enforcement interventions (needs, referrals, uptake and outcomes). We further recommend that MOPAC and the MPS conduct an annual review of the Matrix population, in comparison with the wider London gang and violent offending profiles. MOPAC will also convene partners to discuss options for enabling better collection of data on the impact of non-criminal justice interventions. Recommendation: We recommend that the MPS urgently improve their current Matrix processes to ensure that personal data and information are stored, managed, shared, protected and transmitted safely and appropriately, with particular reference to: • ensuring compliance with data protection principles and legislation, including the completion of a Data Protection Impact Assessment; • who the Matrix information is shared with outside of the MPS; • ensuring that any sharing of personal information is necessary and proportionate; and • implementing recommendations set out in the ICO enforcement notice. Recommendation: We recommend that the MPS improve transparency by producing, by the end of February 2019, publicly available, plain English and accessible information that: • answers frequently asked questions about how the Matrix works and its purpose; • sets out the governance and oversight mechanisms the Matrix is subject to; • explains the training officers receive that helps them implement the Operating Model; • describes the practical effect for an individual of being on the Matrix; 9 • clarifies how information is shared and who with; • explains how an individual is added to and removed from the Matrix; • describes how the multi-agency process works; and • describes the relationship between the Matrix and the response to serious organised crime. Recommendation: MOPAC will oversee implementation of the recommendations in this Review - and transparency in this work - by: • requiring the MPS to report annually on progress against recommendations to the MPS Ethics and Audit panels and publishing those reports; and • ensuring progress against recommendations are reported to the MOPAC / MPS Oversight Board. Recommendation: The MPS should consider whether the lessons learned from this Review are applicable to the operation of, and sharing of information under, other operational tools. 10 Introduction - about this Review Mayor of London Sadiq Khan made a commitment to Londoners in his manifesto and in his Police and Crime Plan to conduct this Review of the MPS Gangs Matrix. The Metropolitan Police Service (MPS) defines a ‘gang’ as a relatively durable, predominantly street-based group of young people who: - see themselves (and are seen by others) as a discernible group; and - engage in a range of criminal activity and violence. They may also have any or all of the following features: - identify with or lay claim over territory; - have some form of identifying structure feature; and - are in conflict with other, similar gangs. This is the definition used by government and was defined in the Centre for Social Justice’s 2009 report ‘Dying to Belong’. However, we note from our engagement with practitioners as part of this Review that nonetheless, the definition of a gang is still the subject of debate. Gang-flagged violence accounts for a relatively small proportion of overall levels of violent crime in London but represents a significant percentage of the most serious and harmful offending and victimisation as the table below illustrates. Gang-related violence is significantly more likely to result in serious injury; 57% of gang-related stabbings featured a serious or fatal injury, compared to 34% of non-gang-flagged stabbings. 2 Table 1: Proportion of serious violence identified as gang related3 2 Internal MOPAC analysis based on 2017 data. Previous analysis from 2015 also found gang-related stabbings to result in more serious injury. 3 Homicide and knife injury statistics are victim counts. Both firearm categories are offence counts. Data shown as 2018 includes offences up to 30/09/2018. 2016 2017 2018* Total All Homicides 112 136 106 354 Gang related 29% 27% 37% 31% Homicide excluding Domestic Abuse & Terrorism 95 115 83 293 Gang related 35% 32% 47% 37% Lethal Barrelled Discharge 334 354 313 1001 Gang related 50% 40% 52% 47% Lethal Barrelled Discharge - Victim Shot 136 144 112 392 Gang related 62% 49% 63% 58% Knife Injury; Victim under 25; excl. Domestic Abuse 1853 2138 1433 5424 Gang related 26% 21% 18% 22% Homicide Shootings Stabbings Crime Type 11 Clearly, this is an issue that demands a police response, one which recognises the unique characteristics of gang violence, enables an understanding of who is at highest risk of involvement in it and can bring to bear the wide range of interventions – from prevention to enforcement – necessary to tackle it. In the aftermath of the serious disorder in the streets of London in 2011, and in recognition of the severity of gang-related offending, in 2012 the Metropolitan Police Service (MPS) created the Gangs Matrix - a tool used to identify and risk-assess the most harmful gang members in a borough, based on violence and weapons offences, and informed by intelligence. This basic concept – collating and using data about individuals known to be involved in criminality or at risk of doing so, is deemed by the police to be a fundamental part of policing and public safety, whether that be a large-scale tool such as the Police National Computer (PNC), down to small-scale, local measures such as the Gangs Matrix (in context, the Matrix has generally included between 3-4,000 individuals at any given time, from a London population in excess of 8,000,000). The Matrix is an important part of the MPS and partner agencies’ activities to reduce gang-related violence and prevent young lives being lost, by taking enforcement action against the most violent gang members and seeking to divert those most at risk of being drawn into gang violence. Once on the Matrix, an individual may be subject to more intensive policing activity or interventions by partner organisations4 to reduce their risk of offending or victimisation. From its inception, the Gangs Matrix has been a controversial issue within the wider - and already often contentious - debate around policing and ethnicity in London. Most recently, reports by David Lammy MP, Amnesty and StopWatch have been highly critical of the 4 Such as Social Services, the London Gang Exit Service or Mental Health services. Identifying gang-related incidents Identifying gang-related incidents is challenging. This is particularly true of incidents where no suspect is identified or where the victims or witnesses are unwilling to co-operate with the police investigation. In line with internal MPS procedures, gang-related incidents were identified using a combination of gang flags on CRIS (Crime Reporting Information System) and offences included on GRITS (Gang Related Incident Tracking System). CRIS gang-flagged reports are defined as 'any gang- related crimes or crime-related incidents where any individual believes that there is a link to the activities of a gang or gangs'. GRITS is an MPS Intelligence tool used to identify the most active gangs in London. Offences in GRITS are defined as any offence where the suspect or victim are shown through intelligence to be a member of a gang or where the motivation of the offence is linked to gang tensions and rivalries. The combination is used due to the ‘gang’ flag not being consistently applied across the MPS. 12 Matrix, accusing the MPS of disproportionality and discrimination, particularly against young black males. In addition, the Information Commissioner’s Office (ICO) has served an enforcement notice against the MPS after an investigation found multiple and serious breaches of data protection law in the use of the Matrix. The MPS have defended the value of the Matrix as a policing tool, arguing that it assists them in preventing crime and protecting some of the most vulnerable people in the city from coming to harm. London continues to face a significant challenge from violent offending and the rises in gun and knife crime over the past two years highlight this very real concern. The MPS know from evidence, intelligence and specific cases that gangs play a significant role in this violence picture, which brings misery and fear to communities across London. They argue that the Matrix has been an essential operational tool to help identify those that have been - or are likely to be - involved in violence and facilitate a range of police and partnership action from enforcement through to the vital prevention work that occurs at a local level. Both sides of this debate are important, and both are persuasive. The data is clear that gang-related violence is a source of tremendous harm; that it disproportionately affects some groups of Londoners more than others, particularly young black males; and that these boys and young men are more likely to be the victim of gang- related violence, as they are more likely to be the perpetrator. It is logical that an understanding of who is involved in gangs or who is on the periphery of involvement is a key asset in reducing the risk of victimisation and offending. The evidence revealed by this Review shows that the Matrix does appear to have a positive impact in this regard, but also that data limitations mean it is not possible to identify the reasons for that impact. Equally, it is fundamental that the police use the powers granted to them by the public in a proportionate way that upholds the rights and privacy of the citizen – and that citizens can see and have confidence that this is the case. This is at the very core of policing by consent. This Review shows that young, black African-Caribbean men are disproportionately represented on the Gangs Matrix. It has also found significant issues around public understanding of the Matrix and a lack of transparency on the part of the MPS in communicating the aims and purpose of the Matrix – not only to the public but also to practitioners. There is also a question of process. The Matrix is a significant database of personal information on several thousand Londoners. It is essential that this data is controlled and processed in a lawful, secure way. This Review has found examples of good practice, but it has also found inconsistencies and gaps in the management, processes and oversight of the 13 Gangs Matrix that will be addressed, including responding to recommendations from the Information Commissioner. It is the nature of policing by consent that it operates in a state of constant – and necessary – tension between securing public safety and protecting civil rights and liberties. That tension is brought sharply into focus in this Review. That is why, even though its structures have changed over the years, there has always been a thread of democratic oversight over policing, ensuring that the public’s wishes and expectations are reflected in the policing they consent to and pay for. Gang databases and official processes for gang member identification have received relatively little academic or public scrutiny in the UK. Only recently has research begun to explore this area, with studies highlighting the range of factors influencing the process and the differing or competing perceptions, priorities and labelling (Medina et al 2009; Smithson et al 2013; Williams 2015). Others have highlighted administrative problems in policing gangs on a borough basis and hinted at the difficulties of centralised mapping of localised problems (Densley & Jones 2013). Building on a growing debate around the policy response to gangs (Densley 2011; Hallsworth 2013; Joseph et al. 2011; Pitts 2017; Shute & Medina 2014), two recent London-focused studies drew on qualitative and secondary data sources to explore how individuals were affected by the Gangs Matrix and other targeted policing activity, with a specific focus on the disproportionate policing of young black males (Amnesty 2018; Williams 2018). In the US, research in this area is far more developed, with mixed findings as to gang database efficacy. The ability to accurately identify gang affiliation has been questioned (Chesney-Lind et al. 1994; McCorkle & Miethe 1998; Spergel 1995), whilst others have suggested that the focus should be on individual and group (as opposed to gang) harm (Kennedy 2009). Further issues have been examined around governance and oversight (Jacobs 2009), including concerns about non-removal (Esbensen et al. 2001; Klein 2009) and punitive criminal justice outcomes (Caudhill et al. 2014). Research has also generated more positive indications, suggesting that differences in severity of offending between gang and non-gang individuals make monitoring worthwhile (Katz et al. 2000). Non-identification carries its own risks and may mean an individual is denied diversionary interventions (Short 2009; Spergel 2009), or at a wider level presents an unacceptable risk to society (Ericson & Haggerty 1997). We have endeavoured to engage with all aspects of the Gangs Matrix, its operations and impacts in London and the different points of view on its usage. In doing so, we have: - engaged with officers and staff of the Metropolitan Police Service up to the highest levels of senior leadership, and with other agencies working to address gangs and 14 gang crime in order to understand their use of the Matrix and their views on its effectiveness5; - conducted an in-depth analysis on the process and population characteristics of the Gangs Matrix since inception, looking at the offending patterns of more than 7,000 individuals on the Matrix, and the patterns of both victimisation and stop and search on over 5,000 individuals in total, exploring issues around disproportionality and potential impact across these areas; - brought together a Reference Group with expertise in this area and/or experience of working with people affected by the Matrix; - conducted direct consultation to hear the voices of young black people, who are disproportionately represented on the Matrix; and - engaged with other interested parties and statutory bodies with perspectives on the Matrix, including Amnesty, Williams and Clarke, StopWatch and the Information Commissioner’s Office (ICO). This is a substantial document, and necessarily so. There are no simple, definitive answers to the questions that the use of the Gangs Matrix approach poses. We have reflected in great depth and taken great caution in developing the recommendations made considering the findings of this research. This is not a theoretical exercise and there is no room for complacency or rashness. We recognise throughout this document that any proposals for change ultimately impact on real interventions with real people in situations of real risk. We are very grateful to everyone who has given their time, experiences and insights as part of the Review process. In particular, we would like to pay tribute to the late Bobby Martin, a tireless supporter of young and Black and Minority Ethnic (BAME) Londoners and a critical friend to the MPS and the Mayor’s Office for Policing and Crime (MOPAC) over many years. He will be greatly missed. 5 Surveys were carried out in Autumn 2017 of police ‘Single Points Of Contact’ who used the Matrix. We received 88 responses across 28 boroughs. Surveys were also carried out of local authority leads working with young people involved in violence. 45 responses were received across 27 boroughs. In Summer 2018, surveys were carried out of voluntary and community sector organisations supporting young people involved in violence. 83 full responses (and 15 partial responses) were received across 21 boroughs. 15 Gangs Matrix Review - Terms of Reference The Mayor’s Office for Policing and Crime (MOPAC) are committed to conducting a Review of the Metropolitan Police Service (MPS) Gangs Matrix. The Review follows the Mayor’s Manifesto pledge in 2016 and commitments in London’s Police and Crime Plan6 and Knife Crime Strategy in 2017. Keeping young people safe in London is a key priority and to that end we want to ensure that any tool being used to manage young people involved in group violence responds appropriately to harm and risk. The Gangs Matrix is an operational tool that was developed by the Metropolitan Police Service and has been used since 2012. The intended audience for the outcome of the review is the MPS and we expect to make our recommendations public by publishing them in Autumn 2018. The review will be conducted by MOPAC, with the support of the MPS. No previous review of the Gangs Matrix has taken place. Background information The overarching aim of the MPS’ Gangs Matrix is to reduce gang related violence and prevent young lives being lost. The Gangs Matrix is an intelligence tool used to identify and risk assess gang members in every London borough informed by intelligence and based on violence and weapons offences. It provides the police with a way of prioritising the most harmful individuals for enforcement and partners for support and intervention. The Matrix scores individuals who are in a gang - it is not the gangs that are scored. Scores are based on an individual’s involvement in violence only - drugs intelligence or drugs offences are not scored. Individuals are also scored and ranked as victims of violence to reflect the vulnerability and victimisation of this cohort. Every borough has their own Matrix and every day these are combined to produce an MPS Matrix. Matrices are owned by boroughs with local Single Points Of Contact (SPOCs) who update the Matrix and make decisions about who is added and removed. Trident and central MPS intelligence7 may make recommendations where necessary for people to be added or removed to any borough Matrix. Following 2011 there was a greater focus on work on gangs, both from the Home Office and the MPS, which resulted in a standardised method of quantifying and prioritising those 6 The Police and Crime Plan states “we will Review the MPS approach to gang crime, including the Gangs Matrix, supporting the MPS to tackle gang crime, gun crime and knife crime more effectively in London.”. The London Knife Crime Strategy says we will target lawbreakers by “Reviewing the MPS Gangs Matrix which identifies the most violent gang members in London and strengthen the identification of young people who are involved in serious youth violence, whether perpetrators or victims.”. 7 Trident is a dedicated Operational Command Unit within the MPS that has responsibility for tackling gang crime and investigating all non-fatal shootings. Met Intelligence relates to the entire MPS intelligence infrastructure. A centralised intelligence team of officers and analysts collect, assesses and interpret a range of information sources, generating analysis to tackle threat, risk and harm in London. 16 involved in serious youth violence. In many areas of London, using the Matrix has meant a more joined-up discussion and targeted response from local police and partners about those involved in gangs and youth violence, co-ordinating not just enforcement to reduce identified risk, but offer diversion, support and intervention too. However, the Matrix has proved to be a controversial tool and is perceived negatively by some communities. Given these concerns it is important to set out the criteria for assessment with the aim of addressing issues of consistency, disproportionality, scope and impact. The police and partner response to gangs and youth violence has evolved significantly with a better understanding of the drivers and needs of those involved as victims or perpetrators. There have been improvements in enforcement tactics, use of legislative tools and services for young people, such as funding for Integrated Gangs Units, London Gang Exit and support for young victims of violence in London’s Major Trauma Centres. The Mayor’s Police and Crime Plan focuses on ‘gang’ violence, vulnerability and exploitation; the specific behaviours that are impacting negatively on our communities – violence and weapon enabled offending – and the exploitation of children and vulnerable adults linked to this – child criminal exploitation, child sexual exploitation and trafficking of young people through ‘county lines’8 drugs activity. David Lammy’s review into the treatment of BAME Londoners in the criminal justice system9 supported the Mayor’s commitment to review the Matrix and asked MOPAC to ‘examine the way information is gathered, verified, stored and shared, with specific reference to BAME disproportionality’ and to consider community perspectives. This Review is part of MOPAC’s response to this. Scope of the review In recognition of these changes, and within the context of increasing knife crime in London, it is therefore timely to review the Matrix to better understand whether the targeted approach taken by the Gangs Matrix is a proportionate and effective tool in reducing youth violence. The Review will focus on the following issues: 1. Impact of the Gangs Matrix • How the Gangs Matrix impacts on levels of gang member offending and youth violence; 8 County lines is a term used to describe gangs and criminal networks involved in exporting illegal drugs in to one or more importing areas [within the UK], using dedicated mobile phone lines or other form of ‘deal line’. They are likely to exploit children and vulnerable adults to move [and store] the drugs and money and they will often use coercion, intimidation, violence (including sexual violence) and weapons. 9 published September 2017 17 • How being on the Gangs Matrix impacts an individual’s involvement in violence and offending; • How the Gangs Matrix has evolved to be able manage risk and vulnerability; • How being on the Gangs Matrix impacts an individual who features on it; from an enforcement perspective; and • How the demographics of those on the Gangs Matrix compare with our understanding of victims and perpetrators of youth and weapon-enabled violence in London to identify any disproportionate impact. 2. Gangs Matrix processes • Assurance that the collection, use, sharing and storage of personal information in the context of the Gangs Matrix is compliant with human rights and data protection legislation; • How individuals are added to and removed from the Gangs Matrix; • Understand which partners have access to Gangs Matrix information and how it is used; • What data sources the Gangs Matrix uses; • How the Gangs Matrix is used locally by the London boroughs; and • How the Gangs Matrix is reviewed and overseen; to assess consistency of use and prevention of disproportionate impact. 3. Perceptions and understanding of the Gangs Matrix • How information about the purpose and use of the Gangs Matrix has been shared and understood by partners, community members and the public; • Assessing whether the Gangs Matrix is understood within the wider context of the other operational tools that inform the threat, harm and risk profile for London; and • Consider what steps can be taken to provide better information to the public about how the MPS tackle youth violence. Guiding principles The recommendations of the Review will be guided by the need to: • Be evidence-led through in-depth analysis of available data; • Recognise the operational independence of the MPS; • Ensure legality and data compliance; • Address any unfairly disproportionate impact; • Be informed by community and partner views; 18 • Increase transparency; and • Be informed by an understanding of the operational challenges faced by police officers responding to youth violence. This Review will not make recommendations related to police intelligence systems as a whole. The Review has been carried out by MOPAC’s Criminal Justice & Commissioning and Evidence & Insight teams. MOPAC facilitated the collection and analysis of professional and community views in relation to the matrix and use these to formulate recommendations. The Deputy Mayor for Policing and Crime is responsible for the completion of this review. The MPS is responsible for deciding what changes are necessary and how to take recommendations forward operationally. Recommendations will be published in Autumn 2018. 19 Impact of the Gangs Matrix 20 This section sets out how the Gangs Matrix operates. The MPS Gangs Matrix Operating Model and Guidance states that the threshold for being included on the Matrix is ‘someone who has been identified as being a member of a gang and this is corroborated by reliable intelligence from more than one source (e.g. police, partner agencies such as local authorities).’ Individuals included on the Gangs Matrix are classified as Red, Amber or Green, depending on how they are scored against a combination of factors. Individuals on the Gangs Matrix will often be themselves at risk of being victims of violent crime but the Red, Amber or Green status of an individual as recorded on the Gangs Matrix is intended to reflect the extent to which that individual poses a risk to others, and not the extent to which the individual is judged themselves to be at risk. The factors considered when scoring an individual on the Gangs Matrix are: • their history of violence in the past three years, taken from the Crime Report Information System (CRIS). Scores are weighted according to the severity of the offence and how recently it was committed, as well as whether the individual is suspected or charged. Victimisation scores are also calculated; • violence or weapons intelligence in the last six months (taken from CRIMINT, a criminal intelligence database); • the judgement of a local gang unit (or equivalent) intelligence manager; and • a partner organisation’s assessment of risk of harm (e.g. from the Youth Offending Service, or an offender manager assessment), if applicable. Weighted scores combine to generate a total ‘harm’ score which determines their harm banding and subsequent levels of police and partner response. A snapshot of the Matrix from September 2018 presents the following breakdown by Harm Banding. The categories reflect the risk of violent offending, with Red being the highest risk and Green the lowest. RAG Live Custody Green 66% 34% Amber 65% 35% Red 49% 51% Total 65% 35% Green 65% Amber 31% Red 4% 21 The Operating Model sets out that the Gangs Matrix should be reviewed quarterly and that individuals remain on the Matrix for no longer than is necessary. Each MPS borough is responsible for reviewing the Gangs Matrix in respect of their local areas to determine whether it is appropriate to retain or remove an individual from the Matrix, and whether any new information linked to the individuals on the Matrix has been taken into account. As a result, the number of individuals on the Matrix changes daily as individuals are reviewed, removed or added. A September 2018 snapshot shows the following breakdown: 3,228 individuals are on the Matrix of which 48% are in either in custody or have a judicial restriction such as a GPS tag or a suspended sentence in place. Individuals on the Matrix are predominantly young – with 74% under the age of 25; 99% are male and 80% are black African-Caribbean10. This demographic has changed little since the Matrix was first made fully operational in 2013. The data on demography is explored further later in this Review. The number of individuals on the Matrix remained relatively stable until August 2017 when the population began to decrease. In August 2018, the number of individuals on the Matrix was at its lowest since July 2013. The reasons for this decrease and the timing of it are 10 According to the MPS Ethnic appearance codes: White proportion is 8.4% White – North European (n=282); White - South European (n=101); BAME proportion is made up of 80% African Caribbean (n=2,696); 5 % Asian – Indian Sub- Continent (n=176), 3% Arabic or North African (n=105) and 0.06% Chinese, Japanese or other South East Asian (n=2). Activity for Red, Amber and Green nominals on the Gangs Matrix Activity against nominals on the Matrix graded as Red, Amber and Green ensures a graded response commensurate with the risk the offender presents and can include: Red individuals Daily activity around these individuals – enforcement, partnership or both. A bespoke, multi-agency partnership plan is put in place to offer pathways to prevent and divert from gang-related activity. Assessment for judicial interventions such as a Criminal Behaviour Order is undertaken. Amber individuals A bespoke, multi-agency partnership plan is put in place to offer pathways to prevent and divert from gang-related activity. Enforcement activities and assessment for judicial interventions are considered. Green individuals A single-agency response plan is put in place and diversion or engagement activities considered with partners. 22 difficult to evidence but may include internal MPS activity to improve practice and continued resource pressures on policing. The factors affecting Matrix population levels form a key part of the following sections. Graph 1: Total Matrix Population by Month Our analysis for this Review explored the population of the Matrix over a five-year period between June 2013 and May 2018. To do this, monthly snapshots of the Matrix were combined to generate a total sample of over 7,000 individuals who appeared on the Matrix at some point during this 60-month period. For further details see the methodology in Appendix 4. Over the period of analysis, individuals spent an average of 28.5 months on the Matrix (including approximately a third of that time in custody). A sizable proportion of the cohort (11%, n=905) appeared on the Matrix continuously across the five-year period, with the majority of these (n=721) also appearing on an early, 2012 iteration. At the other end of the scale, 12.5% (n=1038) of the cohort appeared on the Matrix for five months or less and 390 individuals appeared on the Matrix for just a single month. Taking into account these extremes, the median length of time spent on the Matrix is 25 months. Matrix criminality: overview Our analysis into the offending of individuals on the Matrix was primarily undertaken using records from the Police National Computer (PNC), with emphasis on proven offending, referred to as ‘sanctions’ (offences for which the individual received a conviction, caution or 23 warning).11 Arguably the most robust metric for establishing police recorded offending, it is important to note that this is only one indicator of criminal activity. Offending over lifetime The clear majority of the Matrix cohort have at least one sanction for any offence in their lifetime (96%, n=6,978). Overall, the cohort has been responsible for 98,189 sanctions, an average of 13.45 sanctions each. The cohort is also reasonably criminally versatile; 46% (n=3,389) have sanctions across four or more crime types.12 The majority have been sanctioned for a serious offence in their lifetime (62%, n=4,549)13 Nearly three-quarters of the cohort has at least one sanction for either violence against the person or weapons offences (71%, n=5,160), rising to 78% when robbery is included (n=5,691). Over a third (37%, n=2,735) of the cohort have been convicted of public order offences. 14 77% (n=5,576) of the cohort have been convicted of a drugs offence, a third of which was solely cannabis possession (30%, n=1,672). While the Matrix doesn’t score for drugs offences, drug distribution is central to the business model of many gangs and is often what drives street violence. In this context such findings may be expected, although it is important to note that 40% (n=2,904) of the cohort have at least one sanction for drug supply or trafficking offences. Offending before Matrix inclusion In terms of proven offending before inclusion on the Matrix, 90% (n=6,526) of the cohort have been sanctioned for at least one offence of any kind. Half of the cohort had received at least one sanction for a serious offence (50%, n=3,651) and 67% (n=4,923) had a previous sanction for violence related offences.15 58% (n=4,248) had at least one sanction for a drugs offence; including 23% (n=1,700) for supply and 51% (n=3,722) for possession. Just under a quarter (24%, n=1,770) of the cohort had drugs sanctions which related only to cannabis possession pre-Matrix inclusion. 11 Only 3,468 sanctions relate to non-court convictions (cautions, reprimands or warnings). Fixed Penalty Notices (FPNs), such as those that might be issues for minor cannabis possession offences, are not included in sanctions data (n=964). 12 Most commonly drugs, theft, offences relating to police/courts/prison and weapons possession. 13 Serious offences include homicide, serious violence, firearms possession, robbery, serious sexual offending and drug trafficking. 14 476 of the 13,974 Public Order charges were for offences in August 2011, the month of the London riots. 385 individuals (5% of cohort) were charged with at least one Public Order offence taking place in this month and 181 were convicted. 15 The measure of violence used in the offending and victimisation impact analysis: Violence against the person; weapons possession and robbery. For comparison, for just violence or weapons possession offences, 57% (n=4,139) had at least one sanction for an offence committed before inclusion on the Matrix. 24 How many people on the Matrix are subject to some form of judicial control? Individuals on the Matrix may be subject to certain controls from the courts if they have been found guilty of an offence – such as a custodial sentence or judicial restrictions such as Criminal Behaviour Orders, injunctions or restrictions on licence conditions. The proportion of individuals subject to some form of control has increased steadily over time, but as with populations, there is variation between boroughs in the types of controls used and in the frequency of their use. Individuals in custody remain on the Matrix during their sentence to ensure that police and partners can continue to monitor and re-evaluate their status in the months following their release. The proportion of individuals in custody has seen a small but steady increase over the Review period. Increasing from average of 27.2% in 2013/14 to 32.8% in 2017/18. September 2018 has the highest proportion in custody (35.2%). Graph 2: Proportion of Matrix Individuals in Custody Certain boroughs have much higher proportions of individuals in custody. For boroughs with more than 100 individuals on their Matrix, judicial restriction rates range from 9% to 39%, suggesting variance in practice at the local level. There may be a number of reasons for this. We know that currently borough Matrices are refreshed at different frequencies. This could indicate that on boroughs where the Matrix is regularly refreshed, a criminally active cohort is on the Matrix and so greater proportions of judicial restrictions are achieved, whereas on boroughs that refresh their cohort less often, lower proportions of judicial restrictions are seen. This could also reflect the level of police resources at a borough level focused on gangs and youth violence. 25 Matrix criminal careers: change over time and new additions Individuals on the Matrix are most often diverse and frequent offenders; however, there is significant variation. Our analysis compared criminal careers of a recent Matrix cohort (May 2018) to previous snapshots, as well as considering changes over the five-year period. In this section, individuals are discussed with reference to their harm banding (Red, Amber, Green), indicative of the frequency with which they are coming to police notice for violence and the severity of those incidents. As Table 2 sets out, the proportion of males added to the Matrix has remained the same over the period of analysis, and whilst there is some variance in the proportion of ethnicities added, this has not affected the overall population demographics at any snapshot in time. However, individuals are being added at an earlier age. As we have seen, most individuals on the Matrix have a criminal background. Our analysis explored how, if at all, this has changed over the five-year period of analysis. The age of first charge has remained relatively stable (around 14.5 years old) as has the age of first conviction (around 16 years old). However, more individuals are being added without a previous sanction (though it should be noted that the majority still have one; for new additions in 2017/18, 79% (n=362) had at least one sanction for an offence committed before Matrix inclusion). The decrease in age added is particularly clear for those under 18, where the proportion of individuals added per year has more than doubled from 26% in the base year to 56% (n=340) in Year 5. Practitioner perceptions, both in our own findings16 and from recent Home Office research (Disley & Liddle 2016), suggest individuals are becoming involved in gang or group violence at a younger age. Whilst our analysis can only account for police recorded offending – of which there is no discernible decrease in age of first charge or conviction over the 5-year period – it is clear that individuals are being added to the Matrix at a younger age. Table 2: Demographics of new additions to the Matrix by year17 16 A large proportion of VCS survey respondents thought that involvement of under 18s in serious criminality had increased over the last three years (78%, n=69). 17 MOPAC analysis of MPS data. Additions to Matrix Base Year One Year Two Year Three Year Four Year Five 2012 Jun 13-May 14 Jun 14-May 15 Jun 15-May 16 Jun 16-May 17 Jun 17-May 18 Male 99% 98% 99% 98% 98% 99% White 13.5% 15.2% 17.6% 20.2% 13.0% 17.7% BAME 86.5% 84.8% 82.4% 79.8% 87.0% 82.3% Black African Caribbean 75.9% 75.2% 67.3% 69.0% 78.9% 74.6% U18 25.7% 31.9% 40.7% 48.6% 52.3% 55.6% U25 85.8% 86.2% 84.9% 90.2% 91.5% 93.8% Average Age 20.4 20.2 19.7 18.8 18.4 18.0 26 Harm bandings now and over time Analysis explored changes in the pre-Matrix offending histories of individuals added to the Matrix over time. Table 3 illustrates new additions to the Matrix based on their harm banding when added. Red Those identified as the most harmful offenders. Making up the smallest proportion of new additions, nearly half of ‘Red’ individuals (47%; n=86) were under-18 in May 2018. As expected, Red individuals added to the Matrix are more likely to have a previous sanction, and have on average more previous sanctions overall, and across serious and violent crime types. Their offending is more serious - although this gap is lessening. Over time, the proportion of red nominals added with no previous sanctions has also increased slightly from 5% in 2013/14 to 12% in 2017/18. Amber Over time, the Amber group has become increasingly comparable to Red in terms of their criminal careers before being added to the Matrix. This group is most likely to have drugs, violence and weapons sanctions. The proportion with previous drugs sanctions is decreasing at a greater rate than that of Red individuals, with the proportion added with no previous sanctions also increasing over time; although still a minority (17%; n=38 in 2017/18 – see Table 3). Green The variation in criminal careers is most pronounced for those placed on the Matrix in the Green banding. Overall, Green individuals have the highest sanctions average, but also the greatest proportion without any sanctions - across the five-year period, 20% of new additions in the Green banding had no previous sanction (n=362), compared to 12% of Amber (n=205) and 10% of Red (n=29). As Table 3 demonstrates, the proportion with no previous sanctions has increased over time, with 27% (n=51) of Greens having no previous sanctions in Year 5 (2017/18). Greens are least likely to have a recent sanction. Greens also have the highest proportion of convictions for drugs-related offending. This can be explained by the very different types of people either being placed on or maintained on the Matrix. For example, this may reflect the feedback from practitioners that the Matrix is used as a reference tool, a way of maintaining corporate memory and tracking ‘elders’ who may not be directly committing crime but inciting others to do so. At the other end of the criminality scale, recent operational policy changes to focus more on vulnerability and exploitation may have contributed to an increase in the number of ‘at 27 risk’/periphery young people added to the Matrix, also contributing to the wide range of offending backgrounds in the Green category. Zero harm-scoring Greens 2,318 individuals remained Green for their entire time on the Matrix. At any time, around 38% of individuals on the Matrix have a zero-harm score18. Around 15% of the total cohort had a zero-harm score for the duration of their time on the Matrix. There is no compelling data to demonstrate why these individuals are included on the Matrix. The MPS have told us their rationale for including persons not involved in violence in the Matrix is to identify those who are thought to be in a gang but have not yet been drawn into gang violence. This can then enable prevention activity by local borough police and partners where they carefully consider what support can be provided to ensure they are diverted away from activity that may result in violent offending. This is a key demographic – one which is the focus of a new public health approach to tackling violence in London. The Mayor is funding and leading the creation of a new partnership Violence Reduction Unit for London, intended to take a long-term approach to dealing with the underlying causes of violence. At the time of this Review’s publication, the development of the Violence Reduction Unit is in its earliest stages, and while we cannot speak for the Unit and partners, we pose the question here about how this group of at-risk individuals should best be identified, monitored and supported as part of a public health approach. Table 3: Criminal careers of new additions by harm status Change in an individual’s RAG status whilst on the Matrix can be seen as an indicator of impact on offending. Overall, the picture appears positive in that there is a net movement 18 This proportion has remained relatively stable since Mar 2015, averaging 37.7% 28 towards Green, the lowest harm status (60% of individuals who start Amber become Green, and 54% of individuals who start Red become Green), with very few individuals going from Green to Red (1% of individuals who start Green, and 4% of individuals who start Amber). Exploring the impact of the Gangs Matrix The question of determining robust impact, that is - has a given initiative made a demonstrable and measurable difference on a key outcome measure (i.e. offending or victimisation)? - is one of the most difficult questions to answer within criminal justice analysis. This is especially the case when investigating the impact of the Gangs Matrix. There are thousands of Matrix cases, each with different start, end and time periods; limits to readily available, standardised data on what happened to the individuals during their time on the Matrix; and there are real difficulties in generating a valid comparison group. This final point is especially important, as is it only with a robust comparison group that strong conclusions can be made in terms of impact. However, while there is limited data on partner activity and enforcement activity with Matrix individuals, we have clear data on offending and victimisation. This analysis sought to include a significant proportion of all Matrix individuals over the 5- year period - over 7,000 individuals – making it the most in-depth analysis of the impact of the Matrix yet conducted. For more details on this section, see Appendix 4. Impact on offending To begin with, analytics explored the proven offending of Matrix individuals before, during and after their period on the Matrix. Graph 3 presents the proportion of individuals on the Matrix to have received a sanction in each time period. As can be seen, the cohort presents an increasing proportion of sanctions leading up to inclusion on the Matrix, a sharp decline once on the Matrix, and then a steady decline once removed from the Matrix. 29 Graph 3: Proportion of cohort sanctioned (based on date of offence) To illustrate, in the six months prior to their inclusion, 42% of the individuals on the Matrix had received a sanction for an offence. Within the first six months of their inclusion on the Matrix this declined slightly to 39% receiving a sanction. Finally, in the six months subsequent to removal from the Matrix, this fell to only 20% receiving a sanction. This overall trend also applies to violence.19 This is reflected in the evidence discussed in the previous section, indicating many individuals would move from higher harm bandings to lower bandings during their time on the Matrix. Table 4: Proportions of Matrix cohort sanctioned (all crime) If we explore the average number of sanctions of the Matrix cohort to include the entire length of time spent on the Matrix, we also see positive indications. Average sanctions per month decrease from 0.14 in the two years before the Matrix, to 0.13 whilst on the Matrix and 0.07 in the two years following removal. It should be stated that, when controlling for periods of time spent in custody when on the Matrix, the above findings are similar. Those individuals that were not in custody at any time 19 ‘Violence’ in this case refers to a composite of ‘Offences against the person’, ‘Weapons Possession’ and ‘Robbery’. In terms of proven offending, 16% of the cohort committed a violence offence in the six months before being added to the Matrix, 14% in the first six months following inclusion and 5% in the six months following removal. Period Cohort Sanctions Offenders Avg Sanctions (all) Avg Sanctions (sanctioned only) Proportion Sanctioned Six months before Matrix inclusion 7129 7903 3013 1.11 2.62 42.3% Six months during Matrx inclusion 6585 6770 2551 1.03 2.65 38.7% Six months after Matrix removal 3032 1379 599 0.45 2.30 19.8% 30 during their time on the Matrix still saw a similar decline in offending – so we can say that the decline observed in proven offending is not only driven by custody. As the previous section explained, Matrix individuals are subject to increased police attentions utilising a variety of tactics and controls such as judicial restrictions. However, there is a lack of data on non-criminal justice interventions which mean it is not possible to isolate the potential impact of all these different elements. Exploring offending across the different harm bandings Analysis also explored proven offending across the different harm bandings. In each of the bandings (Red, Amber, Green) a generally similar trend to what we have seen so far is observed, but there are differences. The Red group present the highest level of proven offending before inclusion, followed by Amber, and then Green. This is consistent with the findings on criminal careers of new additions in the previous section. This peak in offending suggests that the inclusion onto the Matrix for the Red and Amber individuals would seem to be appropriate. However, the much lower offending levels for individuals categorised as Green, and the less prominent decrease in offending amongst this group continues to raise questions about the appropriateness of these individuals’ inclusion when non-enforcement alternatives may be more suitable (e.g. youth engagement or safeguarding approaches). Graph 4: Proportion of cohort sanctioned - by harm status on entry 31 Comparing the Matrix cohort to a comparison group So far, findings appear to suggest that presence on the Matrix reduces offending. However, we must remain mindful of the challenges of understanding the reasons behind that impact. Given that we know offending and reoffending in this crime type tends to reduce with age, there is a challenge in understanding if it is the impact of the enforcement and partner activity associated with being on the Matrix which causes this reduction, or whether that would simply have happened over time in any event. In other situations when seeking to test the impact of a policy, the strongest method would be a randomised control trial – comparing one group receiving an intervention with another not receiving it selected by chance. However, in the case of the Matrix, this would not be possible for operational, ethical and public safety reasons. For this Review, we have sought out an alternative approach to testing the impact of the Matrix. MOPAC worked alongside a leading university (University College London, via the Institute of Global City Policing) to explore the feasibility of meaningfully examining impact in this context via a quasi-experimental approach called ‘within-group reference point shuffle’. Put as simply as possible, this method allows us to generate a valid comparison group from within the sample we are seeking to compare with – in this case the population of the Gangs Matrix – by reshuffling the data to give each individual a different, randomly selected date of inclusion on the Matrix, rather than their actual date of inclusion. Findings from this approach can be seen in Graph 5, which presents the proportion of cohort offending by month. The ‘observed’ line represents the overall Matrix sanctions graph at the beginning of this section. The orange line represents an estimate of how the average offending of this group would have appeared if the intervention had not taken place, with the band around it representing the upper and lower limits of normal variation. 32 Graph 5: Proportion sanctioned by month (within-group reference point shuffle) Prior to the Matrix, both groups are similar in terms of sanctions. The difference between the comparison line and that observed for Matrix individuals several months prior to and at inclusion is stark. However, the likely interpretation is that this simply relates to the reason individuals were added (e.g. a recent uplift in offending bringing them to police attention). In this way, it is appropriate that our comparison group do not have this 'peak'. During inclusion on the Matrix, offending decreases, rapidly coming back within expected levels and then dropping below the comparison level. Crucially, there is also a departure below the orange band after removal from the Matrix, suggesting that Matrix individuals are offending In depth – the within-group reference point shuffle This method of analysis is inspired by a technique commonly used in the spatio-temporal analysis of crime, called the Knox test (Knox, 1964). The test is used as a means of identifying space-time clustering; that is, the tendency of incidents to appear close to each other in space and time, commonly manifested as 'near-repeat' victimisation (see Johnson et al. 2007). The current approach seeks to establish a baseline level reflecting the volume and time course of offending that would have been expected if individuals had not been included on the Matrix. It does this by calculating how the offending curve would appear if the true time of Matrix inclusion was replaced by a randomly-chosen alternative; that is, if the curve was calculated with respect to a different inclusion point. Importantly, this method controls for the age-crime relationship (whereby crime increases with age and then subsides over time), which would otherwise provide an alternative explanation for a peak of offending around Matrix entry. 33 at a lower rate in the months after removal than might be expected. It should also be stated that there is no evidence that offending increases during or after removal from the Matrix. This analysis suggests that being on the Matrix reduces offending when compared to what would otherwise be expected for the cohort. This also applied when looking at violent offending.20 The overall approach is innovative and reasonably robust. The work – outlined at greater length in Appendix 4 - has been peer-reviewed by academics at University College London. As with other quasi-experimental designs, absolute statements of causation are not possible, but the approach is able to provide useful insights that could be built upon in further research. Digging deeper into impact So far, we have looked at proven offending but there are other measures in terms of impact that can be explored such as timeliness of arrest; levels of victimisation and the number of stop and searches. This section now considers these issues. Timeliness of arrest Timeliness of arrest relates to the time between when an individual committed an offence and when they were arrested for it, the rationale being that increased attentions resulting from Matrix inclusion mean individuals should be arrested quicker once an offence is committed. Across all three time-periods (i.e. before, during and after), nearly two thirds of arrests of Matrix individuals were made on the same day as the offence was committed. When considering averages, we do see that individuals on the Matrix are arrested, on average, four days quicker than either before or after their inclusion on the Matrix. The time between offence and arrest gradually increases the longer individuals stay on the Matrix, perhaps indicative of changes in offending behaviours and police attentions.21 The quicker speed to arrest gives us a proxy of more intensive policing of individuals while on the Matrix. Importantly, the speed returns to pre-Matrix levels once individuals are removed. 20 Violent offending was defined as ‘Offences against the person’, ‘robbery’ and ’weapons possession’. 21 Offence date and arrest date was used as available within the PNC dataset. This is not intended to be representative of investigative efficiency and it is recognised that many investigative variables are not accounted for (CCTV circulation; suspect identification etc.). 34 Graph 6: Average days between offence and arrest It is also possible to explore the difference between arrest and conviction, a proxy for the wider criminal justice service. There was little difference in the time between arrest and conviction22 across the three periods, suggesting that while the Matrix is having an impact on policing activity, inclusion is not influencing the wider criminal justice service. Through our community engagement work we heard concerns that the Matrix was used at court to seek longer or harsher sentences for those that featured on it. For a case to progress to conviction the facts must be proved to a judge or jury, or in the case of civil enforcement, a pattern of harassment, alarm or distress must be proved. Matrix inclusion alone, without this evidence, would not lead to a conviction, order or injunction, but there is a pertinent question about whether inclusion leads to tougher sentencing. The MPS’ Matrix Operating Model is clear that Matrix inclusion should not be used in court, and 74% of MPS respondents did not believe that being on the Matrix meant tougher sentencing outcomes. However, 42% of MPS practitioners who responded to our survey said they frequently or often used the Matrix as legal evidence for gang membership. Whilst we identify no legal issues with this discrepancy, it is a clear demonstration of inconsistency between guidance and practice, and the need for further training of officers on applying the Matrix Operating Model on a consistent basis. While the information we have had access to highlights discrepancies on the use of Matrix information in sentencing, we note the challenges in coming to firm conclusions around it because the Matrix does not currently capture details of interventions that result from an individual’s inclusion. MOPAC will work with the MPS and local authorities to explore better ways to capture future information on interventions received. 22 Days between arrest and conviction: 125.1 days before going on Matrix; 124.6 days during; 129.7 after removal. 35 Stop and search Stop and search remains a high-profile issue for policing in London and around the country. In recent years, the MPS has made a concerted effort to reduce stop and search, with a resulting 42% reduction in the number of stops and searches conducted between 2014 and 2017. During that period there was a 29% reduction in the number of black individuals stopped and searched, compared to a 54% reduction for white individuals. Currently, black individuals in London are more than four times more likely to be stopped and searched than white individuals.23 Of the total number of people stopped and searched in London by MPS officers between April 2017 and March 2018,24 current Matrix individuals accounted for 2% (n=1,541). This represents 37% of individuals on the Matrix being stopped during this period. In total, 3,585 stops and searches were conducted on these individuals – an average of 2.3 stops per person. Individuals not on the Matrix were stopped an average of 1.4 times.25 Analysis also compared similar groups in terms of ethnicity and age. For black individuals aged 18-21, the difference in average stops remains broadly the same; those stopped while on the Matrix averaged 2.1 stops per person compared to 1.5 stops for individuals not on the Matrix. It is therefore clear that Matrix individuals are subjected to stop and search at a greater rate than the population of individuals stopped who are not on the Matrix. This finding is broadly in line with previous research conducted on stops on gang and non-gang members involved in the London riots, which found that ‘the stop and search tactic is more finely attuned to intelligence into recent criminal activity or known gang membership’ (Stanko & Dawson 2012 p.7). In terms of outcomes, 26% of Matrix individuals stopped and searched were arrested, with No Further Action (NFA) taken on 69% of stops. In comparison, those never on the Matrix were arrested 19% of the time, with a similar proportion NFA’d (67.5%). These similarities in outcomes suggest that Matrix individuals are not being stopped any more arbitrarily than non-Matrix individuals. Analysis also explored the differences in the frequency and concentration of stops before, during and after inclusion on the Matrix.26 The proportion of individuals stopped from the 23 Stop and Search data taken from the MPS Stop and Search Dashboard. Disproportionality is calculated on year projection population figures (source: London Data Store). 24 This relates to records where an individual can be identified. There were 132,699 recorded stops between April 2017 and March 2018, of which 126,118 could be assigned to identifiable individuals (e.g. details not withheld, date of birth provided). See methodology for further details. 25 88,846 individuals over 120,369 stops 26 Due to a change in data management systems, individual level stop and search data for Matrix individuals was only available from 1st January 2016. 36 available cohort was compared in the two years prior to being added to the Matrix, the first two years of Matrix inclusion, and the two years following removal. As expected, stop and search increased the nearer individuals got to inclusion on the Matrix. Once on the Matrix, the proportion of individuals being stopped decreased, replicating the change in offending patterns discussed elsewhere in this Review. This suggests that individuals may not be targeted at the levels some anecdotal evidence suggests; whilst there is evidence that some individuals are repeat stopped, a greater proportion of individuals were not stopped at all (or stops were not recorded) in the period of analysis. To illustrate: • 36% of the cohort were stopped at least once in the three months prior to being added to the Matrix (Average 0.63 stops per person; 1.72 for only those stopped at least once). The maximum number of stops on one individual in this period was seven. • 32% of the cohort were stopped at least once in the first three months after being added to the Matrix (Average: 0.62 stops per person; 1.92 for only those stopped at least once). The maximum number of stops on one individual in this period was ten. • 10% of the cohort were stopped at least once in the first three months after being removed from the Matrix (Average 0.12 stops per person; 1.21 for only those stopped at least once). The maximum number of stops on one individual in this period was four. This latter point suggests a reduction in police attention following removal from the Matrix is a significant one, potentially allaying concerns about Matrix status ‘following’ an individual after removal. Analysis found little difference in the level of stops between ethnicities. The figures also highlight the large variation in the number of stops between individuals; these are often concentrated within a relatively short time period. The differences in levels of stops between the different harm bandings when first added to the Matrix was also explored. Here, the difference between high risk (Red and Amber) individuals and lower risk Green is stark. Only 13% of Green banded individuals were stopped in the 3 months before being added to the Matrix, compared to 51% of Reds and 41% of Ambers. Once on the Matrix, the difference in the level of stops between Green and the higher harm bands continues to be apparent. Thus, analysis suggests that generally the police are targeting those individuals with greater levels of offending, rather than simply using the Matrix as a reason to stop. 37 Graph 7: Proportion of Matrix individuals stopped by harm status on inclusion Levels of victimisation The protective aspect of inclusion on the Gangs Matrix is an important, but often overlooked element. This is especially so given the evidence shown earlier in this Review about the higher severity of gang-related violence. Young black males are disproportionately represented as victims of serious violence in London - the likelihood of victimisation for knife crime with injury for black males aged 18-24 is 5.9 times greater than that for white males aged 18-24.27 Analysis explored victimisation of Matrix individuals before, during and after inclusion.28 Once again, analysis (set out in Table 5) demonstrates increased levels of victimisation - including violent victimisation29 - in the period directly before inclusion to the Matrix, a reduction in the six months after being added and a greater reduction in victimisation after removal from the Matrix. Table 5: Proportion of cohort as victims of violence on MPS crime reports (01/01/2014-31/03/2018) 27 FY2017/18 data. Black males aged 18-24 have a victimisation rate for Knife Crime with Injury of 9.64 per 1,000 population. White males aged 18-24 have a victimisation rate for Knife Crime with Injury of 1.63 per 1,000 population. 28 Victimisation identified through Crime reports between 2014 and March 2018. It should be noted that it is more likely some victims will be missed from the dataset due to the matching and extraction techniques available. 29 CRIS reports relating to Violence against the Person, Robbery and Weapons Possession Offences. Serious sexual offending was low prevalence across the cohort and therefore excluded along with various several other violent crime types of low prevalence. Period Cohort Victim Reports Victims Avg Victims (all) Avg Victims (victims only) Proportion Victims Six months before Matrix inclusion 2536 440 373 0.17 1.18 14.7% Six months during Matrix inclusion 2775 394 341 0.14 1.16 12.3% Six months after Matrix removal 2881 128 119 0.04 1.08 4.1% 38 Victimisation follows a similar pattern to offending, with sustained lower levels both during and crucially after removal from the Matrix. This suggests that the Matrix has an important role in reducing harm amongst those included in it, particularly young black males. Graph 8: Proportion of cohort victims of violence on MPS crime reports (01/01/2014-31/03/2018) Practitioner perceptions - impact Practitioners offered mixed opinions in terms of the perceived impact of the Matrix, with Local Authority practitioners being slightly more optimistic than police officers that the Gangs Matrix had helped reduce gang violence. Police officers generally saw the Matrix as an efficient monitoring system for violent gang members, and a useful tool to direct police resource, although Local Authority practitioners were more reticent. Several respondents emphasised a need not to overcomplicate a tool that served them well as one element of a suite of approaches: “The Matrix is a police tool and used with additional systems is an excellent document. It should not be overly complicated …there does not need to be a vast amount of info … a danger that over-complication would lead to the Matrix not being managed effectively.” (Police Officer) In terms of positive outcomes, all groups were also unsure that being on the Matrix gives individuals a better chance to exit gang life. Several respondents raised concerns around the focus on – and resource dedicated to - gang individuals compared to those not on the Matrix, questioning whether this left a resource ‘vacuum’. 39 “It puts up barriers to resourcing any enforcement against those ‘not on the Matrix’ and suggests that they’re not ‘at risk’ simply because they’re not on.” (Local Authority Practitioner) However, the lack of standardised collection of data on non-criminal justice interventions with Matrix individuals limits the evaluation of the specific drivers of these outcomes. Some police officers and local authority practitioners interviewed noted the low harm bandings – or omission altogether – of influential or high-ranking criminals from the Matrix as an issue of concern. The response below is indicative of a perception that some ‘hands- off’ criminals were not scoring as highly as they should be. “High level criminals/most influential often don’t appear or appear at a very low level - possibly because they are not getting arrested or doing the ‘dirty’ work.” (Local Authority Practitioner) Reflections and recommendations The evidence shows that across various measures, the Gangs Matrix does appear to reduce offending and victimisation amongst the individuals included on it. It is also positive that after removal from the Matrix, victimisation and offending remains low whilst speed of arrest returns to pre-Matrix levels; and based on our analysis there is no evidence of increased police activity or offending once removed from the Matrix. We see positive indications in the movement of individuals from one harm banding to another and that over time, many Matrix individuals will be assessed as being less likely to be the perpetrator or victim of violence than they were when they were first brought on to the Matrix. However, as we are clear, it has not been possible with the data and information available to isolate the elements or combination of factors contributing to the identified effects. It is also not possible to be definitive that the impacts identified are a product of correlation or causation. There are questions around the harm bandings, whether the inclusion of individuals with the lowest harm scores is appropriate and how the risks facing these individuals are best addressed through a public health approach. 40 Analysis indicates that those going on at Red and Amber are most often (but not always) individuals with histories of proven serious offending, and that they are also more likely to be victims. In contrast, fewer Green individuals have previous sanctions for serious offending and a significant proportion do not receive such a sanction when placed on the Matrix, raising questions as to the justification for their inclusion. Recommendation: That the MPS improves systematic data capture across all aspects of the Matrix process. To include but not limited to: • demographics of Matrix individuals (gender, age and ethnicity); • nature and extent of police activity for those on the Matrix; and • nature and extent of non-enforcement interventions (needs, referrals, uptake and outcomes). We further recommend that MOPAC and the MPS conduct an annual review of the Matrix population, in comparison with the wider London gang and violent offending profiles. MOPAC will also convene partners to discuss options for enabling better collection of data on the impact of non-criminal justice interventions. Recommendation: MOPAC will oversee implementation of the recommendations in this Review - and transparency in this work - by: • requiring the MPS to report annually on progress against recommendations to the MPS Ethics and Audit panels and publishing those reports; and • ensuring progress against recommendations are reported to the MOPAC / MPS Oversight Board. Recommendation: We recommend a thorough reappraisal of the individuals in the Green category, with a focus on: those that currently score ‘zero-harm’; those that have never had a harm score or have remained in the Green category for their entire time on the Matrix; and those under the age of 18. This reappraisal should begin as soon as possible and be concluded no later than 31st December 2019. This reappraisal should consider whether: the level of risk they present justifies their continued inclusion; their inclusion is consistent with the published purpose of the Matrix; and whether their inclusion is compatible with Article 8 (2) of the Human Rights Act 1998. Where an individual does not meet these criteria, they should be removed from the Matrix. 41 Gangs Matrix Processes 42 The Gangs Matrix has been the subject of intensive scrutiny and debate as to its lawfulness – as a concept and in its operation. As part of this Review, MOPAC has carefully considered - with particular regard to equality and privacy - reports by Amnesty and StopWatch, engaging with both organisations. MOPAC has also engaged with the Information Commissioner’s Office, who have conducted their own investigation into the Gangs Matrix and given an enforcement notice to the MPS as a result of serious breaches of data protection law. Further, MOPAC has reviewed the legal position on several key issues. Issues arising under the Human Rights Act 1998 (“HRA 1998”) in relation to European Convention on Human Rights (ECHR) Articles 8 and 14. The Authorised Professional Practice (APP) document for police officers states that: Human rights principles underpin every area of police work. From basic probationary training through to senior management courses, officers learn that human rights must sit at the heart of the conception, planning, implementation and control of every aspect of the operations of the police service. Section 3 of the HRA requires that the police must interpret and apply their legislative powers in a manner which is compatible with the ECHR. HRA 1998, section 6(1), provides that it is unlawful for a public authority to act in a way which is incompatible with a Convention right. Article 8 of the ECHR provides as follows: (1) Everyone has the right to respect for his private and family life, his home and his correspondence. (2) There shall be no interference by a public authority with the exercise of this right except such as is in accordance with the law and is necessary in a democratic society in the interests of national security, public safety or the economic well-being of the country, for the prevention of disorder or crime, for the protection of health or morals, or for the protection of the rights and freedoms of others. There is no doubt that the creation and maintenance by the MPS Commissioner of a database such as the Gangs Matrix would involve an interference with the right to respect for private life under Article 8(1). The question in relation to the Gangs Matrix is whether that interference can be justified under Article 8(2). 43 Legal opinion sought by MOPAC is that the Gangs Matrix is capable of being operated consistently with Article 8. However, inclusion on the Matrix leads to a set of specific consequences for individuals in terms of their interaction with the MPS; and the nature of those consequences will vary, depending on the individual’s Red, Amber or Green rating. Therefore, the MPS’ lack of a clear, publicly available policy document specifically setting out how the Matrix operates is an important shortcoming. Article 14 of the ECHR states: The enjoyment of the rights and freedoms set forth in this Convention shall be secured without discrimination on any ground such as sex, race, colour, language, religion, political or other opinion, national or social origin, association with a national minority, property, birth or other status. Article 14 does not provide for a free-standing right not to suffer discrimination. Rather, it relates to discrimination in connection with the enjoyment of the Convention rights themselves. Article 14 will therefore always need to be considered in conjunction with one or more of the other articles in the Convention. The criticism of the Gangs Matrix in relation to Article 14 is that it has a disproportionate impact on the black population of London, and that it therefore discriminates between black Londoners and others in relation to the enjoyment of the Article 8 right to privacy. In other words, black Londoners are more likely than others to suffer an interference with their Article 8 right as a consequence of the operation of the Gangs Matrix. However, we must also consider the other, protective aspect of the Gangs Matrix. Young black males are disproportionately likely to be the victims of violence and there is an argument that any measures to tackle the harm caused by gang crime are likely to disproportionately affect that demographic group. It has also been suggested that the application of the concept of a ‘gang’ in practice involves stigmatising elements of black street culture. As a result, black people are disproportionately represented on the Gangs Matrix. However, it is difficult to determine whether or not this is the case. The MPS has not completed an Equality Impact Assessment of the Gangs Matrix – a serious concern raised through our consultation and engagement and described in more detail later in this Review. We welcome their commitment to urgently rectify this issue. We further discuss the issue of disproportionality and make our recommendations on this issue elsewhere in this Review. 44 Data protection With effect from 25th May 2018, the Data Protection Act (DPA) 1998 was repealed and replaced by a new data protection regime set out in the General Data Protection Regulation (“GDPR”); the Data Protection Act 2018 (“DPA 2018”); and Directive (EU) 2016/680 (“the Law Enforcement Directive”). The new data protection law is complex. The main provisions are set out in the GDPR; but DPA 2018 contains various supplementary provisions which are necessary in order for the GDPR to have effect in the UK. In addition, DPA 2018 complies with the UK’s obligation to give effect to the Law Enforcement Directive: see Part 3 of DPA 2018, dealing with Law Enforcement Processing. The Gangs Matrix falls within the scope of the Law Enforcement Directive and of Part 3 of DPA 2018. DPA 2018 sets out six data protection principles. The most significant are the first and second principles. The first principle requires that the processing of personal data for any of the law enforcement purposes must be lawful and fair (DPA 2018, section 35). In order to satisfy this requirement, the processing must be based on one of two conditions: either the data subject must have given their consent, or the processing must be necessary for the performance of a task carried out for the law enforcement purpose by the competent authority. Clearly, the consent condition will not be satisfied in relation to the Gangs Matrix, and so it is the second condition that is relevant. The second data protection principle is set out in DPA section 36(1). It requires that personal data that is collected for a law enforcement purpose must not be processed in a manner incompatible with the purpose for which it was collected. At first sight this is inconsistent with the fact that the Gangs Matrix involves making use of personal data collected for other policing purposes, so as to generate an individual’s score for the purposes of the Gangs Matrix. However, section 36(1) is subject to section 36(3), which provides as follows. Personal data collected for a law enforcement purpose may be processed for any other law enforcement purpose (whether by the controller that collected the data or by another controller) provided that – (a) the controller is authorised by law to process the data for that purpose, and (b) the processing is necessary and proportionate to that other purpose. 45 Legal opinion is that provided, again, that the requirement of proportionality is met, this would permit policing data that was originally collected for some other purpose to be used in calculating individuals’ Gang Matrix scores. Legal advice suggests that, provided that the Gangs Matrix satisfies the requirements of Article 8, it is likely also to satisfy the requirements of the first and second data protection principles. With regard to the rights of data subjects, a question has been posed about whether the law requires that individuals who are listed on the Gangs Matrix must be specifically informed of that fact. This must be balanced with the operational requirements of policing and keeping the public safe – it is clear that disclosing information about persons on the Matrix to those persons may compromise operational activity to reduce harm and protect the public. However, we return to the point raised earlier in this section about the lack of public information on the Matrix, how it operates and what it means to those on it. In our recommendations, we put forward suggestions for greater transparency, with a view to balancing the rights of the public to access information about policing and understand how individuals’ data is processed and stored,whilst at the same time not compromising operational effectiveness in protecting the public. More broadly, in light of all of the above and in the knowledge that the ICO investigation found contraventions of the data protection principles, we welcome the investigation by the Information Commissioner’s Office around data protection. Their findings in many cases overlap with those of our Review. We have worked closely with them in developing this Review and fully support their recommendations, as well as the MPS’s action plan to rectify the issues identified. Statutory duties to share information Alongside the requirements to act within the human rights and data protection legislation outlined above, there are multiple pieces of legislation that place a duty on the police, local authorities and other bodies such as the NHS to share information for the purposes of crime reduction, safeguarding and promoting welfare and wellbeing. Often it is pursuant to these aims that information about a Matrix individual is shared between statutory agencies. RIPA The Regulation of Investigatory Powers Act 2000 (RIPA) is another consideration as it relates (among other things) to the use of social media as an information source in relation to the Gangs Matrix. There are two types of authorisation that may be required for social media monitoring under RIPA: 46 • authorisation for use of a covert human intelligence source (CHIS); and • authorisation for directed surveillance. It has been suggested that authorisation for use of a CHIS is required where a police officer forms an online relationship with a person (e.g. by “friending” them on social media) without disclosing their true identity. In cases where the only direct interaction between a police officer and a person of interest is that the former sends a friend request (without disclosing that they are a police officer) and the latter accepts it, legal advice suggests that this would not give rise to a sufficient relationship to require authorisation as a CHIS. However, if the social media interaction is more extensive than this, a different analysis might apply. It has also been suggested that where the police view public profiles and access open source material in order to help inform a decision as to whether an individual should be listed on the Gangs Matrix, then in every case this should require authorisation for directed surveillance under RIPA. It is however doubtful that viewing and considering material that has been placed online by an individual and made publicly available by them would usually constitute surveillance for RIPA purposes. If the viewing was intensive and repeated in relation to a specific target individual, then this might perhaps cross the line into being directed surveillance, but it is not easy to define where the border might be. Reflections and recommendations In principle, the Gangs Matrix can operate in accordance with the law. Its purpose – to reduce crime and disorder - is clearly in line with Article 8(2). The Operating Model provides the framework for ensuring that decisions on inclusion are not made in an arbitrary manner. Furthermore, as the analysis in this Review has shown, inclusion on the Matrix is about more than tackling perpetrating behaviour – it also has a protective impact, reducing victimisation. The Commissioner believes the Matrix is essential in tackling serious violence in London. However, it is equally clear from the findings of the ICO investigation that urgent action is needed to strengthen process around the storage, management and sharing of data and information; that the MPS must be absolutely clear on the purpose of the Matrix; and that transparency – where operationally possible - must be at the heart of the MPS’s use of the Matrix going forward. 47 We are clear that, whilst it is vital that these steps are taken, it is also vital that the community has confidence that these changes are being made and that they will have a positive impact on the Matrix. In our view, engaging with an independent third party as we work to deliver this Review’s recommendations – the Equality and Human Rights Commission - will provide important additional insight and assurance for the benefit of all concerned. Recommendation: We recommend that the MPS urgently improve their current Matrix processes to ensure that personal data and information are stored, managed, shared, protected and transmitted safely and appropriately, with particular reference to: • ensuring compliance with data protection principles and legislation, including the completion of a Data Protection Impact Assessment; • who the Matrix information is shared with outside of the MPS; • ensuring that any sharing of personal information is necessary and proportionate; and • implementing any recommendations set out in the ICO enforcement notice. Recommendation: We recommend that the MPS improve transparency by producing, by the end of 2018, publicly available, plain English and accessible information that: • answers frequently asked questions about how the Matrix works and its purpose; • sets out the governance and oversight mechanisms the Matrix is subject to; • explains the training officers receive that helps them implement the Operating Model; • describes the practical effect for an individual of being on the Matrix; • clarifies how information is shared and who with; • explains how an individual is added to and removed from the Matrix; • describes how the multi-agency process works; and • describes the relationship between the Matrix and the response to serious organised crime. Recommendation: That MOPAC and the MPS engage with the Equality and Human Rights Commission as we deliver the recommendations in this Review and the ongoing work of the Gangs Matrix, supporting the MPS’ work to further assess issues around human rights, disproportionality and produce an Equalities Impact Assessment. 48 In practice - how the Gangs Matrix is used at a borough level In some London boroughs – such as Westminster, Hackney, Lewisham and Islington - the response to young people involved in gangs and serious youth violence is co-ordinated through the Integrated Gangs Unit (IGU). The IGU differs on a borough basis but can be made up of roles such as police officers, specialist youth workers, analysts, council enforcement officers, mental health nurses and employment workers; and often has strong partnership links with probation services, youth offending workers, youth centres, the Multi-Agency Safeguarding Hub (MASH) and social workers. The principle of the team is to provide a tailored response to an individual young person who has been highlighted as being involved in youth violence or who is being exploited by a group or gang. This response might involve targeted enforcement of those where evidence and intelligence (the Matrix, or through partner or community intelligence) identifies an individual who is involved in violence. This enforcement is generally led by the police officers in the unit and is designed to reduce the risk of future violence from or against that individual. Examples of responses for each harm banding are included earlier in this document. Where there is evidence of a pattern of harassment, alarm or distress (e.g. witness statements) council enforcement powers may be used to pursue civil enforcement such as anti-social behaviour injunctions. For most individuals this response will include the offer of intensive mentoring and support for an individual, from a council specialist youth worker or a charitable organisation such as St Giles Trust, if they are willing to engage on a voluntary basis. The nature of this work varies depending on the needs of the individual and often is planned with the young person. It might be safety planning work, teaching conflict resolution skills, treatment for mental and physical health problems such as post-traumatic stress disorder, or support to reengage with medical help to treat historic injuries (e.g. stab wounds) that the young person did not seek medical assistance for at the time. The Matrix can play an important role in prioritising those young people who receive this multi-agency planning and resource. In some areas, the Matrix helps to identify the top 15- 20 young people to go for discussion at a multi-agency Gangs Panel – usually those that score most highly on the Matrix. Frontline practitioners that have a role in that young person’s life – this could be a social worker, probation officer, a support worker (from a charity) are invited to a designated time slot at the Gangs Panel and support and enforcement actions are agreed. On some areas, as well as those identified as posing the most risk of violence, the multi-agency Gangs Panel also looks at those that have moved 49 substantially up the Matrix to look at why the risks have increased around that young person and what enforcement, support and/or interventions can be put in place. Practitioner perspectives – aims and purpose of the Matrix Both police and local authority practitioners surveyed for this Review were clear on the overarching aims of the Matrix to identify, prioritise and manage risk of high-harm individuals. However, only a quarter of both groups specifically mentioned targeting violence when describing the key aims of the Matrix. As the fundamental focus of the Matrix, this suggests practitioners might hold different perceptions of harm in terms of violence versus the ‘hands-off’ influence and coercion of higher-level gang members. Police practitioners highlighted the Matrix’s usefulness as an indicator of gang membership and as a tool to inform local analysis and briefings. Nearly two-thirds of police respondents also saw it as a mechanism to assist with providing appropriate support or interventions, though nearly the same proportion viewed it as a performance management tool. Views on its use as an evidential tool for gang affiliation were also mixed, with 35% never using it this way and 42% using it this way frequently or quite often. Additional uses included assisting with the actioning of intelligence such as warrant applications. “The Matrix gives me a steer on violent gang members. It’s really useful for me to direct pro- active work… however…those not on the list shouldn’t be ignored.” (Police Officer) Local authority practitioners saw the Matrix as a useful partnership tool to inform, assess and manage risk. As with police officers, they highlighted its benefits in monitoring offending, directing interventions and to inform various assessments. These included pre-sentencing reports, Youth Offending and Probation risk assessments, Children’s Social Care assessments and housing options, highlighting the breath of the influence of the Matrix. “Presence on the Matrix is one of the risk factor considerations when developing interventions.” (Local Authority Practitioner) Local authority practitioners also suggested that the Matrix provided a useful starting point for inter-agency discussions concerning engagement, interventions and enforcement. Some boroughs were positive around the levels of partnership working: “Our partnership links with MPS are good. [The] MPS co-chair our Youth Offender Management Panel, Gangs and SYV Panel.” (Local Authority Practitioner) For police officers, further benefits included a provision of accountability and a corporate memory for gang related crime. Local authority practitioners recognised its benefits in 50 targeting interventions. Both groups were positive around the way it enabled effective monitoring of offending through a consistent approach to harm scoring. “It is a convenient list of gang membership to enable targeted engagement & enforcement work.” (Police Officer) A key theme emerging was a consistent misrepresentation of Matrix as a comprehensive gangs list. The Matrix is consistently referred as a list of all gang members in London and there exist numerous examples of this in the media, by politicians and in policy documents. Many officers saw the Matrix as a useful tool to confirm gang affiliation, although others recognised there were divergent understandings: “Officers that believe if they are not on the Matrix they are not a gang member - very common with senior officers.” (Police Officer) There was acknowledgement amongst some Voluntary and Community Sector practitioners that the Matrix is used to support at-risk individuals but also a perception that it may be used differently or understood differently by different organisations and that there needs to be a clear set of criteria on its use and purpose. Several respondents believed that it was misunderstood by communities also: “There are big misunderstandings amongst communities, activists and individuals about how the Matrix works and why it is there. These mistakes and misunderstandings have now found their way into popular gossip and belief, but they are often untrue and unchallenged.” (VCS Respondent) This underlines the need for clarity, transparency and proactive communication of the purpose of the Gangs Matrix, who should be on it and why - within the MPS, with partner agencies and with the general public. Addition and removal from the Matrix Practitioners had mixed views as to whether the right individuals were always being selected for the Matrix. Reasons for this were wide-ranging, with many stemming from different interpretations of the overarching aims and purpose of the Matrix as described. Some were more closely linked with differences in local policy and practice as the following section will explore. Nearly two-thirds of police officers surveyed for this Review acknowledged the difficulties in confidently assigning gang affiliation to individuals, and this is an important statistic to keep in mind in relation to the wider issue of unconscious or conscious bias in identifying 51 potential gang members. Five out of 23 boroughs indicated that they did not keep a record of the (two plus) sources used to confirm gang membership, which underlines the need for greater consistency in process and enhanced training for all who use the Matrix. Eligibility criteria varies significantly between boroughs, with both respondents and documentary evidence indicating that inclusion may be based on a number of parameters such as gender, age and offending levels. Most local police practitioners thought that female gang members were under-represented on their local Matrix. Some local authority practitioners felt that the MPS recognised the value of the intelligence they provided concerning gang membership and potential inclusions: “We contribute information/intelligence, that feeds into police intelligence, which contributes to an individual's Matrix score” (Local Authority Practitioner) However, there were mixed opinions on the extent to which their views on inclusion were taken on board, often dependent on the frequency and configuration of local multi-agency arrangements and personal relationships: “We were having meetings every six weeks where we would discuss the Matrix and negotiate who would be discussed at the GMAP meetings. However, over the last three months this has fallen by the wayside, I think due to a lack of resources from the police. (Local Authority Practitioner) Criteria for removing an individual from the Matrix include: • evidence that they are not engaging in gang activity and group violence and have not done so for a period of time; • they are engaging in a diversion program for a period of time (6 months) and have not come to police notice since that engagement started; • they have not come to police notice for a significant period (6 months+); • they are deceased or have been deported; or • they have moved away from London and are no longer believed involved in gang criminality within the Metropolitan Police area.30 Survey findings suggest that, in practice, the decision to remove individuals from the Matrix is far more police-orientated. 30 Most areas outside London use a national tool to monitor those involved in organised crime – the OCGM Organised Crime Group Management tool – and this can be where their higher-level gang individuals are tracked. Some other areas use local violence tracking tools, but these tend to be local and not centrally collated like the MPS Gangs Matrix. 52 “Police usually make the decision as to who is removed and sometimes consult the authority” (Local Authority Practitioner) We found that there is variation across boroughs as to when Matrix members are removed: for instance, time periods for ‘not coming to police notice’ ranged from anything between three and 24 months, though most boroughs did not actually specify a time period. Risk aversion was a common reason given for non-removal, meaning that individuals with no recent offending or victimisation history could be kept on, ‘just in case’ something happened, as one officer describes: “There is an element of fear that if a person is removed completely and they then become a victim or suspect of serious gang-related violence there would be heavy criticism.” (Police Officer) Others indicated that although some individuals should perhaps be removed, management would not allow them to do it: “I have attempted to remove over twenty subjects from the Matrix who do not conform. I have been told I cannot and they will fall away naturally.” (Police Officer) Findings indicate that reluctance to remove individuals also relates to the use of the Matrix as a long-term gang intelligence tool. For many officers, it presented an ‘easy point of reference’ to identify an individual’s gang affiliation. One borough indicated that ‘elders’ were kept on the Matrix despite zero scores, to ensure new officers were aware of their historic affiliations. “[The Matrix] provides a corporate memory - individuals are remembered, highlighted, discussed and either engaged with or targeted.” (Police Officer) Non-police respondents also voiced concerns with the removal process; over half of Voluntary and Community Sector (VCS) practitioners were not confident that individuals are removed from the Matrix if they are no longer involved in gangs. Training on the Gangs Matrix and related areas Perhaps linked to discrepancies in process, respondents indicated a need for training in some areas. Specifically, just under half of Matrix borough Single Points of Contact (SPOCS)31 reported receiving no or inadequate training on Matrix administration. 31 Always police officers – usually a Detective Constable (DC) or Detective Sergeant (DS) 53 Police practitioners also indicated little or no training around the Gangs Operating Model, an overarching suite of documents which outlines all elements of the MPS’ operational approach to gang policing including Matrix processes. Just over half of local authority practitioners reported receiving either inadequate or no training on the Matrix. Several local authority practitioners believed there would be benefits in extending the gang focus wider than those directly involved in the Matrix: “It would be helpful for the gangs unit to give this information to a wider audience than just those involved with the Matrix.” (Local Authority Practitioner) Another suggested that a coherent, transparent, partnership approach to training would be beneficial: “I think the police and partners should undertake the same training re: gangs. When using different providers people learn different things and there is no consistency…we can also to build better networking relationships…” (Local Authority Practitioner) Oversight and governance The Matrix is not centrally controlled; MPS boroughs have ownership of their local gang Matrix and differences in local organisational configuration, partnership integration and policy will impact on the way the Matrix is maintained, including who is selected and removed from the Matrix. The influence of partners on the Matrix and wider gang processes appears dependent on local configuration and integration. For example, local authorities might have different Single Points of Contact (SPOCS); this could be the head of Youth Offending Service on one borough or the Head of Community Safety on another. As highlighted previously, the partnership element of the Matrix selection and removal process generated some apparently contradictory views. Few police officers highlighted or acknowledged the partnership element of the Matrix, with a majority disagreeing that a consensus on removals or additions was usually reached with local authority counterparts. Conversely, most of the local authority practitioners interviewed believed that local authorities and police usually agree on who to add or remove. Local authority practitioner opinion as to whether police give sufficient consideration to their views was evenly split. “…it gives the whole Panel, which includes Youth Offending Services, Probation, Local Authority, Safer London, Youth Services, Education and Health, a sense of ownership and inclusion in the decision making.” (Police Officer) 54 The role of resourcing in dictating numbers placed on local Matrices appears significant. This is true both in terms of dedicated analytical resource to identify potential violent gang members or to build an accurate analytical picture of local problems. Both groups indicated that local priorities or resourcing affected the numbers of individuals placed on local Matrices. Insufficient resourcing was also raised by some respondents as a factor contributing to a perceived lack of accuracy: “Not enough dedicated intelligence or time is allocated to supporting the Matrix. Matrix was run by the BIU and LIT teams supported by central intelligence hubs who now longer exist or to accessed by borough officers. Dedicated - not rotated - staff are needed. This leads to errors and omissions.” (Police Officer) Resourcing can impact numbers on the Matrix both ways. There was some evidence from surveys to suggest that resourcing limited the numbers on their Matrix. In addition, some boroughs reported lacking sufficient resource to adequately refresh and audit their local Matrix, meaning that individuals who should have been removed were staying on the list. Most boroughs indicated refreshing their Matrix on a regular basis monthly or more frequently; however, there was no clarity or consistency in what was meant by ‘refresh’. Of forty-three descriptions of local processes in identifying gang members, only one mentioned analysis at all. For many boroughs, this has meant an increasing use of officers in analysis roles, which appears to be increasingly common across the MPS as resources have become strained. One PC described their work in such terms: “I link violent offences together…produce i2 [analytical software] association charts and assist with sorting through mobile phone data and use MapInfo [analytical mapping software] to produce maps to support investigations undertaken by our gangs team.” (Police Officer) In terms of constructive data sharing and analysis, many local authority respondents held similar views, with some boroughs highlighting barriers created by IT systems access, or high staff turnover making it difficult to build trust required to share sensitive information: “We used to have a meeting between an analyst and police to add / remove individuals. That no longer happens. This reflects the fluctuating police /council relationship based on individuals.” (Local Authority Practitioner) There was little indication of any robust local oversight of the Matrix process other than the multi-agency panels that might decide inclusion. Some responses suggested this 55 responsibility was solely in the hands of a single Matrix SPOC. Local quality-control processes varied considerably, with some boroughs relying purely on central oversight. Reflections and recommendations Differences in perception between police and other practitioners on the purpose and principles of the Matrix are compounded by inconsistency in process and practice in each borough and by weaknesses in overall governance. These are issues within the gift of the MPS to address and improvements must happen. In line with our responsibilities for overseeing policing in London and holding the Commissioner to account for the performance of the MPS, we will continue to scrutinise progress against this issue and all of the recommendations of this Review. Recommendation: We recommend that the MPS comprehensively overhaul the Matrix Operating Model and review the Model annually. We further recommend that all officers using the Model receive training on how to apply the guidance and to do so consistently across all boroughs. Both the Operating Model and the training should have a particular focus on ensuring: • that the right people are on the Matrix; • that people are added and removed in a standardised, evidence-based manner; • that they can be removed and that the ‘gang’ label will not ‘follow’ them; • that local Matrices are refreshed regularly so that individuals don’t stay on any longer than necessary; • that the guidance on the use of social media for intelligence purposes is updated; and • that the Data Protection principles and legislation are fully applied. 56 Perceptions and understanding of the Gangs Matrix 57 As outlined throughout this report, we know that some Londoners have more negative experiences of policing and crime than others, particularly those from BAME backgrounds. The same is true of their views of policing - the percentage of black Londoners who agree that the police treat everyone fairly, regardless of who they are, is 63% compared to the MPS average of 76%.32 This gap in trust is a significant concern in a system of policing by consent, which depends on the support of the public for its legitimacy and for its license to operate. Moreover, people who have trust and confidence in the police are more likely to abide by the law, cooperate with the police (Tyler 2006) and provide intelligence. Quite simply, increasing trust in policing increases its effectiveness. In his Police and Crime Plan, the Mayor has set out his agenda for addressing the disproportionalities we see in policing and crime, with this Review a key element of that work. Whilst we know that the perpetrators and victims of the crime types targeted through the Matrix are largely young black males, our analysis shows that at present, their representation on the Matrix is disproportionate to their rates of offending and victimisation. London-level disproportionality Young black males are disproportionately represented as both victims and offenders in all serious violence. At a London wide level, the percentage of black, African-Caribbean males on the Matrix exceeds that of the general population, as well as across many crime types, including those most associated with serious gang- or group-related violence. 32 https://www.london.gov.uk/what-we-do/mayors-office-policing-and-crime-mopac/data-and-statistics/public-voice- dashboard 58 Graph 9: Offenders - proportion of black African-Caribbean ethnicity by cohort/crime type33 The same is true when looking at victimisation: Graph 10: Victims - proportion of black African-Caribbean ethnicity by cohort / crime type 33 Sources: All data for offenders (PPA – Persons Proceeded against) and victims from CRIS FY2017/18 unless otherwise stated: Population projections - 2018 (GLA/London Datastore); Prolific Firearms Offenders/Habitual Knife Carriers - Nov 2018 (MPS Intelligence); GRITS (Gang Related Incident Tracking System) - 2018 to 05/11/18 (MPS Intelligence); First Time Entrants - 2017 (MoJ); GRITS [Gang Related Incident Tracking System] - 17/18 (MPS Intelligence); Knife Possession; Knife Injury (all sub categories); Homicide (all sub categories); Firearms Discharges (Lethal barrelled weapons only); all Nov 16 to Oct 18 (MPS Performance); Matrix September 18 (MPS Intelligence) 10 13 14 16 17 18 20 22 23 27 30 36 47 48 56 59 60 61 69 80 0 10 20 30 40 50 60 70 80 90BurglaryTheft & HandlingRobberyLondon PopulationOther NotifiableTotal Notifiable OffencesCriminal DamageSexual OffencesViolenceU25 Pop (Matrix>100)Serious Youth ViolenceKnife InjuryKnife Injury Vic U25 Non-DAGRITSFirearms DischargesGRITS (Violence only)U25 HomicideGang Flagged ViolenceU25 Knife HomicideMatrix * All data relates to victims except Population and Matrix 59 Hospital admission data supports police data in identifying young black males as victims of gun and knife related violence - black individuals are over-represented in both the knife injury incidents and gun injury incidents when compared to all injury types. However, although this London-wide comparison has been used to demonstrate disproportionality by previous research (Amnesty 2018; Williams 2016; Williams 2018), it may exaggerate findings for a number of reasons; there are large demographic variations across London, and young black males are disproportionately represented as both victims and offenders of serious violence. A more appropriate comparison, and one we have progressed, takes the BAME and black under-25 populations by borough (as the most likely group to be included on the Matrix), the proportion of individuals charged with various violent and group-related offence types; and compares these to the proportion of black/BAME individuals on the Matrix (Tables 6 and 7). Reading across the tables, the figures highlighted in dark grey relate to the proportions furthest from the Matrix demographic for each borough, those highlighted in yellow are the closest. For BAME individuals (Table 6), there is much variation between boroughs and across crime types. The proportion of BAME individuals charged rarely reflects that of the Matrix. Generally, disproportionality lessens when looking at more serious violent offending towards the right-hand side of the table. However, measurement at this aggregated level of ethnicity masks key differences. Table 6: For selected offence types, proportion of all charged who are BAME 34 34 Includes only boroughs with average Matrix population of > 100. CRIS data from FY2017/18 except ‘Gun Crime’ - aggregated 2014/15 – 2017/18 for total notifiable offences where a firearm feature code is present. Borough Population U25 BAME All TNO Drug Poss Drug Traffick Violence Against Person Serious Wound Robbery Weapons Knife Crime Knife w/injury Gun Crime Matrix BAME Lambeth 56% 62% 65% 69% 61% 62% 76% 78% 74% 64% 69% 95% Westminster 55% 46% 50% 55% 43% 40% 72% 65% 64% 53% 61% 77% Brent 72% 64% 73% 76% 64% 63% 65% 77% 67% 68% 77% 96% Waltham Forest 62% 49% 59% 54% 49% 48% 78% 57% 54% 56% 64% 85% Newham 80% 64% 72% 80% 62% 60% 85% 74% 74% 69% 77% 94% Enfield 53% 43% 50% 52% 42% 48% 60% 57% 62% 67% 65% 83% Barnet 45% 42% 55% 53% 42% 48% 54% 58% 54% 50% 74% 74% Greenwich 50% 39% 44% 55% 40% 42% 51% 62% 52% 49% 45% 87% Southwark 58% 55% 57% 62% 56% 58% 73% 76% 65% 71% 70% 94% Islington 45% 44% 48% 56% 41% 48% 60% 45% 50% 48% 49% 67% Haringey 48% 51% 60% 73% 47% 53% 64% 60% 60% 63% 62% 97% Barking & Dagenham 62% 40% 47% 53% 36% 34% 59% 43% 48% 44% 44% 76% Tower Hamlets 70% 64% 71% 80% 63% 73% 64% 66% 68% 76% 80% 80% Lewisham 62% 57% 64% 76% 56% 60% 79% 72% 70% 75% 76% 97% Hackney 54% 60% 67% 71% 59% 59% 72% 74% 72% 77% 63% 96% Wandsworth 41% 52% 55% 79% 49% 55% 43% 62% 54% 68% 56% 91% Croydon 64% 55% 62% 68% 54% 54% 77% 71% 72% 90% 81% 91% LONDON TOTALS 54% 49% 57% 65% 48% 50% 65% 63% 60% 63% 62% 87% Wider Offending Violent Offending Violence - Weapons 60 Only looking at black African-Caribbean ethnicity, the disproportionality is starker. All boroughs have black Matrix populations at least two times greater than the overall U25 population. Tower Hamlets has a black Matrix population 4.8 times greater than its overall U25 population, and this disproportionality can be seen across crime types. Overall, disproportionality generally becomes less in relation to the crime types that the Matrix is focusing on but is still apparent to some degree across all boroughs. Table 7: For selected offence types, the proportion of all those charged who are black African-Caribbean Do the individuals on the Matrix come from the most violent areas? The Matrix population is not distributed evenly across London’s boroughs. The Review tested whether the distribution of the individuals on the Matrix reflected the levels of violence in boroughs, and overall, analysis demonstrates that this is the case. In some Boroughs there are as few as three individuals on the Matrix, in others 300. The first five columns in Table 8 show the average number of people on the Matrix in each borough from 2013-2018. The last seven columns show the numbers and types of offences, focusing on violence, committed in these boroughs over that time. The colours – red to green – reflect the relative volumes. Overall, the matching of these colours between the number of people on the Matrix and the number of violent offences shows that the distribution of the Matrix population does generally reflect the distribution of violence. Borough Population U25 Black All TNO Drugs Poss Drug Traffick Violence Against Person Serious Wounding Robbery Weapons Knife Crime Knife w/injury Gun Crime Matrix Black Lambeth 40% 56% 58% 67% 55% 57% 74% 74% 72% 61% 63% 93% Westminster 13% 27% 27% 33% 25% 22% 44% 51% 37% 38% 34% 47% Brent 26% 41% 48% 66% 39% 42% 48% 59% 45% 40% 65% 89% Waltham Forest 25% 31% 36% 39% 30% 31% 62% 44% 43% 41% 43% 75% Newham 24% 36% 37% 51% 36% 38% 61% 46% 44% 45% 61% 87% Enfield 29% 35% 42% 48% 32% 40% 60% 51% 57% 56% 63% 79% Barnet 14% 25% 30% 33% 25% 26% 39% 39% 41% 28% 46% 60% Greenwich 31% 32% 36% 43% 32% 33% 50% 54% 49% 47% 42% 82% Southwark 37% 48% 49% 55% 49% 52% 71% 72% 61% 65% 61% 94% Islington 22% 32% 31% 45% 32% 39% 51% 39% 42% 42% 36% 64% Haringey 26% 44% 50% 64% 40% 47% 63% 56% 57% 60% 61% 94% Barking & Dagenham 33% 26% 31% 35% 24% 22% 42% 37% 32% 28% 31% 71% Tower Hamlets 11% 20% 16% 21% 20% 19% 26% 23% 19% 17% 23% 52% Lewisham 41% 51% 57% 71% 50% 53% 75% 67% 63% 70% 75% 93% Hackney 29% 49% 51% 50% 49% 52% 66% 69% 66% 73% 52% 90% Wandsworth 20% 41% 41% 74% 38% 45% 40% 52% 50% 61% 52% 83% Croydon 37% 44% 50% 57% 45% 44% 70% 64% 67% 81% 78% 90% LONDON TOTALS 22% 33% 37% 48% 32% 35% 52% 51% 46% 47% 54% 79% Wider Offending Violent Offending Violence - Weapons 61 Table 8: Matrix Population and Key Offence Statistics Why does disproportionality occur? Therefore, understanding the drivers of this disproportionality is a difficult and complicated task, rooted in wider issues around policing, criminal justice and society as a whole. Several arguments can and have been made. Racial discrimination is a possible factor in disproportionality, and we have considered this carefully in our analysis and in our engagement with police, partners and community representatives. David Lammy’s review into the treatment of BAME Londoners in the criminal justice service supported the Mayor’s commitment to review the Matrix and asked MOPAC to ‘examine the way information is gathered, verified, stored and shared, with specific reference to BAME disproportionality’ and to consider community perspectives. This Review is part of MOPAC’s response to this. We must acknowledge the possibility of conscious or unconscious bias against young black males in London – whether the term ‘gang’ is now heavily racially-loaded and that this perception - that a gang is often comprised of young black males and ergo that young black 62 males are often in a gang – either directly or unconsciously influences the enforcement focus of the police and subsequent actions of the justice service. However, proving or disproving this point in the case of the Matrix is extremely difficult and would require analysis of the process and drivers behind every single action and decision leading to an individual’s addition, including embedded intelligence processes far wider than the Matrix. MOPAC is currently undertaking complimentary analysis looking deeper into structural drivers of knife crime and violence – such as physical and mental health, education and access to opportunities – building on the evidence base established for the Mayor’s Knife Crime Strategy and supporting our work in developing a public health approach to tackling violence through the London Violence Reduction Unit (VRU). It may be that these wider characteristics are more predictive of Matrix membership. Using methodology from the Jill Dando Institute at University College London, the Vulnerable Locality Index35 uses datasets related to crime (burglary & criminal damage rates), deprivation (claimant count rate, GCSE capped point score, average household income) and population (resident population density for the 10-24 age group) to identify the wards of highest vulnerability to crime and community safety problems. On average, the most vulnerable wards have a higher proportion of BAME population than the least vulnerable. The figures below demonstrate signification variation across London: • In the top 10% vulnerable wards the average percentage of population that are BAME is 48%. This ranges from 94% in Southall Broadway (Ealing) to 14% in Cray Valley West (Bromley) • In the least vulnerable wards (90-100%) the population is on average 28% BAME. This ranged from 84% in Hillside (Merton) and 4% in Biggin Hill (Bromley)36 Through the Index, we know that victims are twice as likely to be offended against for both sexual offences and gun crime within the top 10% most vulnerable wards in London when compared with the 10% least vulnerable. For knife crime with injury the disparity is even greater, with victims six times more likely to be offended against within the most vulnerable wards in London. If we score across these three high harm crime types - nearly half of the highest risk wards are located within just six boroughs. The disproportional representation of Matrix individuals living in vulnerable locations is stark. They are 10.6 times more likely to live in the top 10% (n=859) of vulnerable wards than the bottom 10% (n=81) and 3.5 times more Matrix individuals reside in the top 50% (n=2,661) of vulnerable wards than the least vulnerable (n=755). 35 https://www.london.gov.uk/what-we-do/mayors-office-policing-and-crime-mopac/data-and-statistics/london-landscape 36 Based on 2011 census data; average BAME proportions should only be used as an indication only. 63 Whilst we cannot give a definitive answer on the causes of disproportionality, the evidence presented in this Review clearly illustrates that the Matrix disproportionately includes a group that is already disproportionately likely to be perpetrators and victims of serious violence. The recommendations of this Review are aimed at reducing unjustifiable disproportionality and ensuing that the processes that govern the Matrix are more robust and more transparent. Practitioner views of disproportionality Although both police and local authority practitioners believed that the most violent gang members, and those most at risk of violent victimisation were captured on the Matrix, significantly more local authority practitioners saw BAME gang members as over- represented and white gang members as under-represented. Two thirds of police respondents viewed the Matrix as representative of all gang members on their respective boroughs; the same proportion of local authority practitioners perceived the Matrix to be under-representative of this group. Several VCS practitioners questioned the efficacy of the current MPS gang definition, suggesting it could result in groups of young people being incorrectly and unfairly labelled: “The Met definition of a gang can be applied to almost all homogeneous groups of young and other people in theory however in practice and from the data obtained it appears to disproportionately and predominately apply to BME young males.” (VCS Practitioner) Reflections and recommendations The evidence is clear that young BAME men are disproportionately represented as victims and offenders in London, and it appears that this is even more so on the Gangs Matrix. 64 It is equally clear that this disproportionality is keenly felt by communities, and that in turn this affects their overall confidence in policing. The recommendations in this Review are intended to address the issues we are able to evidence, disproportionality being one of the biggest. By implementing the recommendations of this Review: helping to clarify the aims, improve training, bring consistency to practitioners’ usage and strengthen oversight over the Matrix, we hope to see the population of the Matrix align much more closely with the data on violence in London. The community perspective We have made considerable efforts to hear the range of views from all stakeholders; local authorities, the police and voluntary sector organisations working with young people. We know how important this is to Londoners and that there are strong views on either side of this issue, so it has been equally important to ensure we have heard and understood the views and concerns of London's communities, particularly young people. Recommendation: We recommend that the MPS strengthen their governance of the Matrix and the officers and partners that use it, creating single points of responsibility on each Borough Command Unit to: • ensure there is no discriminatory practice; • ensure that risks around data breaches are properly assessed and mitigated; • track progress against recommendations published in this report; • review the intelligence processes and flows that create organisational knowledge around gangs; • assess the Matrix in the light of technological advances to future proof it; • oversee the continuation and expansion of regular borough audits; and • work with MOPAC to conduct new analytics exploring issues relating to disproportionality and the Matrix (e.g. micro-level demographics analysis in gang affected areas) Recommendation: The MPS should consider whether the lessons learned from this Review are applicable to the operation of, and sharing of information under, other operational tools. 65 The Matrix Review Reference Group To ensure this Review was informed by community views, we brought together a Reference Group to work with us in delivering this work (see membership at Appendix 2). The Reference Group included people working in and with communities and young people who have been affected by violent crime, those with experience of working with the Metropolitan Police Service, those who have researched the Matrix and those with expertise in human rights and data protection. The Group had access to the data that underpins this report and were able to add their invaluable views and expertise to this work. The Reference Group met on five occasions. The summaries below represent the majority or consensus view of the group but do not quote individual members. 12 July 2018 – The first meeting explained the scope of the Review and briefed attendees on the process and practitioner findings to date. It also provided an opportunity for attendees to share their views on: what the MPS and MOPAC say to communities about the Gangs Matrix; what it does and how it works; and how best to ‘bust’ some of the myths that surround the Gangs Matrix. There was consensus across the group that the Matrix needed improvement as a tool. The Reference Group made strong representations about the need to ensure the Review is contextualised and recognises the disproportionate nature and impact of violence in our communities, as well as the impact of austerity and deprivation in those same communities (as evidenced in the data above). The Review is not being conducted in a vacuum, and the Group highlighted the impact of what they considered a failure to sufficiently invest in community engagement/relations [in these times of austerity] and the harm that has been caused to community relations by some previous initiatives, such as Operation Shield. The Group felt that the Review should also acknowledge the disproportionate impact of the Matrix on black boys and men and raised concerns about the potential for discrimination with the use of such tools. For the Reference Group, one marker of the success of the Review would be that they and others would in future be better-equipped to be able to articulate responses to the community around their concerns, and that this would start to build confidence in the community and the police. While the Group expressed significant concerns about the impact of the Matrix, with some advocating for its discontinuation entirely, there was a recognition amongst the Group that risk, harm and vulnerability management tools are commonly used across the public sector and that there is a need for organisations, including the police, to be able to have the means to identify and respond to these variables. 66 The Reference Group emphasised the importance of making information available to the community/public so that they can better understand how the Matrix is intended to operate, whether it is effective in tackling inter-borough and County Lines activity and the overarching measures of success. It is also necessary for the police to explain the different terminology they use. For example, the use of the term ‘intelligence’; it is one of the things that drives the Matrix and yet the public do not understand what that represents. It would also be helpful for the MPS to be able to describe the measures of harm and vulnerability that are used to categorise people on the Matrix. In addition, for the community to have confidence in the Matrix, the Reference Group advised that there should be greater transparency and consistency in the process of people going on and coming off the Matrix and any formulas used to determine risk as well as clarity about who Matrix information may be shared with and assurance that any sharing of information is necessary, proportionate and accurate. The Reference Group said that in their experience, there were inconsistencies in how police officers speak to the public about the Matrix, with references being made to both a Gangs Matrix and a Violence Matrix. To address this, the Group thought the MPS should develop consistent and standardised language. Providing more information on the Matrix to the public as a matter of routine (as described above) would help improve consistency in this respect. The Group saw inconsistencies also in how the Matrix was implemented by borough police, leading to variation in how it is adopted, used and implemented. The Group agreed that this was unhelpful and that there must be guidance put in place to ensure consistent usage, terminology and adoption across the MPS. As well as making more information about the Matrix processes available to the public, the Reference Group put forward the view that the police should be explicit about the guidance and training that is given to officers in order that the Matrix be operated consistently and in compliance with information governance standards. The Group was clear that it is important that any information produced for communities and young people around the Matrix is accessible and in plain English format “so they can feel part of the conversation”. 25 September and 8 October 2018 – The second and third meetings focused on the presentation of our emerging analysis of disproportionality in the context of the Gangs Matrix. It provided an opportunity for the Group to give their views on why disproportionality occurs, whether the focus on the term ‘gang’ and how that is defined plays into the disproportionality, as well as the impact of wider organisational processes. 67 The Reference Group was presented with the data described in the previous sections of this document. This was the first time any external partners had had the opportunity to consider such in-depth data on the profile of the Matrix cohort, which clearly evidences significant disproportionality and which some of the Group believed supports the findings of some of the more qualitative research that has been produced by Amnesty and Stopwatch, for example. The Reference Group expressed the view that there would be difficulties in building community confidence in any tool, which as described by one of the Group, is considered by some to be racially-loaded, ineffective and adding to the actual problem. However, some of the Group thought that police officers should be given training on conscious and unconscious bias and that this might impact on their use of the Matrix. The Group expressed several views on what is driving the disproportionality; seeing the wider structural inequalities as a key issue. The relationship between the police and young black people was also seen as a key driver of the violence we are now seeing in our communities. One member was of the view that young black people are feeling targeted by the police and therefore even the ‘good kids’ are arming themselves for protection and drifting towards groups involved in criminality. There were others in the Group that disagreed that this was a factor in weapon-carrying. Some members of the Group expressed concerns about the perceived “racialised nature of the word ‘gangs’,” and felt that this language was labelling all groups of young black men as potential criminals. However, others pointed out that this was very much a London perspective and pointed to Asian youths, white youths and Eastern European youths who may also be affiliated to gangs, and that the composition of groups considered as ‘gangs’ in other parts of the country would differ to reflect the population demographic accordingly. Views were also expressed about the wider context of policing. The Reference Group agreed that the MPS needs to be more diverse in order to have real credibility with communities and to address what they felt to be a lack of cultural awareness. They also felt that there needed to be a greater emphasis on community policing and engagement with young people, rather than enforcement. The Group felt that the development of the Violence Reduction Unit would present an opportunity for a refreshed conversation about community engagement and for the community to be able to contribute to the response to violence. 17 October 2018 – The impact findings were shared and discussed with the Group along with draft recommendations. The Group were asked for their views on how the police could 68 be more answerable to the communities they serve in relation to the Matrix and discussions took place on what mechanisms could be used or created to increase community confidence in developments that occur in response to published recommendations. The Group were aware of the statutory duties under the Crime and Disorder Act (1998), which are attributable to several agencies, including the local authority, public health and the police, and that sharing data is part of this. As such, they were of the view that it should be the collective responsibility of these agencies to ensure Matrix data is verifiable, contemporary, and shared appropriately, and that Community Safety Partnerships, or other appropriate partnership structures, should have oversight of the local Matrix data, with built-in opportunities to review the approach. The Group recognised that the processes and governance of the Matrix differed significantly to that of the Multi-Agency Public Protection Arrangements (MAPPA) that are used to manage high risk offenders and suggested that MAPPA good practice should be applied to the Matrix. The Reference Group had a detailed discussion about the categorisation of people on the Matrix (i.e. Red, Amber, Green ratings). Their view was that there needed to be more clarity about the people who have been categorised as ‘Green’, because this group would include, for example, those at risk with possibly minor offences and others with convictions for violence who simply have a low rating because they are in custody. Clearly these two different groups of people need different approaches, so the Group felt that, where interventions were warranted, agencies other than the police should be leading on working with some of those rated ‘Green’. It was also expressed that, to further examine levels of disproportionality, more detailed data analysis was perhaps required in order to drill down to a local community or ward level to more carefully compare the ethnic composition of the local community with the ethnicity of local youths appearing on their local borough Matrix. The Group felt that the MPS needed to recognise the nature of the vulnerability to violence of individuals on the Matrix as well as the violence they are involved in perpetrating, and to respond accordingly. They suggested that the MPS consider a different approach to children with low harm scores on the Matrix - mostly young black boys - moving away from criminalising them and towards safeguarding them, an approach more aligned with a public health approach to tackling violence. In order to increase public confidence in the Matrix and to provide assurance that the Matrix is fair, transparent, and human rights compliant, the Reference Group felt that there should be an independent oversight body to oversee the Matrix. 69 More broadly, the Group were strongly of the view that, in addition to compliance with Human Rights legislation, there was a need to ensure that the Matrix is also compliant with equalities legislation. They expressed the view that an Equality Impact Assessment should be conducted and would provide a mechanism for ensuring that the issues discussed above were properly considered and appropriate mitigations against any negative impact put in place. On 1 November 2018, the Reference Group met with the Deputy Mayor for Policing and Crime, Sophie Linden, so that she could take them through the draft recommendations and hear their feedback directly. The draft recommendations were themed in relation to the purpose, operation and transparency of the Matrix and governance and oversight. The Group reiterated some of their views that had been expressed at previous meetings, including their concerns about the legality of the Matrix, the disproportionality evidenced in the data, the need for independent oversight of the Matrix and their concerns about the use and application of the ‘gang’ label. The Group felt that in order to secure community confidence in the tool, the recommendations would need to be specific and actionable, with agreed timescales and lead officers to ensure effective accountability for delivery of the necessary improvements. The Group acknowledged that the proposed recommendations addressed many of the key issues and were particularly positive about the focus on governance and transparency, including data management. In addition, the Reference Group welcomed the proposed recommendations relating to the way in which those banded ‘Green’ with zero-harm scores are dealt with. The Group expressed a strong view that the Review would be strengthened with the inclusion of a specific recommendation (or recommendations) relating to the issues of disproportionality - which is evidenced in the data, potential discrimination and the application of equalities and human rights legislation in the context of the Matrix, including specific reference to conducting an equality impact assessment. The Reference Group further suggested that, if the Matrix is to be used as a partnership tool, then there should be a clear partnership governance structure at the local level and this should be reflected in the recommendations. Looking to the future, it was recognised that the Review does not present an evaluation of the efficacy of the Matrix and the Group suggested that MOPAC ought to consider commissioning such an evaluation in the future. Engagement with young people, parents and community groups In addition to the Reference Group and the views of the voluntary and community sector collected through the surveys, we made efforts to conduct targeted engagement activities in communities across London. We were also cognisant of the personal accounts represented in the Amnesty and StopWatch reports on the Matrix, which identified some of 70 the same concerns. The views of the Reference Group were in many cases echoed by the community members we spoke with, particularly in relation to transparency, the need for oversight and clearer communication of the purpose of the Matrix. Engagement events were held in Southwark, Waltham Forest, Haringey and Westminster and we met with 110 young people, parents and community members. These groups included a mixture of those who had direct and indirect experience of the Matrix and those with no involvement at all. This was not a representative sample of all Londoners but did enable us to add further qualitative information to the analytical work already undertaken. These discussions included significant representation from the BAME (predominantly black African Caribbean) community (83%), 62% of those engaged were male and 90% of those present were aged between 11 and 25. In all sessions, strong views were expressed about the current violence taking place across London and the need for effective ways to stop it. There was an acceptance that there could be a role for the use of tools like the Matrix, but that that needed to be fair, proportionate and appropriately targeted. Many raised concerns about racial disproportionality and unconscious bias and how that might play into the use of the Matrix. Those engaged with were also very aware of the wider structural problems facing their communities which have a cumulative impact, such as youth unemployment and school exclusions. There were common themes that emerged across all the engagement events and these are discussed in more detail below. Concerns about transparency and a lack of understanding of what the Matrix is and how it is intended to work. For example, people specifically wanted to know how an individual would know they are on the Matrix and what they would have to do to be removed from it. There was a clearly-held view that individuals should have the right to know that they are on the Matrix because it is their personal information and also so that they could either challenge that status or take steps to change their behaviour as appropriate. There was also a strong view that a parent/guardian should be informed if a child (under 18) is added to the Matrix. This was also linked to the view that there should be interventions to help people on the Matrix and that the public should know what they are and how to access them. One person expressed their view on this very clearly – “Some might think it's a problem if you tell people, but if it's a harm reduction tool then notifying people will modify their behaviour and reduce harm. By not telling people they’re 71 taking a negative approach, which is intended to catch people out and put them in jail rather than help them to stop their behaviour.” One of the young people we heard from had specifically experienced the police taking a more positive approach to dealing with his behaviour, with the police having warned him that he was “mixing with the wrong people” when he had been found associating with people he presumed to be on the Matrix. Data sharing was also an area of concern in terms of how and with whom data is shared. The lasting and damaging impact this can have on people's lives was also raised, as well as the impact on communities’ trust in the police and other authorities. For example, we were told by some young people, or through youth workers that were supporting them, about personal experiences such as being told they were stopped and searched because they were on the Matrix, Matrix information being shared leading to job offers being withdrawn and a housing move being blocked. In one case we were told of an occasion where a young offender’s Matrix status had been shared with an adult associate under the supervision of Probation services. The feedback on the use of the term ‘gangs’ was that it is problematic in two respects and that perhaps it should be re-labelled to focus on violence. Firstly, many felt that those labelled as being in gangs should be active offenders, not just a large group of young people who associate with each other in local areas. Some also questioned why knowing someone who had committed offences should mean that you could, in their words, “be relentlessly targeted for police activity”. One young person said: “Even schools are calling us gangs when we hang out as a large group of school friends - they call us 'thugs'. How we see gangs is different to non-black people.” Secondly, some expressed the view that the ‘right’ people are not on the Matrix and that it was being used to pursue joint enterprise convictions. We heard views that there are “more serious people” that need to be on the Matrix and that the police would have better intelligence if they worked with the community. The matters of oversight and accountability were also discussed and there was consensus on the need for more transparency, with some holding the view that there should be independent scrutiny of the Matrix and how it is used. 72 Transparency Matrix practitioners generally felt that ‘gang-affected communities’ were not aware of the Matrix and those that were did not understand its purpose. Most officers did not perceive the existence of the Matrix as having any negative impact on community relations, also suggesting there was no need for the community to be consulted. Furthermore, other tools used by the MPS to track and tackle criminality – such as the OCG tracker - are not well understood by the wider community and we heard through our engagement that the public wanted to understand this better. However, responses illustrated the varied opinions in terms of transparency. Some argued a more transparent approach might increase public fear of gangs, or give the wrong impression of the extent of the gang problem: “Being a non-gang borough, any suggestion of the Gangs Matrix or its use locally would potentially paint a misleading picture concerning gang activity locally.” (Police Officer) Others saw possible benefits to community relations in terms of legitimising police action; “If the Matrix became accountable and open to all, it could be a useful tool in police and the community targeting gang members, especially if assisted in grounds for stop & search, warrants, and stiffer sentencing.” (Police Officer) Local authority practitioners were more open to community consultation and transparency; “I think there is a huge gap and that if we were to engage communities in the process we would have a far richer intelligence picture.” (Local Authority Practitioner) For VCS practitioners, there was there was strong consensus that the police could both be more transparent in communicating the purpose of the Matrix, and that they could do more to actively engage with communities most affected by serious violence. “There MUST be integration between the police, education services, health services, employment strategies, faith groups, local authorities, voluntary organisations. As the Commissioner said, we cannot police ourselves out of this violence…” (VCS Practitioner) A significantly greater proportion of local authority practitioners believed that the MPS could be more transparent in how it targets organised crime groups. Police, local authority and VCS respondents all indicated common local community perceptions that the Matrix is 73 used to target ‘low hanging fruit’; a continuing community concern highlighted within previous gang intervention projects and emphasised by the Reference Group. 74 Appendices 75 Appendix 1 - Legal Assessment This is the legal assessment commissioned by MOPAC from Tim Pitt-Payne QC, 11KBW. Issues raised 1. I have been instructed by Transport for London Legal, on behalf of the Mayor’s Office for Policing and Crime (“MOPAC”). I have been asked to assess the legality of the Gangs Matrix, a database operated by the Metropolitan Police Service (“MPS”). The data controller for the Gangs Matrix is the Commissioner of Police of the Metropolis (“the Commissioner”). 2. This document discusses: • issues arising under the Human Rights Act 1998 (“HRA 1998”) in relation to Articles 8 and 14 of the European Convention on Human Rights (“ECHR”); • data protection issues; and • RIPA issues, i.e. issues under the Regulation of Investigatory Powers Act 2000. 3. My overall assessment is that there are no legal objections which require the Gangs Matrix to be abandoned altogether, or radically recast. But there are respects in which its operation should be modified in order to ensure that all applicable legal requirements are met. 4. In addition to the points made in this document, any views expressed by the Information Commissioner’s Office (“ICO”) and any regulatory intervention by the ICO will need to be taken into account by MOPAC and the MPS. Human Rights Act issues 5. HRA 1998, section 6(1), provides that it is unlawful for a public authority to act in a way which is incompatible with a Convention right. This general duty would apply both to the Commissioner and to MOPAC. In relation to the operation of the Gangs Matrix, the issue is likely to be whether the Commissioner is acting in breach of this duty. 6. Article 8 of the ECHR provides as follows: (1) Everyone has the right to respect for his private and family life, his home and his correspondence. (2) There shall be no interference by a public authority with the exercise of this right except such as is in accordance with the law and is necessary in a democratic 76 society in the interests of national security, public safety or the economic well- being of the country, for the prevention of disorder or crime, for the protection of health or morals, or for the protection of the rights and freedoms of others. 7. There is no doubt that the creation and maintenance by the Commissioner of a database such as the Gangs Matrix involves an interference with the right to respect for private life under Article 8(1). The question is whether that interference can be justified under Article 8(2). This requires consideration of the following: • Whether the interference is for one of the purposes set out in Article 8(2). • Whether the interference is “in accordance with the law”. • Whether the interference is “necessary”, which in turn gives rise to issues about proportionality. 8. The purpose of the Gangs Matrix clearly falls within Article 8(2), as being for the prevention of disorder or crime. 9. As to whether any interference is in accordance with the law, this requires that the interference must have a basis in domestic law. It also imposes requirements as to the quality of that law, which must be compatible with the rule of law. There must be sufficient protection against arbitrary interference with the rights safeguarded by Article 8(1). The law must be sufficiently clear to give citizens an adequate indication of the circumstances and conditions in which public authorities are empowered to interfere with their rights. 10. In relation to the operation of the Gangs Matrix, the question of whether this is “in accordance with the law” essentially raises two issues: • whether the Gangs Matrix operates in a manner that is arbitrary and over-broad; and • whether there is an adequate publicly-available framework of law and policy governing the operation of the Gangs Matrix. 11. As to the first issue, the decision as to whether someone is to be listed on the Gangs Matrix depends on there being reliable intelligence about their gang membership from more than one source. Once they are listed, their individual status (which currently may be classified as being red, amber or green) depends on the application of a complex scoring system. It does not seem to me that the definition of “gang” adopted by the MPS, or the scoring system used, is so inherently arbitrary as not to be “in accordance with the law”. 77 12. As to the second issue, Gangs Matrix does not have a specific statutory basis; it is based on the common law powers of the police to obtain and store information for policing purposes. In itself, this does not prevent the Gangs Matrix from being “in accordance with the law”. However, consideration should be given as to what further information can be made public about the Gangs Matrix and the way in which it operates. 13. I suggest that there should be a public-facing document covering topics such as the following: • The purpose of the Gangs Matrix, including: encouraging individuals to divert from gang membership; managing the risks presented by the individuals listed; and managing the risk that those individuals will themselves be victims of violent crime. • The criteria for inclusion on the Gangs Matrix. • The basis on which an individual is scored. • The practical consequences of being listed on the Gangs Matrix with a particular score. • The circumstances in which an individual’s listing will be changed, or in which an individual will be removed from the Gangs Matrix altogether. • The arrangements for sharing information from the Gangs Matrix on a London- wide level. • The arrangements for sharing information at borough level. 14. I turn next to the question whether the operation of the Gangs Matrix is proportionate. The requirements in relation to proportionality are usefully summarised in Bank Mellat v HM Treasury [2013] UKSC 39, as follows (see at §74): it is necessary to determine (1) whether the objective of the measure is sufficiently important to justify the limitation of a protected right, (2) whether the measure is rationally connected to the objective, (3) whether a less intrusive measure could have been used without unacceptably compromising the achievement of the objective, and (4) whether, balancing the severity of the measure’s effects on the rights of the persons to whom it applies against the importance of the objective, to the extent that the measure will contribute to its achievement, the former outweighs the latter. 15. In relation to proportionality in the operation of the Gangs Matrix, in my view the key issues are: • whether the Gangs Matrix includes too many individuals; and • whether information is shared too widely about the individuals listed. 78 16. As to the first question, the most important consideration is whether the inclusion of individuals who have a very low score (or a zero score) can be justified. 17. In order to ensure that the Gangs Matrix operates in a proportionate way, I would suggest that the following issues need to be reviewed. • Can individuals be listed on the Gangs Matrix for the first time, even if they have a zero score? If so, why is this appropriate? • Once the score of an individual drops to zero, will they be automatically removed from the Gangs Matrix? If not, why not? • Whatever answers are adopted to these questions, do the relevant policies, training materials, and any public-facing information, all consistently reflect those answers? 18. In relation to the sharing of information, the main area of concern is whether information is being shared at borough level in an inconsistent or excessive way. The following questions need to be addressed. • Whether there are information sharing agreements in place at borough level, across all boroughs where the Gang Matrix is in operation. • Whether these agreements properly reflect the way in which the Gangs Matrix is intended to operate London-wide. • Whether these agreements are consistent with one another. 19. Overall, in my view the Gangs Matrix is in principle capable of being operated consistently with Article 8; but the issues referred to above need to be addressed in order to ensure this is the case. 20. I turn next to Article 14 of the ECHR. This provides as follows: The enjoyment of the rights and freedoms set forth in this Convention shall be secured without discrimination on any ground such as sex, race, colour, language, religion, political or other opinion, national or social origin, association with a national minority, property, birth or other status. Article 14 does not provide for a free-standing right not to suffer discrimination. Rather, it relates to discrimination in connection with the enjoyment of the Convention rights themselves. Article 14 will therefore always need to be considered in conjunction with one or more of the other articles in the Convention. In the context of the Gangs Matrix, Article 14 will need to be considered in conjunction with Article 8. 79 21. The Gangs Matrix has been criticised as having a disproportionate impact on the black or black and minority ethnic (BAME) population of London; the criticism is that black or BAME individuals are disproportionately represented on the Gangs Matrix, as compared with their share of London’s population as a whole. The issue in relation to Article 14 is whether the Gangs Matrix discriminates - as between black and/or BAME Londoners, and Londoners as a whole - in relation to the enjoyment of the Article 8 right to privacy. 22. As a general point, it is important to bear in mind that individuals included on the Gangs Matrix are likely to be at risk both of committing violent offences, and of being victims of such offences. Inclusion provides an opportunity to divert individuals from gang membership, thereby reducing both risks. In assessing the impact of inclusion on the Gangs Matrix, it is important to take account of this potential protective effect, as well as any adverse impacts. This is significant when considering the impact of the Gangs Matrix on different racial groups: I understand that, across London, young black males are disproportionately represented both as victims and as offenders in all serious violence. 23. It does not seem to me that the concept of a “gang” by reference to which the Gangs Matrix operates is inherently race-based, or that it is so vague as to provide an obvious opening for racial basis. Nor is there anything in the material that I have seen to support an assertion that individuals can be listed on the Gangs Matrix merely because of they identify with elements of black street culture. 24. I have seen statistics that compare the following, at borough level: (i) the black or BAME percentage of the total population; (ii) the percentage of those charged with knife crime or knife crime with injury who are black or BAME; and (iii) the percentage of those included on the Gangs Matrix who are black or BAME. By reference to these measures, there is some disproportionality in the black or BAME membership of the Gangs Matrix, at a level that varies significantly from borough to borough. That is to say, the proportion of those on the Gangs Matrix who are black or BAME is greater than: (i) the proportion of black or BAME individuals in the under 25 population; or (ii) the proportion of black or BAME individuals among those charged with knife crime or knife crime with injury. The reasons for the disparity are unclear, and MOPAC intends to conduct further research. To date, the MPS has not conducted an equality impact assessment in relation to the operation of the Gangs Matrix, but I understand that they have now committed to carrying this out. 25. On the basis of the material that I have seen, I think it would be difficult to bring a successful legal challenge to the operation of Gangs Matrix under Article 14. That said, 80 the statistical material referred to above clearly merits further investigation. It is extremely important that the Gangs Matrix should operate, and should be seen to operate, without any element of discrimination. I understand that consideration is being given to seeking assistance from the Equality and Human Rights Commission (EHRC) with these issues, including with carrying out an equality impact assessment. This is a helpful and positive step. In particular, it is important to keep under review – with the assistance of the EHRC – whether there are any material differences in the practical operation of the Gangs Matrix at borough level, and whether this plays any part in explaining the differences in black and BAME statistics as between different boroughs; this is an issue that should be considered as part of any equality impact assessment. Data Protection issues 26. With effect from 25th May 2018 the Data Protection Act 1998 was repealed and replaced by a new data protection regime set out in: the General Data Protection Regulation (“GDPR”); the Data Protection Act 2018 (“DPA 2018”); and Directive (EU) 2016/680 (“the Law Enforcement Directive”). 27. The new data protection law is complex. The main provisions are set out in the GDPR. DPA 2018 contains various supplementary provisions, and also gives effect to the Law Enforcement Directive: see Part 3 of DPA 2018. 28. As far as the Gangs Matrix is concerned, in my view this will fall within the scope of the Law Enforcement Directive, and hence Part 3 of DPA 2018. 29. An individual’s entry on the Gangs Matrix will constitute personal data about that individual. The holding of that data, its use for policing purposes, and its sharing with other organisations, will all constitute the processing of personal data. The data controller in relation to the processing will be the Commissioner. MOPAC itself does not seem to me to be either a data controller or a data processor in relation to the Gangs Matrix. 30. Chapter 2 of Part 3 of DPA 2018 sets out six data protection principles. The most significant for present purposes are the first and second principle. 31. The first principle requires that the processing of personal data for any of the law enforcement purposes must be lawful and fair. In order to satisfy this requirement, the processing must be based on one of two conditions: either the data subject must have given their consent, or the processing must be necessary for the performance of a task carried out for the law enforcement purpose by the competent authority. 81 Clearly, the consent condition will not be satisfied in relation to the Gangs Matrix, and so it is the second condition that is relevant. 32. The use of the term “necessary” in the context of a processing condition will carry with it a test of necessity and proportionality comparable to the test that would apply to an interference with a qualified ECHR right (such as the right under ECHR Article 8). What this means is that: (i) the test of necessity in this context would be a test of reasonable necessity rather than strict necessity; but (ii) the same considerations as were discussed above in relation to proportionality in the context of Article 8 of the ECHR, would apply in determining whether this processing condition was satisfied. 33. Under DPA 2018 section 35(4) and (5), there are more demanding conditions where the processing in question is sensitive processing. In this situation the following additional conditions would apply: • The processing must be strictly necessary for the law enforcement purpose. • It must meet one of the conditions in Schedule 8 to the DPA 2018. • The controller must have an appropriate policy document in place satisfying the requirements of section 42 of DPA 2018. 34. “Sensitive processing” for this purpose is defined in DPA 2018 section 35(8). It means the processing of: personal data revealing racial or ethnic origin, political opinions, religious or philosophical beliefs or trade union membership; the processing of genetic or biometric data, for the purpose of uniquely identifying an individual; the processing of data concerning health; and the processing of data concerning an individual’s sex life or sexual orientation. 35. To the extent that the Gangs Matrix involves sensitive processing, this will be subject to the more onerous requirements summarised above. A test of strict necessity would raise the same proportionality issues as were referred to above, but in addition would require the data controller to establish that the various law enforcement purposes pursued by the Gangs Matrix could not be achieved unless the relevant sensitive processing took place. 36. As to the Schedule 8 conditions, paragraph 1 would be relevant. This requires that the processing (a) is necessary for the exercise of a function conferred on a person by an enactment or a rule or law, and (b) is necessary for reasons of substantial public interest. The relevant “function” of the MPS would not be a statutory function, but rather its common law functions regarding the collecting of information for policing purposes. 82 37. The second data protection principle requires that personal data that is collected for a law enforcement purposes must not be processed in a manner incompatible with the purpose for which it was collected. This is subject to section 36(3) of DPA 2018: Personal data collected for a law enforcement purpose may be processed for any other law enforcement purpose (whether by the controller that collected the data or by another controller) provided that – (a) The controller is authorised by law to process the data for that purpose, and (b) The processing is necessary and proportionate to that other purpose. On this basis – provided again that the requirement of proportionality is met –policing data that was originally collected for some other purpose could lawfully be used in calculating individuals’ Gang Matrix scores. 38. Provided that the Gangs Matrix satisfies the requirements of ECHR Article 8, I think it is likely also to satisfy the requirements of the first and second data protection principles; issues about proportionality will be of central importance in both contexts. 39. Chapter 3 of Part 3 of DPA 2018 deals with the rights of data subjects. Section 44 requires data controllers to provide a range of information to data subjects (whether by making that information available to the public generally, or in some other way). The requirements of section 44 may be restricted where this is a necessary and proportionate measure to avoid prejudicing the law enforcement purposes: see section 44(4)(b). 40. Does section 44 require that individuals who are listed on the Gangs Matrix must be specifically informed of that fact? At first sight, this is what section 44 would seem to require; however, that is subject to the restriction in section 44(4)(b). Hence it is necessary to consider what would be the implications for the Gangs Matrix of informing individuals of their inclusion, and whether this would prejudice the relevant law enforcement purposes. 41. Even assuming that individuals do not need to be informed of their inclusion, the public generally should be given information about the operation of the Gangs Matrix: I would suggest that information should be made public along the lines discussed at paragraph 13 above, together with information that meets the various specific requirements of section 44(1) and (2) (to the extent that this can be done without prejudicing the law enforcement purposes of the Gangs Matrix). 83 42. Chapter 4 of Part 3 imposes a range of supplementary requirements on data controllers: for instance, various records of processing activities must be kept, under section 61. None of the provisions in this Chapter seem to me to provide any insuperable barrier to the operation of the Gangs Matrix, though it is important to check that all of these requirements have been met. 43. I should draw attention to section 64, requiring a data protection impact assessment (DPIA) to be carried out where a type of processing is likely to result in a high risk to the rights and freedoms of individuals. It would be highly desirable at this point for the MPS to carry out a DPIA in respect of the Gangs Matrix. 44. My overall assessment as regards data protection is as follows. • The requirement of proportionality – discussed under Article 8 above – is also an important aspect of the data protection framework. • To the extent that processing for the purposes of the Gangs Matrix is not proportionate, such processing will not satisfy the requirements of DPA 2018. • Otherwise, it seems to me that the requirements of DPA 2018 are in principle capable of being satisfied, though there are various specific issues identified above that will require careful consideration. RIPA issues 45. I have considered whether the use that is made of social media in connection with the operation of the Gangs Matrix, requires RIPA authorisation as constituting surveillance. A failure to obtain RIPA authorisation for surveillance falling within RIPA does not automatically mean that the surveillance is unlawful. However, in practice if surveillance falls within the scope of RIPA and authorisation is not obtained then there is a high risk that the relevant surveillance will be in breach of Article 8. 46. There are two types of authorisation that may be relevant: • authorisation for use of a covert human intelligence source (CHIS); and • authorisation for directed surveillance. 47. If the only direct interaction between a police officer and a person of interest is that the former sends a social media friend request (without disclosing that they are a police officer) and the latter accepts it, then I doubt if this will in itself constitute a sufficient relationship to require authorisation for use of a CHIS. However, if the social media interaction between the two individuals is more extensive then a different analysis might apply. 84 48. In general, I doubt whether the mere viewing of material that has been placed online by an individual and made publicly available by them would usually constitute directed surveillance for RIPA purposes. If the viewing was intensive and repeated in relation to a specific target individual then this might perhaps cross the line into being directed surveillance. Conclusion 49. My overall assessment is that there are no legal objections which require the Gangs Matrix to be abandoned altogether, or radically recast. However, there are a number of areas that need to be addressed, so as to ensure that the Gangs Matrix continues to operate lawfully now and in future. TIMOTHY PITT-PAYNE QC 11KBW 9th November 2018 85 Appendix 2 - Reference Group Members Invitees James Hayward Sarah Moran Information Commissioner’s Office Allan Hogarth Tanya O’Carroll Amnesty Sheridan Mangal Trident IAG Vice-Chair and works with young entrepreneurs Bobby Martin Community engagement & gang intervention advisor Mark Blake Black Training & Enterprise Group Stafford Scott The Monitoring Group Katrina Ffrench Chief Executive, StopWatch Simon Harding Associate Professor in Criminology, University of West London Commander Mark McEwan MPS, Commander for Community Engagement Officers Natasha Plummer (Chair) MOPAC (Engagement) Tom Davies MOPAC (Evidence & Insight) Jeanette Bain-Burnett GLA Community Engagement Team Caroline Tredwell MOPAC (Policy) 86 Appendix 3 - Our Equality Duty as a public body The Macpherson Report into the murder of Stephen Lawrence and the failures of the Metropolitan Police in the subsequent investigation led to a fundamental rethink of how public bodies address discrimination and racism. As a result, in 2001, the Race Equality Duty came into force, changing the focus for public bodies from simply avoiding discrimination to actively and positively promoting equality. Further duties on disability and gender equality followed. The Public Sector Equality Duty was created under the Equality Act 2010, superseding the existing duties and applying them across all protected characteristics - age, disability, gender reassignment, pregnancy and maternity, race, religion or belief, sex and sexual orientation. There are three main aims of the Duty. They are that in the exercising their functions, public bodies must have due regard to the need to: • Eliminate unlawful discrimination, harassment and victimisation and other conduct prohibited by the Act. • Advance equality of opportunity between people who share a protected characteristic and those who do not. • Foster good relations between people who share a protected characteristic and those who do not. MOPAC is committed to fulfilling these aims in all of its work. We address these issues in detail, as they relate to this Review, in an Integrated Impact Assessment published alongside it. 87 Appendix 4 - Matrix Review: Methodology This section provides an overview of the collection methods, cleaning, composition and analysis of the key data sets incorporated into the Matrix Review. As emphasised in the main section, data is always subject to caveats and this is particularly true of police- generated data which may be subject to distortion at various levels (i.e. local recording practices and priorities, police discretion etc. see for example: Kitsuse & Cicourel 1963; Bottomley & Coleman 1981; Coleman & Moynihan 1996) None-the-less, such collected data is the best picture of organisational knowledge available. Surveys As part of the Review, a number of on-line surveys were generated to capture the views of a range of groups. All surveys were disseminated via an email link and hosted by Opinion Research Services (ORS). Survey responses were marked-up and analysed thematically, with relevant quotes extracted in accordance with pre-defined and emergent themes. A breakdown of the surveys is below. 1) MPS Practitioner survey The key aim was to capture key MPS practitioners’ perceptions across a variety issues relating to the Matrix. In particular, the survey was aimed at the borough Single Points of Contact and Matrix leads and the aspiration was to get one or two responses from each borough. The survey consisted ten sections broadly covering topics such as general perceptions of gangs and gang crime; Matrix understanding and use; perceived impact; accuracy; local processes (inclusion, removal and oversight); information sharing and partnership working; community engagement; training; possible changes, benefits and drawbacks. This was distributed in autumn 2017 via a senior Metropolitan Police officer. A total of 88 responses across 28 boroughs, including 47 SPOCS and 41 other officers, predominantly of Constable and Sergeant rank were received. 2) Local Authority Practitioners This survey sought to capture the views of key local authority practitioners working with the Matrix and covered the same thematic areas as described above. The survey was distributed in autumn 2017 via Heads of Community Safety on all London boroughs. We received 45 valid responses were received across 27 boroughs from a wide variety of front line and leadership roles practitioners. 3) Voluntary and Community Sector This survey aimed to capture individuals working within Community and Voluntary Sector (VCS) organisations supporting young people involved in violence, again following the same 88 thematic structure. The survey was distributed via a variety of pathways in summer 2018 receiving 98 responses in total. Matrix Data Analysis This section of the methodology covers the variety of analytics conducted on the Matrix to understand the cohort over time, their criminal background and a number of measures around 'impact' such as upon offending, victimisation and stop and search. The Matrix over Time In order to understand the full picture of the Matrix, monthly snapshots of the Matrix were extracted for a 60-month (5 year) period between June 2013 and May 2018. In addition, an early iteration of the Matrix from September 2012 was also obtained. This covers the first cohort to a recent snapshot in 2018, enabling long-term analytics. All individuals from the 61 Matrix snapshots were combined and cleaned. As a result of cleaning the data, a small number of individuals were excluded from further analysis (i.e., inaccurate PNC number or duplicates). It is important to note that the Matrix has evolved substantially since introduction, with earlier snapshots including a handful of individuals without PNC IDs as well as duplicate entries. Snapshots taken from 2018 indicate that all individuals on the Matrix now have PNC IDs. A total of 8,272 unique PNC-IDs remained post cleaning, inclusive of the 2012 cohort. For the 60-month cohort, 7,556 remained. PNC Analysis, crime and impact upon reoffending PNC data was obtained for the 60-month cohort. The Police National Computer (PNC) records information relating to individuals for the use of the police and law enforcement, including details of all criminal convictions received. Of the 7,556 PNC records requested 7,299 were returned, with the date of extraction 01/05/2018. Following extensive cleaning and coding, overarching criminal careers analysis was conducted on the full dataset, with an emphasis on proven offending (i.e. sanctions as opposed to charges). Using this data enabled a full exploration of the criminal background and offences of those individuals on the Matrix. Offending records were standardised according to the month the offence was committed whether before (pre-), during or after (post-) removal from the Matrix. For impact analysis, a cut-off date of 31/10/2017 for PNC offending data was applied to allow time for sanctions to be recorded on the system, meaning 270 individuals were excluded from the analysis. One caveat to bear in mind when examining the results of the charts presented in the report is that this takes into account all individuals on the Matrix. This brings with it thousands of individuals, each with different start, end and lengths of time on the Matrix. So, the overall offending graphs represent the proportion of the cohort (n=7,129) who were sanctioned for an offence committed within a set period in time, that is in the first month after the matrix, and so on. In this way, the ‘in scope’ or available population gradually decreases the further 89 away from Matrix inclusion and removal time points. For example, whilst all 7,129 individuals had at least one month on the Matrix, only 6,585 had six months, and so on. Likewise, the number in the in-scope ‘post’ cohort is immediately reduced because over half of the cohort remained on the Matrix at the end date of analysis (n=3,253 have 1 month or more ‘post’). For this reason, results at the end of each period should be treated with caution given the low numbers of individuals involved; time periods on graphs were standardised to 4-year periods. Generating the Matrix comparison group As outlined in the main report, identifying, isolating and attributing 'impact' on offending to any given initiative (be it a police intervention or policy) is one of the most difficult aspects of evaluative research. The key question that underpins this is: what would have happened to individuals to have received a given initiative if they had not? This is called developing a counterfactual. For the Matrix review, robust exploration of impact necessitates this counterfactual or comparison group; that is, a group of individuals who present similar demographic and offending characteristics but have not been subject to the intervention in question (i.e. the Matrix). There are numerous experimental methods in developing such comparison groups or counterfactual, each with different levels of robustness. The most robust method and the only one that enables statements on causation is where eligible individuals are randomly assigned to receive the treatment or not and then compared on outcome measures (this is known as a Randomised Control Trial). See the MOPAC Randomised Control Trial on Body Worn Video for an example (Grossmith et al. 2016). In the case of the Matrix, individuals were not randomly assigned, nor would this have likely been ethically possible, and as such this method was not feasible. The next level of robustness in determining 'impact' would be the Quasi-experimental designs. Such designs utilise a range of statistical methods to generate the counterfactual and are seen as strong designs in evaluation research, although not able to prove causation. There are further evaluation designs below these but are not robust in exploring impact and were discounted in exploring impact. Another complication, specific to the Matrix is that of the pan-London nature. Comparison groups are often able to identify similar types of individual from geographic areas not receiving the intervention. With a pan London approach, such as the Matrix, this is not possible as all boroughs have a Matrix cohort. Furthermore, the most problematic or eligible individuals should have already been identified for the Matrix and so any comparison group of individuals identified within London would have inherent flaws. The same was judged to be the case in generating a comparison group drawing upon wider England and Wales data, especially the different working definitions and practices by police forces to identify 'gangs'. In the case of the Matrix Review, in order to explore impact a quasi-experimental approach was developed alongside staff at University College London in a method called ‘within-group reference point shuffle’. This method is inspired by a technique commonly used in the 90 spatio-temporal analysis of crime, called the Knox test (Knox, 1964). The test is used as a means of identifying space-time clustering; that is, the tendency of incidents to appear close to each other in space and time (commonly manifested as 'near-repeat' victimisation; see Johnson et al, 2007). In the test, incidents are first examined to establish how they are separated in observed data: how many occur within 400 metres and 7 days of each other, for example. Once this has been calculated, the test seeks to compare these values against what would be expected if no clustering was present. It does this using a permutation approach, in which the time-stamps of the events are repeatedly shuffled (so each incident is randomly assigned the time of a different incident). After each shuffle, the separation between incidents is calculated in exactly the same way as it was for the observed data, and this process is repeated many times. After doing this, the average separation under the shuffling procedure is compared with what was observed in the original data. The principle of the test is that the shuffling procedure replicates what would be expected if there was no relationship between the spatial and temporal dimensions of the data. The idea is that, if there was no dependence between the locations and times of incidents (i.e. no clustering) then the shuffling procedure ought to make no difference to the observed separation - any proximity is simply the result of random fluctuation. If, on the other hand, there is a discrepancy between the observed data and that obtained by shuffling, this implies that there was something 'special' about the spatio-temporal alignment in the observed data. One of the key properties of the technique is that it controls for the underlying spatial or temporal distribution in the data (i.e. the natural heterogeneity of crime). The approach used to explore the impact of the Matrix essentially substitutes the spatial aspect of the above technique for differences across individuals. Rather than examining the proximity between incidents, in this case the quantity of interest is the proximity of the intervention (e.g. Matrix entry) to sanction events. This is essentially what is being computed when examining sanction rates pre- and post- Matrix inclusion. The shuffling of time-stamps (in this case of Matrix inclusion) is then intended to break down any alignment between the point of Matrix entry and the offending history - just like the original test showed what would be expected if the locations and times were independent, this shows what would be expected if the time of matrix entry and offending history were independent. As before, any departure from this null/baseline distribution suggests a departure from this independence. The overall approach is innovative and reasonably robust. The work has been peer-reviewed by a number of academics at University College London. As with other quasi-experimental designs, statements of causation are not possible, but the approach is able to provide useful insights that could be built upon in further research. 91 Matrix Victimisation A list of Matrix PNC-IDs was sent to MPS Strategy & Governance, requesting details of all CRIS Notifiable Offences where the individual featured as Suspect, Accused or Victim between May 2013 and May 2018. For standardisation purposes, the period of analysis was refined to 01/01/2014 and 31/03/2018. This left 5,107 individuals with at least one of six months pre, during or post Matrix inclusion within the period of analysis; 92% of these (n=4,668) appeared at least once as victim, suspect or accused in this time. Counts of Accused (individuals charged), Suspect and Victim records before, during and after inclusion on the Matrix were calculated, and victimisation analysis conducted using the same methodology as described above. Time between offence and arrest Also referred to as ‘speed to arrest’, this method used the 'offence date' and 'arrest date' as recorded on the Police National Computer to derive an average number of days between offence and arrest. This is only a proxy measure of the offence date - but again is the most accurate data available. Stop and Search Stops data was obtained on the cohort (n=7,299) for the period June 2013 to May 2018, of these stops data was returned on 3,996. As detailed in the main report, individual level data relating to stop and search of Matrix individuals pre-2016 was assessed as incomplete due to changes in recording practices and data platforms. MPS data-matching for stops incidents utilises a variation of the ‘fuzzy matching’ process which is also prone to error. For this reason and to increase accuracy, the period of analysis was shortened to 01/01/2016 to 31/03/2018. This meant that 2,667 individuals were in scope. The methodology as described above for offending was replicated whereby the proportion of in-scope individuals stopped at each time point was calculated (three-month periods). As with offending, this was then split by various groups (harm status on entry, ethnicity, individuals not in custody during Matrix inclusion) An additional dataset of all individuals stopped and searched in the MPS over a one-year period was obtained (April 2017 – March 2018; 4,312 individuals appeared on the Matrix during this period). This allowed a comparison between non-Matrix and Matrix individuals being stopped. Individuals within the ‘stopped’ cohort were identified as being on the Matrix at the point of stop, as well as those who had been on the Matrix but were not on when stopped. Averages for Matrix and non-Matrix individuals were then calculated. Whilst overall average stops in a year for Matrix individuals appears low (2.4) this should be understood in terms of the caveats presented. To allow for fairer comparisons, and with previous findings on ethnic disproportionality in stop and search in mind, analysis also compared stops of Matrix and non-Matrix individuals within ethnic groups for the one-year time period. 92 Appendix 5 - Bibliography Amnesty (2018) Trapped in the Matrix: Secrecy, stigma, and bias in the Met’s Gangs Database. London: Amnesty International United Kingdom Section. Available from: https://www.amnesty.org.uk/files/2018- 05/Trapped%20in%20the%20Matrix%20Amnesty%20report.pdf?HSxuOpdpZW_8neOqHt_K xu1DKk_gHtSL= Bottomley, A.K. & Coleman, C. (1981) Understanding crime rates. Farnborough: Gower. 208: 70–96. Caudill, J., Trulson, C., Marquart, J., DeLisi, M. (2014) On Gang Affiliation, Gang Databases, and Prosecutorial Outcomes. Crime & Delinquency. 63(2): 210-229. Chesney-Lind, M. et al. (1994) Gangs and delinquency. Crime, Law and Social Change. 21(3): 201–228. Coleman, C. & Moynihan, J. (1996) Understanding Crime Data: Haunted by the dark figure. Buckingham: Open University Press. Densley J. & Jones D. (2016) Pulling Levers on Gang Violence in London and St. Paul. In: Maxson C. & Esbensen, F.A. (eds.) Gang Transitions and Transformations in an International Context. Cham: Springer Esbensen, F.A., Winfree Jr., L.T., He, N. & Taylor, T.J. (2001) Youth gangs and definitional issues: When is a gang a gang, and why does it matter? Crime & delinquency. 47(1): 105– 130. Ericson, R.V. & Haggerty, K.D. (1997) Policing the risk society. Oxford: Oxford University Press. Davies, T., Grossmith, L. & Dawson, P. (2017) Group Violence Intervention London: An Evaluation of the Shield Pilot. London: Mayor's Office for Policing and Crime. Available at: https://www.london.gov.uk/sites/default/files/gvi_london_evaluation270117.pdf Densley, J. (2011) Ganging up on gangs: Why the gang intervention industry needs an intervention. The British Journal of Forensic Practice. 13(1): 12–23. Disley, E. & Liddle, M. (2016) Local perspectives in Ending Gang and Youth Violence Areas. London: Home Office. 93 Farrington, D. (1992) Criminal Career Research in the United Kingdom. British Journal of Criminology. 32(4). 521-536. Farrington, D., Ttofi, M., Crago, R. & Coid, J. (2014) Prevalence, frequency, onset, desistance and criminal career duration in self-reports compared with official records. Criminal Behaviour and Mental Health. 24: 241–253. Grossmith, L, Owens, C, Finn, W., Mann, D., Davies, T. & Baika, L. (2015) Police, Camera, Evidence: London’s cluster randomised controlled trial of Body Worn Video. London: College of Policing and Mayor’s Office for Policing and Crime Hallsworth, S. (2013) The gang and beyond: Interpreting violent street worlds. Hampshire: Palgrave Macmillan. Jacobs, J.B. (2009) Gang databases. Criminology & Public Policy. 8(4): 705–709. Johnson, S. D., Bernasco, W., Bowers, K. J., Elffers, H., Ratcliffe, J., Rengert, G., & Townsley, M. (2007). Space–Time Patterns of Risk: A Cross National Assessment of Residential Burglary Victimization. Journal of Quantitative Criminology, 23(3), 201–219. Joseph, I. & Gunter A. (2011) Gangs Revisited: What’s a Gang and What’s Race Got To Do With It? London: Runnymede. Katz, C.M., Webb, V.J. & Schaefer, D.R. (2000) The validity of police gang intelligence lists: Examining differences in delinquency between documented gang members and nondocumented delinquent youth. Police quarterly. 3(4): 413. Kennedy, D.M. (2009) Gangs and public policy. Criminology & Public Policy. 8(4): 711–716. Kitsuse, J.I. & Cicourel, A.V. (1963) A note on the uses of official statistics. Social problems. 11(2): 131–139. Klein, M.W. (2009) Street gang databases. Criminology & Public Policy. 8(4): 717–721. Knox, G. (1964). Epidemiology of Childhood Leukaemia in Northumberland and Durham. British Journal of Preventive and Social Medicine, 18(1), 17–24. McCorkle, R.C. & Miethe, T.D. (1998) The political and organizational response to gangs: An examination of a “moral panic” in Nevada. Justice quarterly. 15(1): 41–64. 94 Medina, J., Aldridge, J. & Ralphs, R. (2009) Youth gangs in the UK: context, evolution and violence. Paper presented to the Global Gangs Workshop. 14-15 May 2009. Graduate Institute, Geneva. Pitts, J. (2017) Whatever happened to the family? England’s new gang strategy. Safer Communities. 16(1): 32–40. Pyrooz, D.C., Wolfe, S.E. & Spohn, C. (2011) Gang-Related Homicide Charging Decisions: The Implementation of a Specialized Prosecution Unit in Los Angeles. Criminal Justice Policy Review. 22(1): 3-26. Short, J.F. (2009) Gangs, law enforcement, and the academy. Criminology & Public Policy. 8(4): 723–731. Shute, J. & Medina, J. (2014) Hunting Gruffalo: “gangs”, unreason and the big bad coalition. Criminal Justice Matters. 96(1): 26–27. Smithson, H., Ralphs, R. & Williams, P., (2013). Used and Abused: The Problematic Usage of Gang Terminology in the United Kingdom and Its Implications for Ethnic Minority Youth. British Journal of Criminology. 53(1): 113–128. Spergel, I. (1995) The youth gang problem: A community approach. Oxford: Oxford University Press. Spergel, I. (2009) Gang databases: To be or not to be. Criminology & Public Policy. 8(4): 667– 674. Stanko, E and Dawson, P. Reflections on the Offending Histories of those Arrested during the Disorder. Policing: A Journal of Policy and Practice. 7(1), 3-11 Tyler, T. R. (2006). Why people obey the law. Princeton, N.J: Princeton University Press Williams, P. (2018) Being Matrixed: The (Over)Policing of Gang Suspects in London. London: StopWatch Available from: http://www.stop- watch.org/uploads/documents/Being_Matrixed.pdf Williams, P. & Clarke, B. (2016) Dangerous associations: Joint enterprise, gangs and racism. London: Centre for crime and Justice Studies. GPS Knife Crime Tagging Final Evaluation Report January 2024 Tim Read, Barry Charleton, & Laura Duckworth. MOPAC Evidence and Insight 2 Contents Contents .................................................................................................................................................. 2 Executive summary ................................................................................................................................. 4 London’s approach to GPS Tagging of Knife Crime Offenders................................................................ 8 GPS Knife Crime Tagging: Service Design ................................................................................................ 9 Evaluation ............................................................................................................................................. 13 Results from Service Delivery: Performance Insights ........................................................................... 17 Results from Service Delivery: Process Insights .................................................................................... 20 Exploring the benefits of the GPS tag ................................................................................................... 25 Discussion.............................................................................................................................................. 33 References ............................................................................................................................................ 37 Appendices ............................................................................................................................................ 39 3 Acknowledgements Thanks to Roxanne Doyle for her assistance with the transcription and analysis of offender manager and service user interviews. Thanks to our colleagues at E&I Abi McNeill and Lynne Conroy for their assistance. Thanks to Dr Krisztian Posch from University College London for his advice on propensity score matching, and to Professor Olivier Marie from Erasmus University, Rotterdam for acting as peer reviewer for this report. Thanks to Tom Dodsworth, Laura Norton and Marcus O’Halloran for their work in running the GPS knife crime pilot and the provision of data from the pilot. 4 Executive summary London’s approach to GPS Tagging of Knife Crime Offenders The Mayor’s Office for Policing And Crime (MOPAC) has been at the forefront in the use of electronic monitoring - first piloting Alcohol Abstinence Monitoring Requirement (AAMR) tags in 2014; then GPS tags in 8 London boroughs with persistent and knife crime offenders on community sentences between 2017-2019; before moving on to the current pilot from February 2019, initially in 4 boroughs, then pan-London later the same year with individuals who had committed knife-related offences released from prison on licence. This final report details the performance, process and impact evaluation of the extended GPS knife crime tagging pilot, covering over a three and a half-year implementation period (Feb 2019 – November 2022). Use of the GPS knife crime tag • Between the pilot’s start in February 2019 and the date of analysis (November 2022), a total of 971 GPS tags had been imposed with 881 individuals - 799 people had been tagged once, with 82 tagged more than once. • Three-quarters of tags (n=726) were imposed on release from prison, the remainder (n=243) as part of a licence variation for those already in the community. • The mean planned tagging period was 138 days, with a minimum of 10 days and a maximum of 365. • All bar one of the tag wearers (n=970) were subject to trail monitoring and over half (54% n=525) were subject to at least one exclusion zone. Characteristics of those on the knife crime tag • Almost all tag wearers were male (98%, n=948), with a mean age of 26.6 years and over half (52%, n=453) are from a Black or Black British background20. In fact, Black or Black British individuals aged under 25 years made up 30% (n=258) of tag wearers, increasing to 32%, (n=309) when including individuals from Mixed White and Black ethnic background. • White tag wearers tended to be older compared to the overall cohort, with 30% (n=59) aged 35 years or over, compared to 16% (n=157) of tag wearers overall, and 12% (n=55) of those from a Black or Black British background. • The mean OGRS1 score (2 year) of tag wearers was 62%, which places the group on average to medium risk of offending (range 4% to 97%). Older tag wearers tended to have a lower OGRS score, with the mean score of those aged 45+ years being 49%. Compliance with the GPS tag • Of the 971 tags imposed by November 2022, 835 had been completed2. Of these 537 ended without recall, yielding a successful completion of 64% overall. 1 The Offender Group Reconviction Scale (OGRS) is a predictor of re-offending based only on static risks – age, gender and criminal history. Howard et al (2009). 2 136 cases were still live at the time of analysis. 5 • Within the 835, there were a total of 298 (36%) that ended due to a recall. However, half (n=153, 51%) were for reasons unrelated to the GPS tag (i.e., missed probation appointment or new incident). • When the research considers the reasons for recall, the effective compliance rate, that is the proportion of cases completed without recall for failure to comply with the GPS tag as a licence condition would be 98% (n=815). Learning from service delivery • The general view from Probation Practitioners (PPs) was the operational delivery and implementation of the pilot improved as the wider London rollout occurred. Examples included improved use of GPS data for case management; not missing eligible cases; better management of risky areas for service users with a gang affiliation; better information provided to probation practitioners by Buddi (the tag provider); and improved guidance to PPs on faulty tags. • Broadly speaking, partnership working functioned well, although there were some examples of differences of opinion across agencies, such as eligibility for the tag or who should be the lead provider. For example, MPS officers expected to see more ‘gang’ related nominals or those with higher OGRS scores and queried the restrictions around access to data. • Most service users who responded to a survey agreed they understood why they received the tag (92%, n=607); knew how to comply (98%, n=645); and how the GPS requirement worked (93%, n=612). However, a small number of individuals reported that they had not been told in advance that the tag was being fitted and how it worked. • For the 964 cases where the location of fitting was recorded, under half (49%, n=476) were at prison, with the remainder at probation (47%, n=454), or another approved premises (4%, n=34). There were a variety of explanations for this – some planned (the distance of certain prisons from London which made fitting there impractical, and those tagged via licence variation who were already in the community), a minority of non-prison fittings were unplanned (for example when an initial prison fitting had not been possible due to issues such as prison access or availability of field staff). • Most surveyed users agreed the tag was comfortable to wear (76%, n=161) although individuals interviewed outlined a different picture, noting it was too bulky and raising other issues, such as charging. Analysis undertaken by Buddi suggests tagging equipment is replaced in around 2% of cases. 6 Exploring the benefits of the GPS tag • Probation Practitioners were overwhelmingly positive about the GPS tag. They identified a range of benefits such as provision of ‘structure and discipline’; enhanced case management; aided supervision; potentially having a deterrence effect and so on. The belief that the tag positively affects offending behaviour was also shared, but there was uncertainty regarding long-term behaviour change. • The majority (60%, n=of 382) of tag wearers who completed the end-point survey were on the positive side regarding their overall experience of the tag. The majority also reported they felt the tag had a positive impact on their offending behaviour (83%, n=529) and towards their life in general (60%, n=382). Two areas, that of ‘financial situation’ and ‘employment situation’ were identified as being ‘worse’ following the tag. Whilst positive, there may be a level of bias here as these individuals not only completed the tag, but they were also engaged enough to complete a survey on its use. • The crime mapping aspect did not have the expected level of influence. Many crimes were uploaded but only resulted in a small number of ‘matches’ (i.e., a link between an individual and crime location), and led to just two additional prosecutions. More work on assessing this aspect would be essential to maximise the value of the technology. • Understanding impact of any initiative is complex, and this is no different for the GPS tag. A variety of approaches were adopted, and analysis undertaken, including predicted versus actual offending comparisons using OGRS; comparisons of offending from the GPS tagged group across time points; and extending these comparisons to a Control group, constructed through Propensity Score Matching (PSM). The time points included offending before the order; on the order (unique to the individual); then fixed periods after the order start date; and after the order end date of 30, 60, 90 days, 12 months, and 24 months. As the Control group were not standardly subject to ‘something’ following prison release, the most useful comparison is from ‘point zero’ (when the GPS individual was tagged, or their match was released from prison) to the standardised number of days – either before or after. • Analysis suggested that GPS tags positively affect offending behaviour, although the full impact cannot yet be robustly determined. This is demonstrated by: - Predicted versus actual offending analysis which indicated the GPS and control groups’ actual offending was lower than their OGRS2 estimates would predict – with both groups reducing from a score of 63%. However, the GPS group (54%) was significantly3 lower than the Control (59%). - Across the time periods the proportion of individuals who reoffend in the GPS and Control groups were very similar – starting at 61% for both groups 24 months before point zero; reducing sharply 30 days after point zero; then slowly increasing over the next 2 years, although 24 months after point zero neither group have returned to their ’24 months before’ levels of offending. 3 p=0.28, Pearson Chi-Square 4.816 7 - Although the GPS group was always lower than the control in terms of the proportion to offend; the number of offences; and the mean number of offences, these were not significantly different within 30-, 60-, and 90-days post point zero, however significant differences were observed at 12- and 24-months. - At 12 months post point zero, for individuals with available data (n=470), the number and mean number of offences for the GPS group are significantly lower to the control (291 vs 4044; and 1.6 vs 2.0 respectively). However, the proportion who offend is not significantly different (38% GPS versus 42% Control). - At 24 months post point zero, for individuals with data still available (n=244), there are significant differences between the GPS group versus the control in terms of the proportion to offend (54% vs 59%5); the number of offences (258 vs. 4036) and mean number of offences (2.0 vs. 2.8). • Overall, there are broadly positive results regarding the impact on offending of the tag. However, caveats within the methodological design (i.e., historical comparison group, and the possible influence of Covid) preclude firm statements of impact. Conclusion • The evaluation of GPS use has spanned a period of 6 years with 5 published reports, prior to this one7. This report explores performance, process, and impact, seeking to bring the evaluation to a conclusion. • Findings indicate the scheme has a high level of compliance; that delivery improved over the duration of the programme; and that both staff and tag users were generally positive about the technology, and whilst there were positive indications as to the impact on offending, methodological limitations preclude firm statements of success. 4 p = 0.050, Mann-Whitney U test = 103288.5 5 p=0.28, Pearson Chi-Square 4.816 6 p = 0.013, Man-Whitney U test = 26102.5 7 All available on MOPAC website: MOPAC academic research | London City Hall 8 London’s approach to GPS Tagging of Knife Crime Offenders The Mayor’s Office for Policing And Crime (MOPAC) has been at the forefront in the use of electronic monitoring, demonstrated by first piloting the Alcohol Abstinence Monitoring Requirement (AAMR) tags in 2014. In 2017 MOPAC published the London Knife Crime Strategy against a background of increases in levels of knife crime nationally (from 2014) and within London (from 2016). The strategy included a commitment to pilot the use of GPS tagging with knife crime offenders, something reaffirmed in MOPAC’s current 2021-25 Police and Crime Plan. Using GPS tags with persistent and knife crime offenders on community sentences was first piloted in eight North and East London boroughs between 2017 and 2019. Since February 2019, MOPAC’s pilot was extended, to include the tagging of individuals being released from prison on licence as part of a knife crime sentence. The pilot was endorsed by the Ministry of Justice (MoJ) and designed to work alongside their national GPS programme. In March 2021 the use of GPS tags on licence was extended to offences of domestic abuse (DA) across London.8 MOPAC’s interest in GPS tags is reflected in national developments. HMPPS started their own national GPS tagging programme in April 2019 for those on Community orders, Suspended Sentence Orders, Court Bail, Home detention curfew and certain cases on parole. HMPPS launched its original alcohol monitoring services in October 2020, for offenders sentenced in the community, but later extending the scheme to those leaving prison in Wales and England (in November 2021 and June 2022 respectively). In April 2021 HMPPS launched an acquisitive crime trial, monitoring imposed locations for prison leavers convicted for theft; burglary; or robbery offences. Later that year, in August 2021, HMPPS launched its Immigration Enforcement project, on behalf of the Home Office, extending the use of GPS tags to Foreign National Offenders subject to deportation proceedings (HoC PAC 2022). In September 2021 the Government announced additional funding of £183m over the next 3 years to kickstart the Electronic Monitoring Expansion Programme, with plans to increase the EM caseload from around 13,500 in 2021 to 25,000 by 2025. Between 2021/22 and 2030/31, HMPPS predicts £1.2 billion will be spent on an enhanced electronic monitoring service, extending to wider groups of offenders (HoC PAC 2022). However, during 2022 three highly critical reports were published regarding HMPPS’s previous implementation of electronic monitoring.9 All three identified the evidence for the effectiveness of tagging was slim. One described it as ‘unacceptable that, despite our previous recommendations, the Ministry and HMPPS still do not have sufficient data to understand the outcomes of tagging…. [they] still do not know what works and for who, and whether tagging reduces reoffending’ (HoC PAC 2022), while another concluded evidence for tagging ‘still remained weak’ (NAO 2022). This conclusion came despite the fact that over the last 15 years there have been a number of studies, undertaken in a variety of countries (albeit not in England and Wales), which 8 Findings for all these pilots are available on MOPAC’s website. 9 By the House of Commons Public Accounts Committee (PAC), the National Audit Office (NAO) and Her Majesty’s Inspectorate of Probation (HMIP). 9 suggest the positive impact of electronic monitoring on recidivism.10 Recent Australian research looking into the use of domestic violence electronic monitoring (DVEM) (Boiteux and Teperski, 2023) found that offenders participating in DVEM were around 7 percentage points (p.p.) less likely to return to custody within a year and significantly less likely to reoffend (for any offence [7.1 p.p.]) and for domestic violence offences [10.5 p.p.] in particular. However, even in this study the authors concluded that they were ‘unable to exclude the possibility that unobserved variables and/or factors related to participation in DVEM’ might be influencing their results. The current report details the performance, process and impact evaluation of London’s GPS tagging of knife crime offenders, over a three and a half-year implementation period (Feb 2019 – November 2022). Whilst crime across London (as elsewhere), temporarily changed during this period due to the covid pandemic, since this time knife offences have subsequently begun to increase in London. The need to evaluate the effectiveness of specialist violence prevention and intervention services has never been greater and this report is a timely comprehensive evaluation of such a service. GPS Knife Crime Tagging: Service Design Using GPS tags for individuals convicted of knife crime offences, released from prison on licence began in February 2019. Initially in four London boroughs (Croydon, Lambeth, Lewisham, and Southwark), the pilot subsequently rolled out to all 32.11 Originally due to finish in April 2020, the pilot has been extended with the agreement of the MoJ in annual increments, with its current end date being March 2024. Key partners in the pilot have included the London Probation Service (previously the National Probation Service (NPS) (London division) and London Community Rehabilitation Company (CRC)), the Metropolitan Police (MPS), Her Majesty’s Prison and Probation Service (HMPPS) and Buddi Ltd (the GPS monitoring provider). At its core the pilot harnessed GPS locational technology, to track individuals released from prison back in the community. Data provided to probation staff enabled monitoring of specific license conditions and the subject’s offending behavior. Probation Practitioner’s (PP) were encouraged to use GPS data openly with service users in supervision meetings, to examine and challenge both lifestyle and risk factors linked to offending, especially knife crime. It was anticipated that GPS could help monitor compliance with ‘exclusion zones’12; and attendance at rehabilitation programmes or treatment interventions. Data could also be provided about bespoke and flexible key locations for the individual in ‘real time’, via notifications and summary reports, such as heat maps and ‘top addresses’ visited. 10 See, for example, Marklund and Holmberg (2009 Sweden), Di Tella and Schargrodsky (2013 Argentina), Henneguelle, Monnery and Kensey (2016 France), Andersen and Telle (2016 Norway), Williams and Weatherburn (2020 Australia), Grenet, Gronqvist and Niknami (2022 Sweden) and Rivera (2023 USA). 11 The initial 4 boroughs were selected because they were amongst the highest boroughs in terms of those with knife crime convictions in prison and because a cluster of boroughs next to each other to made training and engagement easier with police and probation. 12 See Table 1. 10 Aims The overall pilot aim was to test GPS tags, as a tool for reducing the likelihood of reoffending. For this cohort specifically, further weapon-related offending was a key focus, although any impact on the frequency, severity or time taken to reoffend is to be explored. To determine success of the pilot other outcomes were identified, including the use of GPS location data to: • improve the management of risk posed to adults, children, and the public. • improve the enforcement of licence conditions and increase deterrence. • challenge the offenders’ thinking and lifestyle and improve rehabilitation. • enable the MPS to map individual crimes effectively to support crime detection. Eligibility criteria GPS tags are most often fitted as part of the individual’s licence conditions at the point of prison release.13 Licence conditions are nominated by PPs based on what is necessary and proportionate for the management of the case following released, as informed by the pre- release risk assessment. All licence conditions are formally approved by the Prison Governor before they are added to the licence and the pilot followed the existing HMPPS processes for identifying and imposing appropriate licence conditions. The length of the tag is specified by the PP, normally up to a maximum of 6 months.14 To be eligible for the MOPAC pilot an offender must be: • Aged 18+. • Subject to a determinate prison licence.15 • Serving a sentence for either a knife possession offence or another offence which involved the use of a knife or bladed object. • Released from an eligible prison.16 • Being released to a London borough. • Being released to viable accommodation (to enable tag charging). Most tag wearers were fitted with tags at the point of prison release on licence, however some were also fitted with tags at licence variation, having already been in the community for a period of time. In addition, in January 2020 there was agreement in principle for the eligibility criteria to be extended to include cases being considered for re-release after recall by the Parole Board, if GPS conditions were deemed necessary and proportionate. Data provision to Probation Practitioner’s (PP) The GPS tag and monitoring provider (Buddi) supplied a range of data to probation, including information about the tag, as well as locational data for the individual. For example, when the tag reaches 30% battery life, the wearer receives a series of automated text messages and 13 There are a variety of reasons this may not happen, including Buddi being unable to attend, the releasing prison being outside of the immediate London region or agreement that the fitting would be more effective at probation premise, rather than at prison release. 14 In some circumstances (cases of identified risk) the maximum length of the tag may exceed 6 months, but this requires probation line management approval. 15 Where the court sets a fixed length for a prison sentence. 16Eligible releasing prisons from Feb 2019 were Belmarsh, Brixton, Bronzefield, Coldingley, Downview, Feltham, High Down, Highpoint, Isis, Onley, Pentonville, Send, Thameside, Wandsworth and Wormwood Scrubs, expanded from February 2020 to include HMPs Onley and High Point. 11 phone calls reminding them to charge for at least an hour, if the tag has still not been placed on charge after two hours the PP will receive a notification of the battery being flat. Exclusion/inclusion and interest zones As part of the licence conditions for the GPS tag it is possible for the PP to specify an area(s) in which the tag-wearer cannot enter or leave. Zones can vary in type (see table for more details), but the PSI 12/201517 states ‘the purpose of an exclusion zone condition must be clear and necessary, and the size of the exclusion zone reasonable and proportionate’. PPs are provided with real time email alerts in the event of an exclusion zone breach. Interest zones function in a similar way, but are not linked to an enforceable licence condition, instead being linked to the ‘trail monitoring’ condition that allows more flexible monitoring of areas linked to risk or rehabilitation. Table 1: Zone types Zone Type Buddi alert Exclusion zones PP is provided with an alert immediately as the tag-wearer enters the specified area(s).18 Exclusion zone with a ‘grace period’ PP/tag-wearer is not alerted until a specified period has elapsed (typically requested by the PP if, for example, the tag-wearer must pass through the exclusion zone for work/when travelling). Inclusion zones Alert is provided when the tag-wearer moves outside a specified area during a specified time period. Interest zones Zones which generate an alert for the PP but are not linked to an enforceable licence condition. Additional Data In addition, for tag wearers with the ‘trail monitoring’ licence condition a weekly summary email detailing the tag wearer’s charging patterns and ‘top visited addresses’ were provided, enhanced as the pilot progressed to include more location and duration details. If applicable as a concurrent licence condition, PPs also receive respective summary details of curfew compliance (as this is a licence condition rather than a GPS licence condition, PPs do not receive real time alerts for any violations). Additional adhoc data requests were also available for PPs from Buddi, including:  patented ‘heat maps’ - indicate the most popular locations, visited by a service user over a certain period (see Figure 1).  trail maps - showing the tag-wearer’s movements.  specific addresses/location visits. After November 2019 some PPs were given direct access to the Buddi system via the secure customer portal, allowing them to produce the various outputs above as required. This access was initially offered to all NPS PPs who had live cases at the time (15 PPs), about half of whom accepted. More recent data provided by Buddi showed that in the last 12 months 198 PPs had accessed the Buddi Eagle system directly, 31 in the last month (August 2023).19 17 Prison Service Instructions (PSIs) outline the rules, regulations, and guidelines by which prisons are run. Prison Reform Trust. https://prisonreformtrust.org.uk/adviceguide/licence-conditions-and-recall-indeterminate-sentences/. 18 There is also the facility for Buddi to phone the on-call emergency probation manager in the event of an emergency out- of-hours breach of an exclusion zone. 19 This figure covers both MOPAC’s domestic abuse and knife crime tags. 12 The MPS and other agencies can also request specific GPS data through an External Agency Request (EAR) process. This can be used to assist in investigation and conviction. Where justified, a RIPA (Regulation Investigatory Powers Act) request is also possible for police teams to gain direct access to the Buddi system to live monitor the location of an offender on tag. Figure 1. Example of Buddi patented technology Heatmap GPS monitoring & crime mapping Crime mapping allows offenders to have their movements automatically cross referenced against the location of reported crimes. Within the pilot crime mapping is only carried out for offenders assessed as “more likely than not” to reoffend within 2 years, i.e., individuals with an OGRS (Offender Group Reconviction Scale) score of 50%+. Only when a match between an offender and a crime location occurs would data be released to police for initial assessment of the significance of the match. Once this initial match has been assessed, then further data can be requested by MPS to support investigation if required. Crime mapping has been taking place since June 2019, and since November 2019 has been undertaken by a dedicated analyst based in the MPS Central Intelligence team. 13 Evaluation The Evidence and Insight (E&I) Unit - MOPAC’s in-house social research and analytical team - were commissioned to undertake a multi-year evaluation of the GPS Knife Crime Tagging pilot. The broad aim was to gather a holistic understanding of the overall delivery, combining learning from across the three-and-a-half-year period (2019-2022). Findings are split into three main sections: • Performance - monitoring key indicators of the innovation, including numbers receiving the tag; types of orders; compliance rates; violations; breaches; order completions; crime mapping hits; crime mapping outcomes. • Process - understand the context of service delivery; implementation over time; learning about the model design; and experiences of practitioners and tag wearers involved in the pilot through surveys and interviews. • Impact - examining the impact of the GPS tag upon tag wearer offending behaviour (while tagged, and once the tag has been removed) compared against an appropriate counterfactual or un-tagged control group. Methodology The current report draws upon a mixture of quantitative data analysis and qualitative feedback from service providers and tag wearers: Performance Data A range of sources have been brought together, mainly from the tagging provider Buddi and MOPAC delivery staff. Information about the tag wearers; the types of orders they were on; and their compliance with the tag. In addition, data was provided by the MPS about the crime mapping element of the pilot (i.e., numbers of crimes up-loaded, numbers of matches, and subsequent actions taken by the police). Tag wearer survey Those tagged during the pilot were asked to complete a short survey at the point of fitting and again at removal. The survey, designed by E&I, consisted of 3 topics: the wearers’ anticipated/actual perception of the impact of wearing the tag on their behaviour; their relations with friends and family; and their lifestyle. Data was available from 657 individuals during fitting (where the tag wearer had agreed to take part in the survey), and 448 individuals during tag removal20. Surveys were administered by the tag fitter (Buddi staff). Interviews with offender managers Thirty-four semi-structured interviews were undertaken between the start of the pilot and August 2020, with PPs supervising those on GPS tags. Individuals were asked about specific cases details; how they had used data from the tag in supervision; and for their general views of the tag. Interviews with PPs were undertaken regardless of whether the individual(s) they supervised was/were recalled or not (see appendix for various interview schedules). 20 The difference between the numbers of tag fitting and removal surveys is accounted for because of the number of recall cases, and the fact that some tags are still active. 14 In addition, 6 focus groups were undertaken with probation teams during May and June 2021. Teams were selected either because they had high/medium/low caseloads; or high/medium/low take-up of tags. The primary purpose was to ascertain why take-up varied across PPs, and to try and establish why some PPs had not used GPS knife crime tags for their supervised offenders. Interviews with tag wearers Semi-structured interviews with a small (10 people) convenience sample of service users were completed between November 2019 and March 2021. Individuals were asked about their experiences wearing the tag; their understanding of why they had been tagged; and views of the tag overall. Tag-wearer interviews were only undertaken with those who had not been recalled and were still in the community, which clearly introduces a degree of bias into the sample. Interviews with MPS officers Interviews in relation to the crime mapping element of the GPS pilot were undertaken in July 2020 and November 2021 with 16 MPS officers, some who were borough single points of contact (SPOCs) for the crime mapping process and others who were members of the central analysis team that did the crime matching. Data considerations There were several data challenges for the evaluation. The criteria for selection for inclusion on the tag, particularly the broadness of the offence category (‘serving a sentence for either a knife possession offence or another offence which involved the use of a knife or bladed object’) and the level of discretion provided to PPs over selection, produced a treatment group that was difficult to match for subsequent analysis. The specific difficulties in creating an appropriate control group for the impact analysis undertaken for this research are discussed in the section below. Impact analysis To examine the impact of the GPS knife crime tag on offending, the research compared offending patterns whilst on and off tag for two matched groups (treatment, who had received the tag, and control, who had not) created using propensity score matching (PSM). PSM consists of four phases: ‘estimating the probability of participation, i.e., the propensity score, for each unit in the sample; selecting a matching algorithm that is used to match beneficiaries with non-beneficiaries to construct a comparison group; checking for balance in the characteristics of the treatment and comparison groups; and estimating the programme effect and interpreting the results’21. The creation of the control group for the impact analysis was not a simple task. The knife crime tag had a staged roll-out across London. In addition, as has been mentioned elsewhere in this report, offenders released from non-London prisons were not eligible for the GPS knife crime tag, but data was not available to use them as a counterfactual group. Initially the 21 https://www.betterevaluation.org/methods-approaches/methods/propensity-scores. 15 intention for the control group was to take account of the time lag in the rollout of implementation and the non-eligible prisons to create a natural control group of those who would have been eligible for the tag elsewhere in London. Therefore, the initial request to probation was for all those released on determinate licence between 1/1/2019 and 31/3/2020 from specific prisons and from non-participating boroughs for a range of specified offence types (possession of weapons/blades, robbery, violence), aged 18 or older. 22 However, analysis of the control group showed that there were insufficient cases (n=69) to allow appropriate matching to the treatment group. As a result, it was decided to increase the size of the control pool by revising the date criteria for inclusion to cover the period between 1/7/2017 and 31/12/2018, i.e., the period before the GPS knife crime tag had been implemented. While this provided a larger pool for comparison, it meant the control and treatment groups were taken from different time periods, which was not ideal, and had implications for the analysis that are discussed later. The same selection criteria were used, and data fields requested, returning records for 3,869 people. Exclusion of cases where data was missing from CRIS, or where there was no knife crime feature code, or knife crime prior to release (identified from offending histories), reduced the size of the control dataset down to 1,951 cases, that was matched against the 600 treatment cases. The planned core variables for the propensity score matching were age, ethnicity, sex, OGRS2 and offence type. For the last, it was initially planned to use 10 Home Office offence groups (burglary, criminal damage, drugs, fraud, other offences, robbery, sexual offences, theft, violence, and blank/no crime/crime related incident). However, concern about the broadness of these groups, and the presence within them of offences which were qualitatively quite different led to the decision to produce bespoke offence groupings that took into consideration relatable behaviours, intent, type of victim and / or victim interaction. As a result, the following 15 offence groupings were used in the analysis. • Commercial robbery. Any robbery or aggravated burglary where the victim was a business/commercial rather than personal. • Drugs possession. Any drugs offence separated from those where the intent was to produce, sell, supply. • Drugs supply. Any production, actual or intent to sell or supply drugs, separated from possession only offences. • Kidnap, blackmail et cetera. Offences commonly grouped with violence, but having specific characteristics which indicate specific intentions potentially linked to organised crime. • Lower threat and all other offences. • Murder. • Other assault and violence. Any personal violence offence not included under murder or serious violence, plus public disorder. • Other sexual offences. Any sexual offence which does not involve a physical assault. • Other Theft, Acquisitive and Fraud Offences. Any offence whereby the intent is acquisitive (such as Burglary, Motor Vehicle theft, Other theft), but there is no or minimal direct interaction with a victim (such as during a robbery, artifice/distraction burglary). 22 The data fields requested were date of birth, gender, ethnicity, offence type, sentence type, sentence length, OGRS (1- and 2-year score), borough (OM team), release data, releasing prison, CRN, PNC number, whether the individual had been recalled and, if so, the date of recall. 16 • Personal Robbery. All robbery and aggravated burglary offences where the victim is a person/s, not business or commercial. • Serious Violence. Violent offences where injury is caused / intended / threatened. • Sexual Assault. All forms of sexual assault and rape. • Stalking, Harassment et cetera. Any form or stalking or harassment which may be either sexually, racially, religiously or otherwise personally motivated. • Theft Person and Artifice / Distraction Burglary. Acquisitive offences where there is a direct interaction with a victim. • Weapons. Any weapons offence. Three sets of variables were created based upon the 15 offence classification groupings. • a binary variable to show whether each subject had any accused record on CRIS for that offence type. • the number of accused records on CRIS for each offence type. • the accumulated severity or harm score for each offence type – the Cambridge Harm Index (CHI) score.23 Ahead of the PSM, a series of binary logistic regressions were conducted on various iterations of the data. Following this analysis, the decision was taken to remove the OGRS2 variable from the analysis because of its limited contribution to the model, and because of the overlap with the offence variables (detailed above) being used as alternatives. In the subsequent regression analysis, 3 core variables were used in all (age, sex and ethnicity), together with all binary offence variables, total frequency of offence variables and CHI offence variables (as a measure of severity). Propensity score matching was initially undertaken using the 3 core variables and each of the offence variables (binary, volume and CHI). Matching was done with randomised case order on, without replacement, and using a range of match tolerances (0.25, 0.2 and 0.1). Subsequently the distribution of the matches was examined within different bandings for the propensity score (between 0.00001- 0.2) (0.20001 – 0.4) (0.4001 - 0.6) (0.600016-0.8) and judged to be sufficiently equal (see appendices for distribution). A final decision was taken to use the matching from the number of offences variable rather than the CHI variable, because the latter had less model strength and there were concerns about the large variation in offence groups in the CHI scale. T-tests were undertaken on all the variables. Some offence groups, whilst significant were deemed to be irrelevant to the analysis, as they were felt to not be theoretically tied to the eligibility criteria of the pilot (e.g., lower threat offences, other theft/acquisitive offences, other sexual offences, and theft person including artifice burglary offences) and were therefore excluded. PSM matching was re-run for offence totals at 0.1 match tolerance, producing 600 fuzzy matched cases. Mann-Whitney U tests were run on the output. There were still significant differences on age and ethnicity, but after applying a standardised mean difference test, these were not significant and allowed the analysis of reoffending to proceed (appendix A provides comparative details for the control and treatment groups of the distribution of the variables used in the analysis). 23 There are several measures of offence severity/harm available. Variations have been produced by the Ministry of Justice (MoJ), the Office for National Statistics (ONS) and the Cambridge Harm Index (CHI). While none were ideal, ultimately the decision was taken to use the CHI because it was felt to be more robust. 17 Results from Service Delivery: Performance Insights This section provides an understanding of what the service delivered over its duration, drawing upon basic performance data; tag wearer demographics; and information about compliance with the tag. GPS tags imposed Between the pilot’s start in February 2019 and the date of analysis (November 2022), a total of 971 GPS tags had been imposed with 881 individuals - 799 people had been tagged once, with 82 tagged more than once.24 Three-quarters of tags (n=726) were imposed on the individual’s release from prison, the remainder (n=243) as part of a licence variation for those already in the community. When tags were imposed at prison release, individuals were from Isis (19%, n=154), Feltham (13%, n=104), Brixton (12%, n=95), Thameside (10%, n=80), Wormwood Scrubs (8%, n=66) and Pentonville (8%, n=64). The number of tag wearers released from Feltham and Isis, which focus on young adult offenders, reflects the fact that over 55% of the tag wearers were aged under 25 (see ‘who was tagged’ for more information). Where recorded, the mean planned tagging period was 138 days, with a minimum of 10 days and a maximum of 365. Offences for which the tagged individual was serving a current sentence include violence and the use of weapons (22%, n=214 had possession of an article with a blade or point in a public place; 10%, n= 93 intend to do grievous bodily harm; 9% n=85 were convicted of wounding or inflicting grievous bodily harm). All bar one of the tag wearers (n=970) were subject to trail monitoring and over half (54% n=525) were subject to at least one exclusion zone. 24 77 tagged twice, 3 tagged 3 times, 1 tagged 4 times and 1 individual tagged 5 times. 971 tags imposed 75% (n=726) imposed at point of prison release. • 19% (n=154) HMP/YOI Isis. • 13% (n=104) HMP/YOI Feltham. 138 days mean planned tagging period. Main offence of tag wearer • 22% (n=214) possession of article with a blade or point in public place. • 8% posession of an offence weapon (without lawful authoirity or reasonable excuse). • 14% (n=139) robbery. • 10% (n= 93) shooting, wounding etc with the intent to do grevious bodily harm or to resit apprehension. • 9% (n=85) wounding or inflicting grevious bodily harm. Enforceable licence conditions • All bar one of the tag wears were subject to trail monitoring (n=970). • 54% (n=525) of tags were subject to exclusion zone(s) •27% (n=258) requirement to attend specified location. •24% (n=231) requirment to report to approved premises. • 2% (n=15) were required to report to a police station. 18 Who was tagged? Almost all tag wearers were male (98%, n=948), with a mean age of 26.6 years25 and most (52%, n=453) are from a Black or Black British background26. In fact, Black or Black British individuals aged under 25 years made up 30% (n=258) of tag wearers, increasing to 32%, (n=309) when including individuals from Mixed White and Black ethnic background. White tag wearers (23%, n197) tended to be older compared to the overall cohort, with 30% (n=59) of individuals from a White background being aged 35 years or over compared to 16% (n=157) of tag wearers overall, and 12% (n=55) of those from a Black or Black British background. While the percentages of Black individuals in the GPS caseload are clearly disproportionate to proportions of Black Londoners in the general population, they do reflect the picture the criminal justice system more widely, in Metropolitan Police Service statistics, or the probation caseload (which clearly is an issue in itself). As a comparison, the proportion of people with Black or Black British ethnicity within the pilot was 8 percentage points higher than the proportion of those of the same ethnicity in the closest comparison group used for benchmarking. The proportion of tag wearers whose ethnicity was recorded as Mixed was two percentage points lower in the pilot compared to the benchmark sample. The mean OGRS score (2 year) of tag wearers was 62%,27 which places the group on average at medium risk of offending28, but there was a wide range of scores across the cohort (a range from 4% to 97%). Older tag wearers tended to have a lower OGRS score, with the mean score of those aged 45+ years being 49%. The OGRS score is also important because the crime mapping facility is only available for tag wearers with a score over 50. 25 Where the age of the tag wearer is recorded (n=970), most tag wears were aged under 35 years (84%, n=813), with 57% (n=552) aged 18-24 years, 27% (n=261) aged 25-34 years, 10% (n=94) aged 35-44 years, 5% (n=48) aged 45-54 years, and 2% (n=15) aged 55+ years. 26 Where recorded (n=869), 23% (n=197) were from a White background, 13% (n=116) were from a Mixed ethnic background, 7% (n=61) were from an Asian or Asian British background and 5% (n=42) were from an ‘Other’ ethnic background. 27 Mode = 78%, and median = 65% 28 OGRS scores are banded into ‘low’ (49% or below), ‘medium’ (50-74), ‘high’ (75-89) or ‘very high’ (90-99). • 98% (n=948) male Gender • 84% aged under 35 years (n=813) • Mean = 26.6 years Age • 52% (n=453) Black/Black British • 23% (n=197) White Ethnicity • 32% (n=309) aged under 25 years and Black/Black British or Mixed White and Black ethnicity Age x Ethnicity • Mean OGRS score was 62% OGRS score 19 Compliance with the GPS tag 29 Considering the reasons for recall, the effective compliance rate (the proportion of cases completed, without recall for failure to comply with the GPS tag as a licence condition) is 98% (n=815). 30 This relatively high level of compliance can be seen as a positive – although is not a measure of ‘impact’. It needs to be understood as only considering individuals recalled solely for failure to comply with the GPS tag not those recalled for multiple reasons including failure to comply with the GPS tag. 29 This compares against a compliance rate of 66% for licence and post-sentence supervision cases supervised nationally by the CRC (July-September 2019). Figures are not available for the NPS. 30 As a compliance rate, this is higher than that found by MOPAC in the use of GPS tagging with an integrated offender management (IOM) cohort (56%) and alcohol abstinence monitoring requirement (AAMR) (94%). It should be noted however, that the characteristics, and offending behaviour of the individuals who receive these various interventions differ. 971 tags imposed to November 2022. 136 cases were still live. 835 tag wearers completed period of GPS monitoring. 298 ended due to recall 537 ended without recall 64% rate of completion Where data available: • mean period an individual was tagged was 125 days (n=473). • Mode 184 • median 121 median 121. Where data available: • mean period an individual was tagged was 55 days (n=290). • mode = 1. • median = 43.5. • 153 cases (51%) recalled for reasons unrelated to the GPS tag (e.g., poor behaviour, missed probation appointments, or a new incident). • 53 cases (18%) were recalled for multiple reasons with GPS a secondary reason. • 37 cases (12%) were recalled due to a breach of GPS monitored exclusion zones. • 20 cases (7%) were recalled primarily for GPS related non-compliance. • 18 cases (6%) were recalled for a new incident detected through crime mapping or evidenced through an external agency request. • 17 cases (6%) recalled because of a breach of other conditions identified using GPS (for example, breach of curfew or failure to reside at an approved address). 20 Results from Service Delivery: Process Insights This section provides learning from across the service, including the process of implementation; adherence to the original service delivery model; and perceptions from those involved (qualitative research carried out with PPs and tag-wearers on their perceptions of the tag’s operation; and police staff about the crime mapping aspect of the pilot). Implementation improved as the pilot embedded. All services encounter implementation delays and challenges as they take time to embed. Using GPS tags for monitoring knife offenders was no exception, with practitioners suggesting as the wider London rollout occurred, the operational delivery of the pilot improved. Initially, when limited to a few boroughs, it was felt that some eligible cases were being missed due to complications with ‘gang’ affiliation and approved premises. For example, high- risk younger offenders, with peripheral or entrenched involvement with serious group offending were unable to be housed outside the pilot area but were also not allowed to return to a home area with their ‘gang’ affiliation. As the pilot rolled out, most approved premises (Aps) were covered, allowing individuals to be placed in safer areas and have a GPS tag to monitor whether they were going back to risky areas. Other areas of improvement observed during the pilot’s implementation, often following feedback from probation practitioners, were changes in the routine information provided to PPs by Buddi, particularly in relation to the most popular locations frequented by tag wearers, improved guidance to PPs about how to ensure that tag wearers were charging, and improved procedures to allow PPs to check whether tags were faulty. There was a recognition on the part of the MPS staff that probation had been subject to staff resourcing concerns and had experienced high levels of staff turnover, as well as the impact of Covid and organisational restructuring during the period covered by this evaluation. Respondents suggested that this had sometimes had a knock-on effect in terms of probation’s awareness of the scheme, their proactivity, and willingness actively use GPS data when managing those on tag. There was debate as to whether the correct people were tagged & where the function should sit within the MPS. When asked, most PPs appeared clear about the eligibility criteria for the pilot and why the tag-wearer had been recommended for inclusion, tending to refer to the individuals current and/or previous knife-related offending. However, views from MPS officers on the extent to which the ‘right’ people were on tag varied. Some suggested probation had not identified the correct individuals (anticipating more gang nominals),31 and that those on the tag ought to have higher OGRS scores. There were mixed views from interviewed MPS officers about whether the GPS knife crime pilot should be police rather than probation led.32 Justifications were that the police would be more proactive and had a better understanding of who should be on the tag (a reflection of the varied willingness from probation to include police 31 Although explanations for their non-inclusion were suggested (principally not meeting the age and offence criteria for the GPS tag). 32 The legal basis for MOPAC’s GPS KC tagging pilot means probation, rather than the police, lead the nomination of cases and receive most of the data. 21 suggestions of who should be tagged). However, there was caution, as it was felt police access to the data would lead to ‘fishing expeditions’ and it was therefore better left with probation as a result. There was more consensus from those interviewed that IOM was the ‘natural home’ within the MPS for responding to the crime mapping work, although some boroughs decided it sat better with the gangs unit and a specialist knife crime SPOC. There may be communication challenges for individuals to be tagged. While the pilot’s aim was that the service user would be informed that they would receive a tag prior to release, in a small number of cases it was suggested that this had not happened. While the majority (92%, n=607) of respondents to the tag-fitting survey agreed they understood why they had received the GPS tag, what they had to do to comply with the tag order (98%, n=645), how the GPS requirement worked (98%, n=644) and how the tag fitted with their other licence requirements (93%, n=612), there were a smaller number (45 individuals) who reported that they were unaware they were getting tagged, or that the fitting of the tag had not previously been discussed or explained to them. 33 Similar views were expressed in the interviews with tag wearers, with half of the ten participants reporting that they had not previously known about the tags; one saying he ‘didn’t even know there was a GPS tag’, and two suggesting that they were ‘told that it was voluntary, but it wasn’t’. Likewise, three of those interviewed suggested that they had not understood why they were on tag, with one saying it was the only way of getting out of prison.34 PPs stated that informing individuals that tags would be used as part of their supervision usually took place during pre-release interviews/visits in prison, also via video link, although these visits were impacted by Covid.35 In cases where the PP had not been able to provide details in advance, the information had usually been provided by the tag-fitter. However, problems with communications about the installation of tags were not new, having been highlighted in E&I’s previous interim report, and although attempts have been made to improve information provision via the production of booklets about the tag for tag users, there remain examples where tag wearers were not as prepared for the tag fitting as they should have been. Nevertheless, despite a potential lack of warning that they were to be tagged, when asked how confident they were about successfully completing time on it, 97% (n=640) of respondents to the tag-fitting survey were ‘very’ (66%) or ‘quite’ (31%) confident. The majority (77% n= 358) also still believed the imposition of the tag had been ‘very’ or ‘quite fair’, compared to 8% for ‘very’ or ‘quite unfair’. Tag fitting did not always occur at prison. While it was always anticipated that the main route for people being tagged would be at prison release, there were certain cases where this could not apply. For those tagged via 33 The 2nd most popular negative response in the survey’s free text field, only behind the 65 responses where people were generally unhappy about being tagged. 34 In contrast, 5 interviewed tag wearers suggested they fully understood what they had to do to comply with the tag’s conditions (charge battery, not go into certain areas), and 4 said they understood the police would be able to monitor their movements (2 said they did not). 35 Although in one case the individual had been in segregation, so the PP had been unable to see them and there were several cases where the current supervising OM had ‘inherited’ the tag wearer, so was unaware what the pre-release arrangements had been. 22 licence variation (n=243) prison fitting was clearly not an option. Similarly, for those being released from prisons furthest away from London (Onley and Highpoint) it was agreed with Buddi that fittings would take place instead at the probation first appointment). As a result, of the 964 cases where the location of fitting was recorded, under half (49%, n=476) were at prison, with the remainder at probation (47%, n=454), or another approved premises (4%, n=34). From the PPs interviewed, similar numbers of tags were fitted in prison compared to the probation office, hostel, or other approved premises (13 compared to 11). When asked, six of the ten wearers had the tag fitted in prison; one was fitted in a hostel the day after release; and three at probation (one ‘a few weeks’ after release). Reasons for the change from the original planned process tended to be pragmatic - the unexpected release of the tag recipient; breakdown in communications between Buddi, MOPAC and the prison; a ‘mix up’ with the prison fitting, resulting in a same-day fitting at the approved premises. Although in one case it was reported to be a pre-planned decision due to risk e.g., fitting occurred in a probation office out of area, because the PP’s office was inside the tag-wearer’s exclusion zone and deemed too risky. Perhaps unsurprisingly, there were fewer issues associated with tag removal, although one wearer cut off his tag when it had not been removed at the appointed time. There was varying use of Exclusion Zones. Exclusion zones existed for 525 (54%) of the 971 tags, and of these 118 had precise location details recorded; they covered a wide range of locations. Some exclusion zones were extremely large - one was for an area that comprised South West London, North Surrey, North Hants and East Berkshire, 2 were for counties outside London and 9 covered cities and towns outside London (including Gateshead, Nottingham and Peterborough as well as towns in the Southeast). Six exclusion zones were for the Notting Hill Carnival, and a couple for the Wireless Festival. Most exclusion zones (41) referred to specific London boroughs – 32 of these being for a single borough, while 8 applied to 2 boroughs, and one to three. 15 exclusion zones related to areas that were part of a London borough (in one of these cases, 12 locations in a West London borough). Twenty-two exclusion zones covered London postcodes36 (a further exclusion zone covered a Brighton postcode). Four exclusion zones were for estates, and 3 for town centres. In comparison, 12 exclusion zones referred to a street or streets, and only 8 were for specific addresses (including one outside London). Three exclusion zones related to stations or (14 named) tube stations, and 2 to supermarkets. Other exclusion zones referred to a specific gang territory, 5 pubs in a specific London postcode, a hospital and civic centre, local bookmakers, educational establishments, and an athletic stadium. The responses from the PPs interviewed reflected the findings above, in terms of proportions of tags with EZs and the variety of levels used (although the proportion of tag wearers with exclusion zones had increased noticeably for those PPs interviewed most recently). One PP indicated that while she had set up an exclusion zone, she had let the tag-wearer pick an area 36 Postcode here refers to the first part of the postcode reference ie E11. Of the 22, 16 referred to a single postcode, 3 to two postcodes, 2 to 3 and one had 5 excluded postcodes identified. 23 of his own to emphasise the benefits of an area for his own safety. PPs who had not established exclusion zones stressed that this was because they did not feel it was appropriate (because the tag was ‘purely for monitoring’, and because of the ‘mobile nature’ of the offence). PPs identified the need for a degree of negotiation or flexibility in the enforcement of exclusion zones. One tag-wearer, who had been moved outside London but had to travel into London, always came into a terminus station within his exclusion zone; ‘having it on record to say there was a breach and that there is no other way it forced me to use my own initiative to contact the police and say that we need to have a sort of negotiation here because I am not prepared to prosecute someone for passing through a train station that is unavoidable’. There are some practical challenges to wearing the tag. In common with other electronic monitoring evaluations,37 respondents’ perceptions of the comfort of the tag differed. While a large majority of respondents from the ‘removal’ survey agreed that the tag was comfortable to wear (76%, n=161), 6 of the ten interviewed tag wearers felt it was uncomfortable or too bulky, and some PPs reported individuals on their case load had remarked on the uncomfortableness and size of the tag. The most frequent ‘negative’ free text response in the tag removal survey was in relation to difficulties in charging (mentioned by 27 respondents). PPs reported mixed reflections from those they were supervising, with some tag wearers having no problems with charging throughout, others initially struggling, but subsequently able to successfully charge after reminders from the PP or via Buddi. However, both tag wearers and PPs reported there were situations where tag wearers received notifications from Buddi stating they were not charging the tag when they were. PPs expressed concerns initially that it was difficult to establish whether the fault lay with the equipment or with the tag wearer, which had led to Buddi producing improved guidance for PPs during the pilot. Anecdotally, PPs suggested the reality was probably a combination of tag wearers tampering/abusing the equipment, and some faulty equipment.38 37 For example, Pepper and Dawson (2016). 38 As an illustration, between the 1st of January and the 31st of March 2023, Buddi undertook 15 follow-up appointments with tag wearers, 4 of which led to equipment being swapped out. Extrapolating these figures over the course of 12 months would mean Buddi swapping out 2% of equipment installed on offenders. This compares to the company’s overall percentage (August 2023 - derived from worldwide figures) of less than 1% of active devices returned and replaced (Communication from Buddi). “[Charging] wasn’t a problem for me, you had the home thing and the charger at home. I ensured that every day an hour before I got home from work, I would tell my sister to put it on charge, by the time I got home, I had a routine, that was the best way.” GPS tag wearer “Buddi replaced the battery twice, after the second time I said that anything else that happens to the battery would mean a warning letter, because from the first time I gave him a second chance and by the second I knew that he was doing something to [the battery], I gave him a chance so that he couldn’t say that I was giving him warnings for no reason, but that was the only issue that we had”. Probation Practitioner 24 Probation practitioners’ perceptions of Buddi. When asked specifically about having direct access to the Buddi system most PPs felt this was/would be an improvement, leading to less ‘to-ing and fro-ing’ between the PP and Buddi and particularly useful for difficult cases or after hours. It was also suggested that direct access to the system would be useful during supervision, to clarify tag-wearers’ movements quickly in the event of an exclusion zone alert, and as an ‘extra safety net’. Those PPs who did not feel live access was necessary suggested that there was enough information from the existing reports and notifications already. In fact, access to the Buddi portal was available to PPs at any time, something that was explained in training and via other communications, so the gaps between some PPs wanting this option but not getting it was more a reflection of them not being aware of or understanding the option, rather than any restrictions that were in place. PPs’ level of understanding of the data available from Buddi, its meaning and what could be requested, varied, and a couple of PPs referred specifically to the need for improved/continued training about the system and its capabilities. Three PPs mentioned the ‘burden’ of notifications from Buddi, referring to multiple/duplicated emails that could be confusing and time-consuming. The introduction of password protection on the reports was also described as a source of frustration. Problems experienced with the set up and administration of the tag (details of a curfew address being recorded incorrectly, difficulties in fitting/removing tags, blank reports being sent to the PP) necessitated contact between the PP and Buddi which could also be time consuming. Overall, the PPs interviewed were complimentary about their contact with Buddi. Interviewees frequently mentioned the ‘helpful and responsive’ nature of the staff (whether via email or telephone),39 and the tag fitters were particularly singled out for praise in explaining the requirements of being on tag, being described as ‘very professional’. One stated: ‘They were very helpful, the minute he breached an exclusion zone I would get an automatic email, if they were contacting him to locate him, I would get an email, if they were sending alerts to his phone they would let me know, then at the time that I wanted further information, they would find it’. 39 Although one PP was critical about the level of knowledge about PPs’ roles on the part of Buddi’s customer service staff. 25 Exploring the benefits of the GPS tag This section focuses on understanding if the pilot achieved its intended benefits. Primarily the aim was to test the GPS tags’ ability to reduce the likelihood of reoffending, specifically weapon-related, although any impact on the frequency, severity or time taken to reoffend is to be considered. Other outcomes were stated to be improved management of risk; improved enforcement of licence conditions; challenging an offenders’ thinking and lifestyle; improve rehabilitation; and enable the MPS to effectively map crimes to support detection. Staff outlined the benefits of GPS tagging including for other offences. PPs overwhelmingly expressed positive views about the GPS tag. It was suggested they: • Provided ‘structure and discipline’ for the wearer. • Enhanced case management. • Gave more information - extending monitoring beyond the functionality of a home detention curfew (HDC) tag. • Aided supervision - particularly to check individual movements; general monitoring of where they had been; and oversight of exclusion zones. • Facilitated discussions during supervision, enabling conversations about movements and safety. • Had a deterrent effect. • Provided the ‘perfect alternative to recall’. • Gave the PP a greater ability to protect the victim(s). Whilst a few PPs (and a few interviewed wearers) did not think the tags made any difference to the nature of their supervision, their reasons were often linked to an early recall of a wearer, meaning the use of GPS data in supervision meetings was not possible. For the majority of PPs, it was felt tags had changed their supervision, for a variety of reasons. Checking the movements of tag-wears was commonplace and data was even used to inform decision making regarding recall. Of the ten interviewed tag-wearers, four stated there were conversations / challenges about their movements in supervision meetings, particularly into or near to exclusion zones. One PP said they ‘would not be able to manage the same way without the tag’40. A knock- on benefit of checking movements was the ability to check tag-wearer’s contacts with associates, in relation to suspected engagement in County Lines, or to monitor the tag-wearer’s return to areas where what were termed ‘risky’ groups/peers resided. 40 Two OMs indicated that they were now supportive of the GPS tag despite having initial concerns. Indeed, one of these individuals suggested that one of the dangers with the use of the tag was that OMs would be become over-reliant on it. “what the tag enabled me to do, and what he knew the tag enabled me to do, was monitor exactly his route that he took in … that was really, really helpful because also it builds trust, as well as be there to catch any return to a lifestyle” Probation Practitioner 26 A few PPs suggested location data had been used to challenge or discuss the tag-wearer’s behaviour, suggesting the data had ‘opened up’ that conversation. Others PPs did not feel it had been appropriate to use the data in this way, because they could clearly see where the tag wearer had been, or it was not feasible (in cases where the tag-wearer had been non- compliant or recalled). However, many stated the tags set the tone of better/clear communication. Certainly, the tags had facilitated increased contact between the PP’s and the wearer and provided data in which additional questions could be made and progress could be measured. One tag-wearer suggested that the presence of the tag had led him to contact the PP in advance because he knew he would be going into an exclusion zone to see a family member. Whether the opinion that increased communication assisted in building trust varied across the PPs, as some felt the imposition of the tag initially soured the relationship, which took some time to overcome. However, one interviewed wearer suggested, because he had previously absconded, the tag played a role in rebuilding trust between him and his PP as his movements could be better monitored. PPs also stated that they used the tag to manage risk, either to or from the wearer. This included setting limits; dissuading risky decision making; or extending the tagging time to continue monitoring. In interviews MPS SPOCs suggested the tag had a deterrent effect on offending – one respondent described wearers as being as ‘good as gold’. PP’s suggested they were particularly effective at restricting movement into exclusion zones aimed at keeping the tag wearer away from an ex-partner, one remarked that when this failed, the tag-wearer was worried because they knew they were being monitored. There was limited reported use of heat maps to check whether the tag-wearer had been in areas where crime had been committed (only one PP mentioned this). Where negatives were expressed by PPs, these related to frustrations about charging and equipment issues, or the ‘time consuming’ nature of monitoring the tag. Some queried why the use of GPS tagging had initially been limited to knife-crime offenders, with a general feeling they could be expanded to cover other offence types, the most popular offence types being domestic abuse and sexual offences (PPs stressing this was primarily “His offending was gang-related, there was weapons offences, possession of knives, firearms previously. It was my thinking of let me have a bit of a measure on his movements and then if I could see a lot of erratic movements around the place, then that might suggest that he may be up to something – acquisitive offending, whether that be dealing or at worst going to seek out altercations with rivals. I couldn’t see a lot of other protective elements that I had to work with this guy. I felt that this was the only possible way that I would have tabs on what he was getting up to, I found it almost a bit of a godsend in that way because I wouldn’t have had anything else” Probation Practitioner “There was a lot of times when the tag wasn’t picking up the charge and so there were warning things, so I was in regular contact with him to let him know that the thing wasn’t charging, but the contact was on both sides and I would contact him so that I didn’t breach” Tag Wearer 27 for the management of risk to the victim)41; county lines and/or drugs supply; serious group/gang-related offending; burglary; and harassment/stalking. Other offences/uses mentioned were for serious organised crime; firearms; and other weapons offences and for those on post sentence supervision (PSS). Probation practitioners believe the tag positively affects offending behaviour, but are unsure about long-term behaviour change. Overall PPs felt that the tag had had an impact on the tag-wearer’s behaviour whilst they were on it - citing as examples that the individual had not been into exclusion zones or had breached less frequently than previously. In circumstances where the PP did not feel the tag had had an impact on short term behaviour this tended to be attributed to the personal characteristics of the tag-wearer; [he’s] ‘not that sort of person, not bothered’. PPs were more ambivalent when asked if they thought the tag would have an impact on the tag-wearer’s longer-term behaviour, particularly in light of the tag’s relatively short duration. Some PPs felt that there was potential for continued improvement, suggesting the tag had provided ‘positive reinforcement’ and reflection on the tag wearer’s part, others felt either it was impossible to say or not the case - ‘definitely not the thinking thing’, pointing to the tag- wearer’s subsequent conviction for the same offence once off-tag, or improvements while on the tag which had ceased once it had been removed. Tag wearers reported beneficial effects on their offending behaviour. When asked to reflect on the impact that wearing the tag had had on general aspects of their life at the point the tag was removed, tag wearers felt that the tag had had a particularly beneficial impact on their offending behaviour (83%, n=310) feeling it had had a positive impact), ‘life in general’ (60% positive, n=226) and relationships with family (44%, n= 165). However, respondents were not persuaded about the benefit of wearing the tag on their relationships with family and friends, financial situation, employment, housing circumstances, education, physical health and mental well-being, most indicating that the tag had had no impact. In only two of the questions (‘financial situation’ and ‘employment situation’) did the 41 Something that has subsequently occurred, with the introduction of the domestic abuse DA GPS tag. “Hard to say because he was on it for only five or six weeks. I don’t think it impacted on long-term positive change but I think that what it did do is give him a five- to six-week period where he didn’t have to even consider going back to a risky area, where he could actually take time to focus on doing what he wanted to do … it gave him a period of time to recognize that he’s on licence, he went to prison for a reason, he can’t do back to an area where his peers might pressure him, so he might as well do something positive, which was really good for him. Since he’s come off the tag, his motivation has nosedived. He hasn’t really done much at all. He’s been rearrested. He’s almost regressed. But I acknowledge it was a really short case, really short period of time.” Probation Practitioner 28 number of ‘worse’ responses exceed the number of ‘better responses’ (details in chart 1 below).42 Chart 1 Wearers opinions of being on tag Tag-wearers were also asked to describe their experience of being on the GPS tag overall (from 1 (‘Very negative’) to 7 (‘Very positive’) at the time the tag was removed. Chart 2 shows the responses, with 60% of responses being on the positive side of the scale (scores of 5, 6 and 7, the mean being 4.6 and the mode 5). Chart 2 Tag wearers overall experience of being on tag (n=381) 42 Tag wearers had been asked the same set of questions when the tag was initially fitted, and the pattern of responses was very similar – no impact being the largest category for all questions, apart from impact on offending behaviour, which 88% of respondents thought would improve, life in general (66% improve) and relationship with family (46% improve) with the tag on. 60 44 35 11 83 31 34 16 9 28 9 28 44 47 76 16 57 46 80 74 47 89 12 11 18 12 1 12 20 4 16 25 3 LI F E I N GE NE RA L (N376) RE LA T I ONS H I P W I T H F A MI LY (N374) RE LA T I ONS H I P W I T H F RI E NDS (N373) MONE Y /F I NA NC I A L S I T UA T I ON (N368) OF F E NDI NG BE H A VI OUR (N373) P H Y S I C A L H E A LT H (N371) ME NT A L W E LLBE I NG (N374) H OUS I NG S I T UA T I ON (N361) E MP LOY ME NT S I T UA T I ON (N364) A BI LI T Y T O GO OUT /S OC I A LI S E (N369) E DUC A T I ON S I T UA T I ON (N338) % BEING ON THE GPS TAG MADE MY..... Better No impact Worse 4 4 12 20 31 22 7 0 5 10 15 20 25 30 35 1 2 3 4 5 6 7 % 1 (very negative) to 7 (very positive) 29 A question was added to both the tag fitting and removal surveys in April 2021 which asked respondents to agree/disagree with the statement ‘wearing the tag made me less likely to offend’. Of the 243 respondents who answered this question in the tag removal survey, 80% (n=195) agreed or strongly agreed with the statement, while only 4% disagreed or strongly disagreed.43 In terms of the perceived benefits of the tag recorded either during interview or via the tag removal survey, the most popular was it provided the wearer with the opportunity/justification to keep away from trouble and/or certain people, (‘no one wants to be with someone on GPS, you don’t know who is monitoring it’). Other positives mentioned were the tag being more comfortable than expected, preferable to a curfew/EMS tag, better than going to prison, had made the tag wearer think about their movements/location, provided a reminder/made them think about their behaviour, enabled them to keep/get accommodation, and provided an alibi by establishing that the tag-wearer was not in a certain area at a certain time.44 GPS Tags positively affect offending behaviour when compared to a counterfactual. Understanding the impact of any initiative is complex, and this is no different for the GPS tag. As a result, a variety of approaches were adopted, and analysis undertaken, to try and measure impact. These included predicted versus actual offending comparisons using OGRS; comparisons of offending from the GPS tagged group across time points; and extending these comparisons to a Control group, constructed through Propensity Score Matching, as described in the Methodology section. The difficulties in establishing a suitable control group for the tagged cohort has already been described in the Methodology section of this report, particularly the inability to generate a suitable control group from the same time period as the treatment group, and the resulting reliance on a control group taken from an earlier time period (1/7/2017 to 31/12/2018 for the control group, and 1/1/2019 to 31/3/2020 for the treatment group). The fact that the control and treatment groups’ periods where reoffending was being measured were at different points chronologically was always a caveat, but this was compounded by the COVID pandemic occurring within the time when the data was collected. The concern was that the two groups were going to be differentially impacted by COVID, and the level of reduction observed in the treatment group had arisen because of the impact of COVID, rather than the GPS tag.45 Essentially, this concern is that offending by the treatment 43 Those who had agreed or strongly disagreed were subsequently asked to agree or disagree with statements as to why the tag had a deterrent effect; 48% agreed it was because the tag reminded them that they were being monitored, 43% because the tag served as a reminder of the consequences of offending, 37% with the statement that the tag meant they had been less likely to mix with people they might get into trouble with, but only 20% agreed it was because they had been worried they would be caught if they reoffended. 44 Conversely, negatives expressed, again either in interview or via the on-line surveys, were the discomfort, the perception that the tag was an invasion of privacy or a breach of human rights, and a means of allowing the police to target young black males. 45 The Crime Survey for England and Wales (CSEW) estimates for the year ending June 2021 compared with the pre-COVID year ending June 2019 showed a 14% decrease in total crime excluding fraud and computer misuse, largely driven by an 18% decrease in theft offences. While there was little change in the total number of incidents of violence ‘there was a 27% 30 group had been suppressed by COVID, something that would not have occurred with the control because their ‘window’ for re-offending occurred before the advent of the pandemic. To determine if a ‘COVID effect’ could be found national reoffending data was examined in the pre/during and post COVID periods (between 2016 and 2021) using published data from the MoJ, comparing this to the reoffending patterns for the GPS KC control and treatment groups over the same period – the results are shown in appendix B. What was found was not a sudden abrupt decline at the point that COVID occurred, either for the treatment or control groups (or nationally) but a steady decline for all the groups that predated COVID and continued throughout it, largely unchanged. This reduction also sits against the rising knife crimes observed in Police Data since 2016. However, while the impact of COVID was not as marked as expected, it cannot be excluded as a possible alternative explanation for the decreases observed in the treatment group. Chart 3 below shows monthly offending by control and treatment groups between 2016 and 2022, and the time periods when the two were selected. Chart 3: KC tagged group versus control - Number of offences per month 2016-2022 decrease in the number of victims of violent crime, largely driven by falls in violence where the offender was a stranger, in part reflecting the closure of the night-time economy for several months of the year’. Similarly, a study of the impact of the COVID-19 pandemic on the number of presentations of penetrating injuries to a UK major trauma centre (Kings College Hospital) compared the number of patients attending the emergency department with a penetrating injury (gunshot or stab wound) during the ‘pandemic year’ (1 March 2020–28 February 2021) with the equivalent time period in the previous year. The results showed that during lockdown 1 there was a 48.45% reduction in presentations in the ‘pandemic year’ compared to the previous year, lockdown 2 showed a 31.25% reduction; however, lockdown 3 showed an 8.89% increase in the number of presentations, suggesting a return to normal levels by this point (Hickland et al, 2022). Control Treatment 31 A range of analysis was undertaken across different time points, including before the licence; on the licence; after the licence start date; and after licence end date. As the Control group was not standardly subject to ‘something’ following prison release, the most useful comparison is from ‘point zero’ (i.e., when the GPS individual was tagged, or their match was released from prison) to the standardised number of days – either before or after the date of release. The time points used included offending before the licence; on the licence (unique to the individual); then fixed periods after the licence start date; and after the licence end date of 30, 60, 90 days, 12 months, and 24 months. Initially a predicted versus actual offending analysis was undertaken comparing the actual levels of offending to OGRS2 estimates. The average OGRS2 score for both groups was 63%. Results indicate both groups actual offending is lower than their OGRS would predict and there is a significant difference between the GPS group and the Control (GPS 54% vs. Control 59%). The results show that the two groups mirror each other – with the mean number of offences dropping between 24 months prior - to 30 days after, then slowly increasing over the next 2 years, although 24 months after point zero neither group have returned to their 24 months prior levels of offending, see chart 4. Chart 4: Offending by treatment & control group during period 2 years prior to 2 years after licence start date. Although the GPS group was always lower than the control in terms of the proportion to offend; the number of offences; and the mean number of offences, these were not significantly different within 30, 60 and 90 days of the licence start date. However, significant differences were observed at the post 12- and 24-month stages. Comparing the 12 months following the start of the licence for individuals with available data (n=470), there is a significant difference between the GPS group versus the control in terms 2.9 1.2 1.3 1.2 1.6 2.0 3.5 1.4 1.6 1.5 2.0 2.8 24 months before start date 30 days from start 60 days from start 90 days from start 12 months from start 24 months from start Mean number of offences Experimental Group Control Group 32 of the proportion to offend (38% versus 42%), with the GPS group conducting fewer overall offences by this time (291 vs 404) and with a lower mean number (1.6 vs 2.0). Likewise, at 24 months post the start of the licence, for individuals where the data is available (n=244), there is still a significant difference between the GPS group versus the control in terms of the proportion to offend (54% vs 59%); the number of offences (258 vs. 403) and mean number of offences (2.0 vs. 2.8). See chart 5. Chart 5: Offending by treatment and control group during period 2 years prior and 2 years after licence start date. 61 4 9 12 38 54 61 6 11 16 42 59 0 10 20 30 40 50 60 70 24 months before start date 30 days from start 60 days from start 90 days from start 12 months from start 24 months from start % of cohort offending Experimental Group Control Group 33 Discussion The GPS knife crime tag pilot forms an important part of the London Knife Crime strategy, developed in the face of increased levels of knife crime in London and heightened public concern about the problem. It takes place at a time when, according to HMIPs 2022 report ‘there is significant political interest in the use of electronic monitoring to enhance the management of people on probation. Legislation over the last 20 years has enabled it to be used in increasingly inventive and intrusive ways’. The report goes on to conclude that ‘that electronic monitoring has significant potential to bring value to the Probation Service’s work. However, more work is needed before these benefits are fully realised’ and that there is ‘an overall lack of evidence about the longer-term effectiveness of such interventions’. In this environment, MOPAC’s knife crime (and DA) tag pilots, and their evaluations, have an important part to play. Between the beginning of the knife crime pilot in February 2019 to November 2022 (the date of analysis) a total of 971 GPS tags had been imposed, most at the point of prison release (75%, n=726), with a quarter (25%, n=243) being imposed as part of a licence variation. In terms of the enforceable licence conditions monitored by the GPS tag, all bar one of the tag wearers were subject to trail monitoring, 54% (n=525) were subject to at least one exclusion zone, 27% (n=258) were subject to a requirement to attend a specified location, and 24% (n=231) were subject to a condition to report at an approved premise. The mean planned tagging period was 138 days (minimum 10 days, maximum 365 days). Of the 835 cases which had completed their period of GPS monitoring up to November 2022, 298 ended due to recall and 537 ended without recall, giving an overall rate of completion without recall of 64%. Of the 298 cases ended due to recall, 153 cases were recalled for reasons unrelated to the GPS tag (e.g., poor behaviour, missed probation appointments, a new incident), 53 cases were recalled for multiple reasons with GPS a secondary reason, 37 cases were recalled due to a breach of GPS monitored exclusion zones, 20 cases were recalled primarily for GPS related non-compliance, 18 cases were recalled for a new incident detected through crime mapping or evidenced through an external agency request, and 17 cases recalled because of a breach of other conditions identified using GPS (for example, breach of curfew or failure to reside at an approved address). Taking these reasons for recall into account, the effective compliance rate, that is the proportion of cases completed without recall for failure to comply with only the GPS tag as a licence condition, is 98% (n=815). Exploring impact. One of the main aims of the evaluation was to add to the literature in terms of electronic monitoring and impact. As outlined in the report’s introduction, this is an area that requires more work. Within the report, a variety of approaches were adopted, and analysis undertaken, to try and measure impact. These included predicted versus actual offending comparisons using OGRS; comparisons of offending from the GPS tagged group across time points; and extending these comparisons to a Control group, constructed through Propensity Score Matching. Even though the matching process adopted was successful in generating a strong model, the fact that a retrospective comparison group had to be employed remains an important caveat, especially given the potential influence of COVID during the time period of the analysis. 34 While the above precludes firm statements on impact, there are positive results as to the impact on offending of the tag, particularly at the 12 and 24-month post-tag periods, which should be seen as a positive foundation to build upon. Probation practitioners were overwhelmingly positive about the GPS KC tag. The qualitative research undertaken during this pilot indicates that PPs have a positive view of the GPS tag. There was widespread support for the GPS tag as a means of monitoring the movements of tag-wearers, providing information that could be used in supervision, and the improved ability to manage risk (both to victims and tag-wearers) precisely because of this increased certainty about tag-wearers’ movements. Overall, PPs felt that availability of GPS data had an impact on the nature of supervision between them and the tag-wearer, and on tag-wearers’ behaviour whilst on the tag. A different type of discussion with the tag-wearer was feasible, because of the type of data additionally available (i.e. it was difficult for the tag- wearer to challenge the location data). Given the value seen in the data generated by the GPS tags, there was qualified support from PPs for direct access to the IT system provided by the tag contractor (Buddi), although it was clear that PPs’ understanding of the GPS data and its potential uses varied. PPs recognised this in asking for ongoing training and support to promote the value of the data and enhance knowledge and confidence around the ways and formats in which data can be requested/accessed. There was also support from PPs for the extension of the use of the GPS tag for other offence types, particularly for domestic abuse, county-lines and gang-related offences, indeed awareness of this support provided part of the justification for MOPAC to extend the use of tags to DA offences in April 2021. However, this extension needs to be considered in the light of concerns around net-widening and the ethical integrity of utilising GPS tags, with the increased scrutiny (and likely detection) they provide, for some offences but not others – a point already made by one of the service users interviewed. It should also be recognised that the knife crime tag is being used disproportionately against young black males something which MOPAC have identified in their training for practitioners and have sought to address.46 PPs suggested that wearing the tag may have a short-term deterrent effect although were not convinced the tag would bring about longer-term change in offending behaviour. The findings from the impact analysis undertaken for this evaluation did not support these views. In the opinion of PPs, any change for the tag-wearer was largely dependent upon the tag- wearer’s motivation and their willingness to change – to take advantage of the ‘window’ that the tag provides. Offenders were largely positive about the tag. The majority (60%, n=of 382) of tag wearers who completed the end-point survey were positive regarding their overall experience of the tag. The majority also reported they felt the tag had a positive impact on their offending behaviour (83%, n=529) and towards their life in general (60%, n=382). There does appear to be a recognition on the part of some tag-wearers of the benefits the tag provides in providing a means of avoiding risky situations/individuals, if they have the inclination to do so, echoing the perception of PPs about the importance of 46 Figures for March 2022 show that 51% of the recipients of knife crime tags were Black, with 24% White, 13% Mixed, 7% Asian and 5% Other. This compares to 54% of tag wearers in February 2021 being Black. Young adults comprised 57% of the cohort in March 2022, compared to 49% in February 2021. 35 motivation among tag-wearer. The tag-wearers also demonstrated a wide range of understanding about the tag, and while the provision of information to them has got better during the pilot, there remains scope for improvement prior to and during the order, although the impact of Covid during the course of the pilot on the ability of PPs to contact (potential) tag wearers should be borne in mind. Improvements during the pilot’s implementation. The evaluation also identified some issues around the administration of the tag, particularly in the early days of its operation, which have been addressed to a greater or lesser extent as the pilot has progressed. There were initial difficulties in arranging tags to be fitted in prison on the day of release. Concerns were expressed by PPs, particularly in the early stages of the pilot, about the utility of some of the outputs routinely provided to them (which has been amended in the light of these concerns) and which has led to PPs being given live access to the Buddi system. Clearly the extent to which PPs engage with, and use, the data from the tag wearers varies from individual to individual, depending on their level of competence, and the individual being supervised. Issues arising from the charging of the tag, and the ability of PPs to establish whether the tag was faulty/not charging in the face of claims from the tag-wearer have also been identified. The latter has led to improved guidance from Buddi/MOPAC but still appears to be a concern in some cases. The reliance of the PP on the tag in making decisions about breaches/recalls also points to the importance of the technical accuracy of the system, and the importance of PPs having clear guidance and access to specialist advice on the reliability of the data and the interpretation of this at key decision points. Nonetheless, the PPs and tag wearers were happy overall with the service received from Buddi. The value of crime mapping. The results from the police crime mapping process were less promising than anticipated. Between November 2019 and August 2023 232,353 crimes were processed, which generated 3,418 initial matches. However, following subsequent analysis by the MPS team, only 80 (2%) of these were referred on to BCUs for subsequent investigation, resulting in 2 prosecutions (one in October 2020 and one in December 2022).47 These findings are consistent with those found in the previous interim evaluation report, and in E&I’s 2018 Interim Report on the use of GPS tagging for an integrated offender management (IOM) cohort. Results from the interviews with MPS staff suggested that they valued the ability to obtain information about the movements of those on tags via EAR requests above the automated generation of hits. There are several possible explanations for the low number of ‘significant matches’ generated by the crime mapping process. One possible explanation for the low number of significant matches is that offenders subject to GPS monitoring are committing fewer offences whilst on the tag, meaning that the lack of matches reflects the deterrent effect of the tag (consistent with PP feedback). The impact analysis undertaken for this evaluation provides limited support for this supposition – the difference between the offending on tag by the treatment group and that of the control group at a similar point post release was not significantly different48 but both groups were offending at a lower rate than their 2 year OGRS would 47 The percentage of initial matches against the number of crimes input has never been higher than 2.75% (February 2023), and at its lowest fell to 0.42%. 48 It was lower for the treatment group, but not significantly so, and as discussed elsewhere in this report, this cannot be solely attributed to the impact of the tag. 36 suggest.49 Other explanations for the less than expected numbers of significant matches and subsequent convictions could include that tag wearers are committing offences unlikely to be reported to the police or that issues with how crimes are recorded and logged (in terms of a specific time and location) may have impaired the matching process. Given that tag wearers, as well as police and probation practitioners, all report a perceived deterrent effect from the tag wearing, it is possible that this may have influenced tag wearers, where they did offend, to change their type of offending to be less likely to be reported to the police in a way that could be crime mapped. Conclusion. MOPAC’s evaluation of GPS use has spanned a period of 6 years with 5 published reports, prior to this one. This report explores performance, process, and impact, seeking to bring the evaluation to a conclusion. Findings indicate the scheme has a high level of compliance; that delivery improved over the duration of the programme; and that both staff and tag users were generally positive about the technology, and whilst there were positive indications as to the impact on offending, methodological limitations preclude firm statements of success. 49 A fuller account of the crime mapping process will be found in MOPAC’s 2020 Interim Evaluation Report for the GPS knife crime tag available at https://www.london.gov.uk/sites/default/files/gps_tagging_knife_crime_on_licence_final_for_publication.pdf 37 References Andersen, S., and Telle, K. (2016) “Electronic Monitoring and Recidivism: Quasi-experimental Evidence from Norway”, Discussion Paper No. 844. Statistics Norway Research Department. Boiteux, S. & Teperski, A (2023) An evaluation of the NSW Domestic Violence Electronic Monitoring program. New South Wales Bureau of Crime Statistics and Research, Crime and Justice Bulletin, number 255 May 2023. Di Tella R, Schargrodsky E. (2013) Criminal Recidivism after Prison and Electronic Monitoring. Journal of Political Economy. 121(1):28–73. Forrester, V. and Read, T (2020) ‘GPS Knife Crime Tagging Interim Evaluation Report’. MOPAC Evidence and Insight Unit. Grenet, J., Grönqvist, H., & Niknami, S. (2022) “The Effect of Electronic Monitoring on Offenders and their Families” Swedish Institute for Social Research Working Paper 3/2022. Henneguelle, A., Benjamin Monnery, B., & Kensey, A. (2016) “Better at Home than in Prison? The Effects of Electronic Monitoring on Recidivism in France,” The Journal of Law and Economics, 59(3): 629-667. Her Majesties Inspectorate of Probation (2022) ’The use of electronic monitoring as a tool for the Probation Service in reducing reoffending and managing risk’. Hickland., M. M., Massouh, P., Sutthakorn, R. E., Greenslade, C., Jennings, C., Cantle, F. & Bew, D. (2022) The impact of the COVID-19 pandemic on the number of presentations of penetrating injuries to a UK major trauma centre. J Public Health (Oxf). 2022 Mar; 44(1): e126–e132. Hobson, Z., Harrison, A. & Duckworth, L. (2018c) ‘Alcohol Abstinence Monitoring Requirement: A review of process and performance from Year 2’. MOPAC Evidence and Insight Unit. House of Commons Committee of Public Accounts (2022) ‘Transforming electronic monitoring services.’ Twenty-First Report of Session 2022–23 Report, together with formal minutes relating to the report. Howard, P., Francis, B., Soothill, K. & Humphreys, L. (2009) OGRS 3: the revised Offender Group Reconviction Scale. Ministry of Justice Research Summary 7/09. Marklund, F., & Holmberg, S. (2009) “Effects of Early Release from Prison using Electronic Tagging in Sweden”, Journal of Experimental Criminology 5 (march): 41–61. MOPAC (2018a) ‘GPS Tagging First Year Interim Report’. MOPAC Evidence and Insight Unit. 38 MOPAC (2018b) ‘Review of the Metropolitan Police Service Gangs Matrix’. MOPAC Evidence and Insight Unit. National Audit Office (2022) ‘Electronic monitoring: a progress update. HM Prison & Probation Service’. Office of National Statistics (2018) ‘Crime in England and Wales: year ending December 2017’ Office of National Statistics (2021) ‘Crime in England and Wales: Year ending June 2021’. https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/crimei nenglandandwales/yearendingjune2021 Pepper, M., and Dawson, P. (2016) ‘Alcohol Abstinence Monitoring Requirement A process review of the proof of concept pilot’. MOPAC Evidence and Insight Unit. Prison Reform Trust, ‘Licence conditions and recall: Indeterminate sentences’ (accessed 16/03/23) at https://prisonreformtrust.org.uk/adviceguide/licence-conditions-and-recall- indeterminate-sentences/ Rivera, R. (2023) “Release, Detain, or Surveil? The Effect of Electronic Monitoring on Defendant Outcomes” Working Paper. Williams, J., & Weatherburn, D. (2020). Can Electronic Monitoring Reduce Reoffending? The Review of Economics and Statistics. 1–46. Appendices Appendix A: PSM matching - comparison of distribution of variables for experimental and control 40 41 42 43 Appendix B – Re-offending patterns 2016 - 2021. MoJ national data – proportion of offenders who reoffend for specific offences MoJ national data – prison releases (quarterly) MoJ national data – Mean number of reoffences for specific offences 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 Jan - Mar 2016 Apr - Jun 2016 Jul - Sep 2016 Oct - Dec 2016 Jan - Mar 2017 Apr - Jun 2017 Jul - Sep 2017 Oct - Dec 2017 Jan - Mar 2018 Apr - Jun 2018 Jul - Sep 2018 Oct - Dec 2018 Jan - Mar 2019 Apr - Jun 2019 Jul - Sep 2019 Oct - Dec 2019 Jan - Mar 2020 Apr - Jun 2020 Jul - Sep 2020 Oct - Dec 2020 Jan - Mar 2021 Apr - Jun 2021 Robbery Possession of weapons VAP Overall 0 5,000 10,000 15,000 20,000 25,000 Thousands 44 MOPAC data – KC tagged group versus control: Number of offences per month MOPAC data – Knife crime tagged versus control 45 MOPAC data – Knife crime tagged group versus control. Mean offences per month MOPAC data – Knife crime tagged group versus control – mean number of offences (3 months) 46 MOPAC data, knife crime tagged group versus control. Proportion of offenders per month. MoJ and MOPAC data combined 47 1 GPS Knife Crime Tagging Interim Evaluation Report February 2020 Valerie Forrester and Tim Read MOPAC Evidence and Insight 2 Contents Executive Summary .................................................................................................................. 3 1. Introduction ......................................................................................................................... 6 GPS tagging evaluation ........................................................................................................... 9 2. Methodology ...................................................................................................................... 11 3. Results ................................................................................................................................ 12 Performance .......................................................................................................................... 12 Qualitative research ............................................................................................................... 14 4. Discussion ........................................................................................................................... 24 5. References…………………………………….………………………………………….27 3 Executive Summary In 2017, the London Mayor’s Office for Policing And Crime (MOPAC) published the London Knife Crime Strategy. Set against the backdrop of increases observed within Knife crime (both Nationally and London), the strategy took a public health approach to tackling knife crime and included a commitment to pilot the use of GPS tagging with knife crime offenders. The overall aim of the pilot is to test GPS as a tool for reducing the likelihood of offenders committing further offences, especially weapon-related offences. Specifically, this includes the use of GPS location data: • to improve the management of the risk posed to known adults, children and the public, • to improve the enforcement of licence conditions and increase deterrence, • to challenge the offenders’ thinking and lifestyle and improve rehabilitation; and • to enable the MPS to map individual cohort crimes effectively to support crime detection. To be eligible for the MOPAC pilot, an offender must be: • Aged 18+; • Automatically released on a determinate prison licence. In addition, in January 2020 there was agreement in principle for the eligibility criteria to be extended to include cases being considered for re-release after recall by the Parole Board, if GPS conditions were deemed necessary and proportionate; • Serving a sentence for either a knife possession offence or another offence which involved the use of a knife or bladed object; • Released from an eligible prison; • Released before the pilot end date (currently 30th April 2020); • Being released to one of the pilot boroughs (as detailed below); and • Being released to some type of viable accommodation (to enable tag charging). The pilot began in February 2019, in four London boroughs and, following a Mayoral decision, has subsequently been rolled out to cover 24 London boroughs. The pilot is due to finish in April 2020 for new releases, with existing cases being monitored until October 2020. From the beginning of the pilot (February 2019) to the date of analysis (early January 2020) a total of 102 GPS tags had been imposed, the majority at the point of prison release (90%, n=92), with 10 being imposed as part of a licence variation in the community. Most cases were held by the National Probation Service (NPS) (71%, n=74) and the remainder (n=28) held by the London Community Rehabilitation Company (CRC). All but one of the tag wearers were male (n=101) and the mean age of the tag wearers was 26.5 years (SD=9.8). Almost all tag wears were aged under 40 years (89%, n=91), with 64% (n=65) aged 18-24 years, 25% (n=26) aged 25-39 years, 9% (n=9) aged 40-54 years, and only 2 aged 55+ years. Where the ethnicity of the tag wearer was known (n=92), 54% (n=50) were Black or Black British, 24% (n=22) were White, 12% (n=11) were Mixed, 7% (n=6) were Asian or Asian British and 3% (n=3) were Other. 4 In half of cases (50%, n=51) the main offence of the tag wearer was possession of weapons, most commonly having an article with a blade or point in a public place (n=34) and possession of offensive weapons without lawful authority or reasonable excuse (n=14). In 16% of cases (n=16), the main offence of the tag wearer was robbery. In a quarter of cases (25%, n=25), the main offence of the tag wearer was violence against the person, most commonly involving actual bodily harm, wounding or grievous bodily harm (n=17). At the point this report was written, there had been 46 completed cases; 27 cases ended due to recall and 19 cases ended without recall, giving an overall rate of completion without recall of 39%. However, of the 27 cases ended due to recall, 16 cases were recalled for reasons unrelated to the GPS tag (e.g., poor behaviour, missed probation appointments, a new incident), 6 cases were recalled due to failure to comply with licence conditions directly related to the GPS tag (e.g., refusal to have the tag fitted, failure to charge the tag), and 5 cases were recalled due to breach of an exclusion zone being monitored through the GPS tag. Taking these reasons for recall into account, the effective compliance rate, that is the proportion of cases completed without recall for failure to comply with the GPS tag as a licence condition, is 87% (n=40). Fifty-one offenders were subject to exclusion zone monitoring, with five offenders being recalled specifically for breaching this condition, giving an exclusion zone compliance rate of 90%. The research undertaken indicates that probation offender managers (OMs) have a very positive view of the GPS tag. There was widespread support for the GPS tag as a means of monitoring the movements of tag-wearers, providing information that could be used in supervision, and the improved ability to manage risk (both to victims and tag-wearers) precisely because of this increased certainty about tag-wearer’s movements. OMs also felt that there was scope to extend the use of GPS tags to other offences. There was qualified support too from OMs for direct access to the IT system provided by the tag contractor (Buddi), although it was clear that OMs’ understanding of the GPS data and its potential uses varied. Given this, OMs would likely benefit from further training and support to promote the value of the data. The findings from tag-wearers were more ambivalent. Although there appeared to be support for the tag from tag-wearers at the point that the tag was fitted around topics such as their offending behaviour and the impact of the tag on their lifestyle, it was not possible in this report to examine the level of support at the point of removal. However, we also found tag wearers and OMs express the need to receive more information about the tag. This is something that should be addressed as the pilot progresses – especially in terms of further communication to tag-wearers. However, there does appear to be a recognition on the part of some tag-wearers of the benefits that the tag gives in providing a means of avoiding risky situations/individuals, if they have the inclination to do so. The evaluation has also identified some issues around the administration of the tag, particularly in the early days of its operation. There were initial difficulties in arranging tags to be fitted in prison on the day of release. Issues, too, relating to the charging of the tag, and the ability of OMs to establish whether the tag was faulty/not charging in the face of claims from the tag- wearer. It would also appear that there is scope to improve communication with offenders about 5 the pilot to ensure that they are aware that they are going to be tagged. Concerns too, were expressed about the utility of some of the output routinely provided to OMs (which has been amended in the light of these concerns). Nonetheless, the OMs appeared to be very content overall with the service received from Buddi. Overall, OMs felt that availability of GPS data had an impact on the nature of supervision between them and the tag-wearer, and on tag-wearers’ behaviour whilst on the tag. A different type of discussion with the tag-wearer was feasible, because of the type of data additionally available (i.e., it was difficult for the tag-wearer to challenge the location data). OMs also suggested that wearing the tag may have a short-term deterrent effect during the monitoring period, although they were less convinced that the tag would bring about longer-term offending change. In the views of OMs, any change for the tag-wearer was largely dependent upon the tag-wearer’s motivation and their willingness to change – to take advantage of the ‘window’ that the tag provides. Interviews with tag-wearers appeared to support the importance of motivation. Findings from the police crime mapping process are inconclusive at this point. Despite the large number of crimes uploaded and matches assessed, there have been relatively few “significant matches” to date and the GPS information shared with local BCUs has not led to any additional convictions so far. The MPS is refining the process for identifying and triaging automatic matches to ensure that they focus on the cases where there appears to be the best potential for a meaningful match with a greater chance of this being used for investigation and arrests. There will also be a great focus on more serious offences (excluding minor property crimes and focusing on contact crimes). The final evaluation report, to be completed within six months of the pilot’s conclusion, will continue to collect data on the number of tags fitted and levels of compliance by wearers, as well as feedback from tag wearers and OMs. Additional performance and qualitative data on crime mapping will be collected, including throughput of crimes uploaded for mapping to outcomes of matches and feedback from police staff. If the sample size (i.e., the number of tags fitted) permits, it will also examine the impact of the tag on tag-wearers’ offending behaviour against an appropriate control group. 6 1. Introduction In 2017, the London Mayor’s Office for Policing And Crime (MOPAC) published the London Knife Crime Strategy. This was against a background of increases in levels of knife crime nationally (from 2014) and within London (increases begin during 2016). To illustrate, knife crime rose steeply within 2017 - England & Wales (+22%), London (+30.5%) (ONS 2018, Met Stats). Even with a stabilisation during 2019 (when there was a 6% increase), the issue remains key for London. Londoners’ perceptions of knife crime as a problem in their local area has also increased over recent years (i.e., from 23% in Q2 17-18, to 38% in Q3 19/20 – as measured by the MOPAC Public Attitude Survey). The Strategy took a public health approach to tackling knife crime and included a commitment to pilot the use of GPS tagging with knife crime offenders. The pilot was endorsed and supported by joint work with the Ministry of Justice (MoJ) and designed to work alongside the MoJ’s national GPS programme, ensuring there was no overlap or duplication between the MoJ and MOPAC GPS provision. Details of the differences between the MOPAC and MoJ programmes are identified below. Table 1. Differences between the MOPAC and MOJ GPS pilots MOPAC knife crime GPS pilot MoJ national GPS programme Only available for knife crime offenders on determinate licence Available for any offence type and for sentence types including community and indeterminate prison sentences Only available for offenders returning to pilot London boroughs Available nationally Pilot end date currently 30th April 2020 for new releases No end date for this programme. The pilot began in February 2019 in four London boroughs and has subsequently been rolled out to cover 24 London boroughs as detailed in table 2 below. The pilot is due to finish in April 2020 for new releases. Table 2. Rollout of the GPS knife crime pilot Launch date Boroughs Total pilot boroughs February 2019 Croydon, Lambeth, Lewisham, Southwark 4 June 2019 Greenwich, Hackney, Newham, Tower Hamlets, Waltham Forest, Westminster 10 August 2019 Barking & Dagenham, Camden, Enfield, Haringey, Islington, Redbridge 16 September 2019 Barnet, Brent, Ealing, Hammersmith & Fulham, Kensington & Chelsea, Wandsworth 22 November 2019 Bromley, Hounslow 24 The GPS tag is part of the tag-wearer’s licence conditions and is usually fitted at release from prison. The length of the tag is specified by the OM, up to a maximum of 6 months. 7 Eligibility criteria To be eligible for the MOPAC pilot an offender must be: • Aged 18+; • Automatically released on a determinate prison licence. In addition, in January 2020 there was agreement in principle for the eligibility criteria to be extended to include cases being considered for re-release after recall by the Parole Board, if GPS conditions were deemed necessary and proportionate; • Serving a sentence for either a knife possession offence or another offence which involved the use of a knife or bladed object; • Released from an eligible prison1; • Released before the pilot end date; • Being released to one of the pilot boroughs; and • Being released to some type of viable accommodation (to enable tag charging). Aims of the pilot The overall aim of the pilot is to test GPS as a tool for reducing the likelihood of offenders committing further weapon-related offences. Specifically, this includes the use of GPS location data: • to improve the management of the risk posed to known adults, children and the public, • to improve the enforcement of licence conditions and increase deterrence, • to challenge the offenders’ thinking and lifestyle and improve rehabilitation; and • to enable the MPS to map individual cohort crimes effectively to support crime detection. Key partners in this pilot include the National Probation Service (NPS) (London division), London Community Rehabilitation Company (CRC), the Metropolitan Police (MPS), Her Majesty’s Prison and Probation Service (HMPPS) and Buddi Ltd (the GPS monitoring provider). Background to the pilot GPS monitoring was piloted by MOPAC with persistent and knife crime offenders on community sentences in eight North and East London boroughs between 2017 and 2019. Building upon the learning, MOPAC, with support from MoJ, extended the use of GPS tags to individuals convicted of knife crime offences released from prison on licence, providing location data to probation staff to assist them in monitoring specific licence conditions, addressing the subject’s offending behaviour and managing risk of serious harm. In the pilot, OMs are encouraged to use GPS data openly with service users in supervision meetings, to examine and challenge lifestyle and risk factors linked to offending, especially knife crime. It is anticipated that GPS will be used to monitor compliance with exclusion zones and attendance at rehabilitation programmes and 1 Eligible releasing prisons are - Belmarsh, Brixton, Bronzefield, Coldingley, Downview, Feltham, High Down, Highpoint, Isis, Onley, Pentonville, Send, Thameside, Wandsworth, and Wormwood Scrubs 8 treatment. Data can also be provided about bespoke and flexible key locations for the individual via real time notifications and summary reports, such as heat maps and top addresses visited, provided by the monitoring provider to the OM. Data provided to OMs by the tagging provider OMs are provided with real time email alerts in the event of their service user breaching an exclusion zone or entering an interest zone. When the tag reaches 30%, the wearer will receive a series of automated text messages and phone calls reminding them to charge the tag for at least an hour, if the tag has still not been placed on charge after two hours the offender manager will receive a notification of the breach.2. In addition to the alerts, OMs are sent a weekly summary report by email which provides details of the service user’s charging pattern, and the ‘top addresses’ visited by the tag-wearer during the week (the format of the latter was amended in the light of comments received from OMs to provide more detail about the location, and the duration of the tag-wearer’s visit to the address). If applicable, OMs also receive respective details of curfew violations (as this is a licence condition rather than a GPS licence condition, OMs do not receive real time alerts for these breaches). Figure 1. Example of Buddi patented Heatmap The OM is also able to request data on an ad-hoc basis from Buddi. This could be in the form of Buddi patented heat maps (indicating the most popular locations visited by a service user over a certain period in a specific area – see Figure 1), trail maps (showing details of the tag-wearer’s movements, or details of the tag-wearer’s visits to specific addresses/locations). Since November 2019 some OMs have been given direct access to the Buddi system via the secure customer portal, allowing them to produce the various outputs above as required without having to request this from Buddi. This access was initially offered to all NPS OMs who had live cases at the time (15 OMs), about half of whom accepted. 2 There is also the facility for Buddi to phone the on-call emergency probation manager in the event of an emergency out-of- hours breach of an exclusion zone. 9 Exclusion/inclusion and interest zones As part of the licence conditions for the GPS tag it is possible for the OM to specify an area or areas which the tag-wearer cannot enter or leave whilst on licence. PSI 12/20153 states that ‘the purpose of an exclusion zone condition must be clear and necessary, and the size of the exclusion zone reasonable and proportionate’ (Prison Reform Trust). These zones can be a variety of types. Exclusion zones are where the OM is provided with an alert immediately as the tag-wearer enters the specified area(s). A variation of this is the exclusion zone with a ‘grace period’ attached – where the OM/tag-wearer is not alerted until a specified period has elapsed (typically requested by the OM if, for example, the tag-wearer has to pass through the exclusion zone for work/when travelling). There are also inclusion zones, where there is an alert when the tag-wearer moves outside a specified area, inclusion zones with a grace period, and interest zones (zones which generate an alert for the OM but are not linked to an enforceable licence condition). GPS monitoring and crime mapping Crime mapping allows eligible offenders to have their movements automatically cross referenced against the location of reported crimes. Data is only released when a match is confirmed, and this is then forwarded on to local police for investigation. Crime mapping is only carried out for offenders assessed as “more likely than not” to reoffend within 2 years, that is individuals with an OGRS (Offender Group Reconviction Scale)4 score of 50%+. Crime mapping has been taking place since June 2019, and since November 2019 has been undertaken by a dedicated analyst based in the MPS Central Intelligence team. The post is funded by MOPAC until October 2020. The MPS and other agencies can also request specific GPS data through an External Agency Request (EAR) process. This can be used to assist in investigation and conviction. Where justified through a RIPA (Regulation Investigatory Powers Act) request it is also possible for police teams to gain direct access to the Buddi system to live monitor the location of an offender on tag. GPS tagging evaluation The Evidence and Insight team at MOPAC were asked to evaluate the GPS knife crime pilot, capturing learning around the implementation and design of the pilot, gaining insight into the experiences of practitioners and tag wearers, and measuring success. The overarching aims of the evaluation are to examine: • Performance - monitor the key performance indicators of the innovation, including numbers receiving the tag, types of orders, compliance rates, violations, breaches, order completions, crime mapping hits, crime mapping outcomes. 3 Prison Service Instructions (PSIs) outline the rules, regulations and guidelines by which prisons are run. 4 The Offender Group Reconviction Scale (OGRS) is a risk assessment tool used to estimate likelihood of re-offending using static factors such as age, gender and criminal history. It gives a score, which shows the likelihood of someone re-offending over a 1 year or 2-year period, expressed as a percentage or on a range from 0 to 1. A lower score means a lower likelihood of re-offending. 10 • Process - understand the experiences of practitioners and tag wearers involved in the pilot through surveys and interviews. • Impact – the final evaluation will seek to explore the impact (if any) of the GPS tag upon offending behaviour (assuming the number of tags sufficiently large). The evaluation will also capture indicators of success and will use qualitative methods to gain learning about the success of the pilot. This report shares findings and learning from the first year of the GPS tagging pilot. It consists of performance and process aspects, including feedback from OMs and tag wearers. The final evaluation report, to be completed within six months of the pilot’s conclusion, will continue to collect data on the number of tags fitted and levels of compliance by wearers, as well as feedback from tag wearers, OMs. Additional performance and qualitative data on crime mapping will be collected, including throughput of crimes uploaded for mapping to outcomes of matches and feedback from police staff. If the sample size (i.e., the number of tags fitted) permits, it will also examine the impact of the tag on tag-wearers’ offending behaviour against an appropriate control group. 11 2. Methodology This section outlines the approach undertaken in the evaluation. Performance Performance data was gathered from a range of sources, mainly the tagging provider Buddi and MOPAC delivery staff. This data included information about the tag wearers and the types of orders they were on, and their compliance with the tag. In addition, data was provided by the MPS about the crime mapping element of the pilot (i.e., numbers of crimes up-loaded, numbers of matches, and subsequent actions taken by the police). Tag wearer survey Tag wearers are asked to complete a short survey at the point the tag is fitted and again when removed. The survey was designed by E&I and consists of 3 topics; the surveys aim to collect the wearers’ anticipated and actual perception of the impact of wearing the tag on their behaviour, their relations with friends and family, and their lifestyle. Data were available from 50 individuals where a tag had been fitted, but only 5 cases where the tag had been removed, the difference in the numbers being due to many tags still being active. Because of the small number of removal cases, these data have not been analysed for this report but will be included in the final evaluation report as numbers increase. Surveys are administered by the tag fitter (staff members from Buddi). Interviews with offender managers and tag wearers Semi-structured interviews were undertaken with 12 OMs who had supervised those on the GPS knife crime tags. Those interviewed were asked about details of the specific case they had supervised, how they had used data from the tag in supervision, and for their views of the GPS tag in general. In addition, semi-structured interviews were completed with a small number of service users (three individuals) who had been tagged. The interview asked about their experiences wearing the tag, their understanding of why they had been tagged, and their views of the tag overall. 12 3. Results This section of the report presents the early findings from the evaluation, covering aspects such as basic performance, tag wearer demographics, compliance with the tag, as well as the findings from the qualitative research carried out with OMs and tag-wearers about their perceptions of the tag’s operation. Performance As of early March, 152 GPS tags have been imposed. For the purpose of this interim report, the analysis below covers GPS tags imposed up to the week commencing 6th January 2020. GPS tags imposed Since the beginning of the pilot in February 2019 to the date of analysis, a total of 102 GPS tags had been imposed, the majority at the point of prison release (90%, n=92), with 10 being imposed as part of a licence variation. Most cases were held by the NPS (71%, n=74) and the remainder (n=28) held by CRC. Tag wearers were most commonly released from Thameside Prison (n=21), followed by Feltham (n=16), Pentonville (n=14), Isis (n=12) and Brixton (n=11). The number of tag wearers released from Feltham and Isis, which focus on young adult offenders, reflects the fact that almost two-thirds the tag wearers were aged under 25 years. In terms of the enforceable licence conditions monitored by the GPS tag, 80 tag wearers were subject to trail monitoring, 51 were subject to at least one exclusion zone, 35 were subject to a requirement to attend a specified location, and 14 were subject to a condition to report at an approved premise. Characteristics of tag wearers All but one of the tag wearers were male (n=101) and the mean age of the tag wearers was 26.5 years (SD=9.8). Almost all tag wears were aged under 40 years (89%, n=91), with 64% (n=65) aged 18-24 years, 25% (n=26) aged 25-39 years, 9% (n=9) aged 40-54 years, and only 2 aged 55+ years. Where the ethnicity of the tag wearer is known (n=92), 54% (n=50) were Black or Black British, 24% (n=22) were White, 12% (n=11) were Mixed, 7% (n=6) were Asian or Asian British and 3% (n=3) were Other5. In half of cases (50%, n=51) the main offence of the tag wearer was possession of weapons, most commonly having an article with a blade or point in a public place (n=34) and possession of offensive weapons without lawful authority or reasonable excuse (n=14). In 16% of cases (n=16), the main offence of the tag wearer was robbery. In a quarter of cases (25%, n=25), the main offence of the tag wearer was violence against the person, most commonly an offence involving actual bodily harm, wounding or grievous bodily harm (n=17). The mean OGRS score (2 year) of tag wearers was 62%, which places the group on average at medium risk of offending6, but there was a wide range of scores across the cohort (minimum of 9, maximum of 97). Older 5 In terms of disproportionality, although BAME individuals represent 41% of the London population overall, they represent 54% of the population under 25. Figures from MPS CRIS data for the FY 2017/18 indicate that those of black African- Caribbean ethnicity comprise 46% of all knife crime offenders, and 50% of knife crime possession offenders (MOPAC 2018b) 6 OGRS scores are banded into ‘low’ (49% or below), ‘medium’ (50-74), ‘high’ (75-89) or ‘very high’ (90-99). 13 tag wearers tended to have a lower OGRS score, with the mean score of those aged 40+ years being 48%. The OGRS score is also important because the crime mapping facility is only available for tag wearers with a score over 50. Compliance with the GPS tag Up to the week commencing 6 January 2020, of the 102 total tags - 56 cases were live. Looking at the 46 cases who had completed their period of GPS monitoring, 27 ended due to recall and 19 ended without recall, giving an overall rate of completion without recall of 39%. Of the 27 cases ended due to recall, 16 cases were recalled for reasons unrelated to the GPS tag (e.g., poor behaviour, missed probation appointments, a new incident), 6 cases were recalled due to failure to comply with licence conditions directly related to the GPS tag (e.g., refusal to have the tag fitted, failure to charge the tag), and 5 cases were recalled due to breach of an exclusion zone being monitored through the GPS tag. Taking these reasons for recall into account, the effective compliance rate, that is the proportion of cases completed without recall for failure to comply with the GPS tag as a licence condition, is 87% (n=40)7. This is a relatively high level of compliance8, and in all likelihood can be seen as a positive - although in itself is not a measure of ‘impact’. Fifty-one offenders were subject to exclusion zone monitoring, with five offenders being recalled specifically for breaching this condition, giving an exclusion zone compliance rate of 90%. Diagram 1. Completed cases and compliance with GPS tag as a licence condition MPS crime mapping To date, nearly 160,000 crime reports have been uploaded for crime mapping, with the reports mapped against the movements of those subject to a GPS tag, with any matches being automatically identified (where the tag-wearer had come within a 200m radius of the crime 7 As a compliance rate, this is higher than that found by MOPAC in the use of GPS tagging with an IOM cohort (56%) but slightly lower than that for AAMR (94%). It should be noted however, that the characteristics, and offending behaviour of the individuals who receive these various interventions differ. 8 This compares against a compliance rate of 66% for licence and post-sentence supervision cases supervised nationally by the CRC (July-September 2019). Figures are not available for the NPS. 46 Completed cases 40 Cases completed without recall for failure to comply with GPS tag 19 Cases ended without recall 16 Cases recalled for reasons primarily unrelated to GPS tag 5 Cases recalled due to breach of monitored exclusion zone 6 Cases recalled due to failure to comply with GPS tag as a licence condition 14 location during the timeframe of the offence). Over 1,005 automatic matches have been made and examined so far, to make an initial assessment of how likely the tag wearer was to have been actually involved in the reported offence. During this assessment the dedicated MPS analyst took in to account factors such as the suspect description or tag-wearer’s speed of movement, for example, to see whether they had been at the scene of the crime long enough to commit the offence. Following this ‘triage’, plausible ‘significant matches’ were forwarded on to basic command units (BCUs) for further investigation. As of an MPS update in December 2019, 96 ‘significant matches’ (10% of the total matches) had been forwarded to BCUs. The most common offence types for significant matches were burglary (48%, n=46) and vehicle crime (25%, n=24). 84% (n=81) of matched offences were theft offences (including taking from/of motor vehicles). To date these this process has not led to any charges or conviction. Factors that might explain the low level of outcomes from the crime mapping procedure are identified in the discussion section of this report. There have been 15 External Agency Requests (EARs) for GPS data made by the MPS to Buddi, and there has been one case of the MPS having direct access to the Buddi system to live monitor the location of an offender subject to a GPS tag granted through a Regulation of Investigatory Powers Act (RIPA) request. Qualitative research The following themes are taken from qualitative research undertaken with tag-wearers and OMs at different stages of the tagging period; a survey completed with tag-wearers at the point that the tag was fitted9, followed by interviews undertaken with OMs and tag-wearers shortly after the tag. Interviews with tag-wearers were only undertaken with those who were not recalled and were in the community, interviews with OMs were undertaken regardless of whether the individual they supervised was recalled or not. General views of the tag OMs overwhelmingly expressed positive views about the GPS tag.10 It was suggested that tag ‘provided structure and discipline’ for the tag-wearers; provided more information and enabled more monitoring than a home detention curfew (HDC) tag; allowed better management of the case; opened up conversations with the tag-wearer (around their movements and safety); provided the ‘perfect alternative to recall’; and gave the OM a greater ability to protect the victim. Interestingly, two of the OMs indicated that they were now supportive of the GPS tag despite having initial concerns. Indeed, one of these individuals suggested that one of the dangers with the use of the tag was that OMs would be become over-reliant on it. Views from the tag-wearers were more considered, although the responses from the tag-fitting survey were positive. When asked whether they thought having the GPS tag would, in general, make their life better, worse or have no impact, the majority answered it would make their life 9 There was also the facility in the tag fitting survey for the service user to provide general comments about the GPS tag. Responses were received from 35 of the 50 respondents. 10 There was one OM who said that her case, had dissuaded her from using a GPS tag in future, but this was a service user who had been recalled after 3 weeks for never engaging and had never charged his tag. 15 better (n=35) and only six thought the GPS tag would have a negative impact. In terms of the anticipated impact of the tag on their offending behaviour, again, most respondents felt it would have a beneficial impact (n=39); no respondents thought the GPS tag would make their offending behaviour worse. Asked about the impact of having the GPS tag on specific aspects of their life, however, respondents were more ambivalent, mostly answering that it would have no impact for all aspects except for their ability to socialise (see chart 1 below). In terms of the perceived benefits of the tag, the most popular response to emerge from the free text on the tagging survey was that the GPS tag would provide the tag-wearer with the opportunity/justification to keep away from trouble and/or keep away from certain people, something that also emerged from the interviews. Other benefits mentioned in the free text were that the tag was better than a curfew tag, better than being in prison, would assist in retaining (or obtaining) secure accommodation, and would establish that the tag-wearer was not in a certain area, and hence unable to commit crimes. The last point was something that was stressed also by OMs in interview, for example: “we would say 18 16 14 7 7 18 14 25 4 7 9 2 1 3 1 9 5 1 24 24 33 40 38 29 25 19 42 1 1 1 2 2 2 2 1 3 0 5 10 15 20 25 30 35 40 45 50 Relationship with family Relationship with friends Financial situation Physical health Mental wellbeing Housing situation Employment situation Ability to socialise Education situation Chart 1. Tag wearers' perception of anticipated impact of wearing GPS tag on ... Better Worse No impact N/A “‘It had a good effect on me in terms of the whole element of people I knew before and would meet and would end up getting into trouble with, and it was like ‘You’re on a tag, I’m not meeting you! Fuck that! I’m not coming to see you, bro, you’re OPO [on police observation]’ and then I was like ‘Cool, don’t come and see me then bro, I’m not doing anything that I shouldn’t….’ That just shows me that all of them people that were saying that they didn’t want to meet me, means that they are probably still doing the same shit, you know what I mean?” GPS tag wearer 16 to them … if there was an incident that took place, I don’t know, in Brixton, and somebody said they saw you, well actually we could provide evidence to say, well, no you weren’t there because GPS says that you were here ... I think they liked that side of it, that bit they took on board definitely.”. Other factors, while identified as benefits by some survey respondents, were considered costs by others. Thus, while some respondents felt that the tag would increase their employment opportunities/allow them to retain employment, others felt that the tag would have costs as far as their employment was concerned (preventing them wearing work boots and prospective employers refusing work because of the GPS tag). Similarly, while some respondents mentioned that they felt that wearing the tag would have a positive impact on their family or family relations (for example, improving contact with their daughter, allowing them to care for their mother), a similar number suggested that wearing the tag would be a source of embarrassment (either to themselves when out in public or to their family). The interviewed tag wearers held markedly different views as to the overall benefit of the tag. One was highly supportive (for the reasons outline above), the others were strongly opposed to the tag (uncomfortable, invasion of privacy and a breach of human rights). One of the opposed tag-wearers was also concerned with how the data would be used by probation and the police once he had completed his term on licence. The interviewed tag-wearers’ views about the perceived fairness of the tag also differed markedly. One felt that the GPS tag should be reserved for terrorists, murderers and rapists, ‘not people like me’. In addition, he observed that other people he knew had committed more serious offences than him but had not received a GPS tag. He did not think that tags would prevent knife crime, rather they would just allow the police to arrest people subsequently, young black males in particular. Conversely, the individual who had stressed the benefits of the tag in keeping him away from previous associates, described his feelings in the following terms: Interestingly, OMs’ perceptions of the views the tag-wearers they supervised held about the tag also varied. While five interviewees felt that their (six) service users were ‘fine with’ or ‘not bothered’ by the tag, seven tag-wearers were felt to be unhappy being on the tag. However, for four of these seven, the OMs felt that initial hostility had been replaced by later acceptance. Operation of the pilot Most OMs appeared clear about the eligibility criteria for the pilot, and why the tag-wearer had been recommended for inclusion, referring to the tag-wearer’s current and/or previous knife- related offence – ‘[he] screened perfectly for it’. Several OMs referred to the additional oversight “At first, I was like, woah, it’s weird that someone can see where you go 24/7, but then when I looked at the situation … I understood it is only right that I am let out of prison on a tag. I was let out of prison and I absconded for a year, so of course I need to be on something so that they can see what I am doing and that I am not out here doing things that I shouldn’t be doing.” GPS tag wearer 17 that the GPS tag provided, particularly around exclusion zones, while one OM stated that the tag had been recommended because it avoided the tag-wearer having to be placed in an approved premise (AP). As one OM said: “His offending was gang-related, there was weapons offences, possession of knives, firearms previously. It was my thinking of let me have a bit of a measure on his movements and then if I could see a lot of erratic movements around the place, then that might suggest that he may be up to something – acquisitive offending, whether that be dealing or at worst going to seek out altercations with rivals. I couldn’t see a lot of other protective elements that I had to work with this guy. I felt that this was the only possible way that I would have tabs on what he was getting up to, I found it almost a bit of a godsend in that way because I wouldn’t have had anything else”. Some OMs queried why the use of GPS tagging had been limited to knife-crime offenders, and most felt there was the scope to extend the use of the GPS tag to other offence types. Amongst the offences identified were domestic abuse and sexual offences (the most popular responses - although a couple stressed this was primarily for the management of risk to the victim); child sexual abuse; serial burglary; serious organised crime; serious group/gang-related offending; firearms offences; and harassment/stalking. One OM suggested that the rollout of the pilot across London boroughs had improved the operation of the tag. In the early stages, when limited to a few boroughs, a lot of otherwise eligible cases (for example, high-risk younger offenders with peripheral or entrenched involvement with serious group offending) were being missed because their return to home areas would not be supported if that was where their ‘gang’ affiliation was, and it was not possible to place them in boroughs outside the pilot areas. As the pilot was rolled out to a greater number of boroughs, most approved premises (APs) were covered, allowing these offenders to be placed in safer areas with a GPS tag to monitor whether they were going back to risky areas. In terms of the level of information the tag-wearer knew about the tag itself, the evidence we found was generally mixed. To illustrate, all respondents to the tag-fitting survey (50) indicated they understood how the GPS requirement worked and what they had to do to comply with the tag order. The clear majority agreed that they understood both how the tag fitted in with the other licence requirements (n=48) and why they had received the GPS tag (n=43). Most respondents reported feeling very confident (n=29) or fairly confident (n=17) that they would successfully complete the tag order. This also was found within our interviews, to illustrate, one tag-wearer highlighted the value of the document he had been given to read - ‘Before I came out of prison I had a little booklet to read about it and I assumed that my Probation Officer was going to ask me about where I had been … there was one time that I was going to visit my nan but it was in my exclusion zone, so I had to contact my Probation Officer to try and sort something out'. However, in contrast, the research also heard from many tag-users whom were unclear around many aspects of the tag. To demonstrate, in the tag fitting survey, nine of the respondents stated they had not known about the tag and were unaware that it was going to be fitted. A similar theme emerged when talking to OMs, with a number highlighting their tag-wearers did not know about the tag prior to fitting, or not told enough about the tag. 18 These mixed findings would seem to indicate that the good practice there is (i.e., the booklet for tag users), is not yet fully embedded in prison pre-release working practices, resulting in an inconsistent knowledge base. This is something that should be addressed as the pilot progresses – especially in terms of further communication to tag-wearers. Administration of the GPS tag Tag fitting OMs provided details of where the GPS tag had been fitted in 13 cases. Notwithstanding the stated aim of the pilot that tag-wearers should be tagged in prison prior to release, this had only happened in five of the 13 cases11. For the remainder, the tag had been fitted in the probation office, occasionally to the OM’s surprise (” It was the same day [as his release] and they came and tagged him in my office, but I wasn’t expecting that. As far as I knew he was getting tagged on release”). In one case the tag had been fitted in a probation office out of area because the OM’s office (inside the tag-wearer’s exclusion zone) was deemed too risky for the tag-wearer concerned. During the interviews, a couple of OMs suggested that Buddi were not fitting the tags in prisons or appeared unaware that the tag could be fitted in prison. A number of explanations were given for the failure to tag in prison: that Buddi staff had been declined entry, that the tag-wearer had been released early and missed the appointment to be tagged. One OM identified the possible ramifications of this: ‘I just think it would work better if it was done in the prison, because they’re in a different space at that point and they’re just quite keen to come out … It took me almost a month to get the tag on him because he missed … the first appointment for the tag’. Perhaps unsurprisingly, there were few, if any, issues associated with tag removal (one tag-wearer had cut off his tag when it had not been removed at the appointed time). Using the Buddi system Buddi is the service provider that fits and removes the GPS tags, and is responsible for providing monitoring information and details of charging and zone breaches to OMs. Probation staff perceptions of the usefulness of the reports routinely provided by Buddi varied. While some OMs described the weekly report that outlined the tag-wearers’ top locations as useful/sufficient, a slightly larger number described it as ‘poor’/ ‘not useful’. Other OMs praised the ability to obtain bespoke information from Buddi about offender movements and popular locations, and the usefulness of data about curfew and inclusion zone compliance. It was clear that OMs’ understanding of the data that was available from Buddi, it’s meaning, and what could be requested, varied (one OM stated that they did not know how to use the data they received, another that they had only ever received charging information). It is important to stress that OMs had different levels of access to the system and its output. In addition, the format of the reports provided by Buddi changed over time (notably the top locations report), largely in the light of comments received from OMs through this evaluation, to provide more detail about the length of time tag-wearers spent at the ‘most popular’ addresses. 11 The most recently available figures (late February 2020) suggest that the pattern emerging from the OM interviews differs slightly to the pattern overall. Of the 138 tags fitted to date, 82 have been fitted in prison, and 53 at probation (3 cases have not been tagged – either refused or recalled at the gate). This change probably reflects the improvement in Buddi gaining access to prisons to fit tags. 19 When asked if they would like to have direct access to the Buddi system OMs’ opinions similarly differed. While most felt that having this access would be an improvement, suggesting that access to the system would be useful during supervision, and to clarify tag-wearers’ movements quickly in the event of an exclusion zone alert, a couple of OMs, however, did not feel this was necessary. One felt that there was enough information from the existing reports and notifications already. The other said ‘just send me the emails, I can’t do anymore’. The responses tended to reflect the OM’s ease with and understanding of the outputs, and points to the need for improved training and support around the system for OMs. However, the OMs interviewed were extremely complimentary about their contact with Buddi overall, with all the comments received being positive. Interviewees frequently mentioned the ‘helpful and responsive’ nature of the staff (whether via email or telephone) and the tag fitters were described as ‘very professional’. One stated: ‘They were very helpful, the minute he breached an exclusion zone I would get an automatic email, if they were contacting him to locate him, I would get an email, if they were sending alerts to his phone they would let me know, then at the time that I wanted further information, they would find it’. Practical issues of wearing the tag Comfort of the tag In relation to the practical issues of wearing the tag, most tag survey respondents stated that it was comfortable to wear (although it should be stressed that this was at the point that the tag was being fitted). The view from the tag-wearers interviewed after removal were different. Two of the three felt it was uncomfortable (too big and bulky, particularly when the charger was attached), with one interviewee identifying difficulties charging it at night: ‘at night you have to charge it and the wire is very short so if I am asleep and charging it, if I move when it feels like it’s going to come off, so that makes it hard to sleep’12. Some OMs also stated their tag-wearers had remarked on the uncomfortableness and size of the tag. These findings bear some comparison to other electronic monitoring evaluations which record wearer perceptions of the size and fit of tags (Pepper & Dawson, 2016). Charging the tag The ability to keep the tag charged is critical as it allows the location of the tag-wearer to be monitored, and tag-wearers can be recalled in the event of failing to keep the tag charged. This being the case, tag wearers (surveyed and interviewed) and OMs were asked to identify any problems they had experienced with keeping the tag charged. Ten respondents (out of 35) to the tag fitting survey anticipated that it would be easy to use the charger/charge the tag, although again it should be stressed that this view was being expressed at the point that the tag was being fitted and where they would have just been provided with instructions by the Buddi fitter. Equally, even at the point of fitting (where the problems were presumably anticipated 12 Throughout the pilot tag wearers were issued with wireless chargers, which clipped on to the ankle tag, without needing to be simultaneously plugged in. As the quote above shows, some tag-wearers used the equipment in ways that had not been foreseen or recommended. Further guidance for offenders and probation staff has been developed through the pilot so far to try and steer subjects towards a more regular routine of compliance with tag charging. 20 rather than experienced) there were four survey respondents (out of 35) who felt that keeping the tag charged would present difficulties for them because of their circumstances. OMs were similarly asked if their tag-wearers had experienced any difficulties charging the tag. Their experiences differed, although slightly more OMs said this had not been the case than those who did. OMs may have received occasional low battery alerts but not to the extent that warranted warning letters or threats of breach (although in one case Buddi had come out to the probation office to check the tag because there was a suspicion it had been tampered with). The OMs who reported that their tag-wearers had problems charging the tags attributed this to their chaotic lifestyle. Only one OM identified that their tag-wearer had problems charging the tag due to not having settled accommodation. Interestingly, while one OM stated a benefit of the tag had been to assist the tag-wearer in getting a place in approved premises, another OM indicated that the tag had meant the tag-wearer avoided being placed in such accommodation, which he was keen to do so because of his concerns about the risk posed to him by other residents. Another OM suggested that the tag-wearer was claiming that the unit was broken and was deliberately running down the tag to avoid being monitoring. The OM had Buddi replace the battery twice: “after the second time I said that anything else that happens to the battery would mean a warning letter because from the first time I gave him a second chance and by the second I knew that he was doing something to [the battery], I gave him a chance so that he couldn’t say that I was giving him warnings for no reason, but that was the only issue that we had”. In addition, there were a number of OMs who reported problems with faulty tags – two because they or the tag-wearer would be receiving charging alerts when the tag was on charge, the other because of water damage. Several OMs observed that it was difficult to establish the veracity of the tag-wearer’s claims that the tag was faulty, particularly the claim that it was not charging when it should be. The answers given by the three service users to the same question about charging broadly reflected the answers provided by the OMs. Two said they had experienced problems, one that he had not. Experiences, unsurprisingly, were linked to the home circumstances and level of self- organisation of the individual, illustrated by the response from the tag-wearer who said he had experienced no difficulties. “It was getting them to keep the unit charged, that was a struggle, because I constantly got emails saying that the units not being charged … I would go on home visits and find half the units on the table, it’s plugged in to the wall but not switched on, so it was about teaching them over a course of like a month that this thing needed to be charged and how to charge it” OM “[Charging] wasn’t a problem for me, you had the home thing and the charger at home. I ensured that every day an hour before I got home from work, I would tell my sister to put it on charge, by the time I got home, I had a routine, that was the best way.” GPS tag wearer 21 Use of GPS data in supervision When asked how they had used location data in supervision the most popular response from OMs was to check on the movements of the tag-wearer, primarily for general monitoring (to see where the tag-wearer had been), or to ensure they had not been in exclusion zones (one OM mentioned the use of heat maps to check whether the tag-wearer had been in areas where crime had been committed). Other uses of tag data mentioned by OMs were to check on the tag-wearer’s contacts with associates (in relation to suspected engagement in County Lines, or to monitor the tag-wearer’s return to areas where what were termed ‘risky’ groups/peers resided). Several OMs mentioned that they intended the tag to act as a deterrent (particularly around the establishment of exclusion zones - to keep the tag-wearer away from an ex-partner for example). One OM remarked that, when this deterrent had failed and the tag-wearer had gone into his exclusion zone, this had worried her precisely because he knew that the OM would be able to see his movements. OMs also used the tag to manage risk, either the risk posed by the tag-wearer (by ‘set(ting) limits’, ‘provide dissuader from risky decision making’) or the risk to the tag-wearer (one OM had extended the length of the tag by a couple of weeks on this basis). A few OMs suggested that they had used location data to challenge or discuss the tag-wearer’s behaviour, one suggesting that the data had ‘opened up’ that conversation, although others stressed that it had not been appropriate to do this (because they could clearly see where the tag wearer had been) or not feasible (in cases where the tag-wearer had been non-compliant or recalled). Exclusion zones had been set up by half the OMs interviewed, at a variety of levels (boroughs, an ex-partner’s address, a hospital, a housing estate and the Notting Hill Carnival area). One OM indicated that while she had set up an exclusion zone, she had let the tag-wearer pick an area of his own to emphasise the benefits of an area for his own safety. OMs who had not established exclusion zones stressed that this was because they did not feel it was appropriate (because the tag was ‘purely for monitoring’, and because of the ‘mobile nature’ of the offence). OMs identified the need for a degree of negotiation or flexibility in the enforcement of exclusion zones. One tag-wearer, who had been moved outside London but had to travel into London, always came into a terminus station within his exclusion zone; ‘having it on record to say there was a breach and that there is no other way it forced me to use my own initiative to contact the police and say that we need to have a sort of negotiation here because I am not prepared to prosecute someone for passing through a train station that is unavoidable’. OMs were also asked if having the tag had made any difference to the nature of supervision. While a couple said not (based on the early recall of the tag-wearer, which meant that it had not been possible to use GPS data in supervision) most felt that it had, for a variety of reasons (the ability to check the tag-wearer’s movements, there being more contact between the OM and tag- wearer, the tag providing data against which the tag-wearer’s progress could be measured, and the tag opening up discussions about the tag-wearer’s behaviour). Notably, even one OM who suggested that the tag had ‘not necessarily’ changed the nature of supervision then went on to add that they were ‘receiving information about where [the tag-wearer] was going and certain addresses, which was useful in seeing if there was going to be a conflict’. 22 Views on the impact of the tag on the relationship between the OM and the tag-wearer varied. While a few OMs felt that it served to build trust – ‘what the tag enabled me to do, and what he knew the tag enabled me to do, was monitor exactly his route that he took in … that was really, really helpful because also it builds trust, as well as be there to catch any return to a lifestyle’ – other OMs felt that the imposition of the tag initially soured the relationship, which took some time to overcome. The 3 tag-wearers interviewed were also asked about the impact of the tag on their supervision. One felt it was clearly beneficial and had led to improved relations with his OM: ‘There was a lot of times when the tag wasn’t picking up the charge and so there were warning things, so I was in regular contact with him to let him know that the thing wasn’t charging, but the contact was on both sides and I would contact him so that I didn’t breach’. One of the other tag-wearers mentioned that he and his OM would talk about his movements at the end of supervision: ‘the GPS conversation would probably come at the end of the meeting, and during the start we would probably talk about other stuff…. she would ask what I was doing in [redacted] for so long, but I was with someone … parties and stuff she would ask me’. The OMs provided details of breaches/recalls (or threats of the same) that had occurred during their supervision. The range of responses they provided reflects the pattern found in the performance section of this report. Two OMs said that they had had no issues in relation to the tag wearers, another that they had received charging emails and threatened a breach but the tag- wearer had completed, and a fourth that there had been some minor and one major breach of the tag-wearer’s curfew conditions, which had been met by ‘informal challenge’ but that the tag- wearer had always attended and completed successfully. The remaining tag-wearers had all been recalled. OMs were asked if they had extended the tag length or had considered doing so. One of the OMs had extended the tag for an additional month because of a ‘little incident’ between her tag- wearer and a ‘couple of boys’ but stated that the he had been fine with the extension and it had passed off successfully. Another tag-wearer had had his tag period extended for an additional month because it emerged that, due to faulty equipment, he had been unmonitored for that period. In that case, the OM was clearly concerned about the ethicality of the extension. Three other OMs indicated that: they would have asked for longer on the tag, with hindsight; they would have extended if they could (the tag-wearer had been recalled); and they would have done so if it had been necessary. Influence of the GPS tag Staff generally felt that the tag had had an impact on the tag-wearer’s behaviour whilst on the tag, because it had acted as a deterrent, or due to then tag-wearer having an increased awareness of being monitored. As an example, one OM cited a case where the tag-wearer had not ventured outside his inclusion zone (the M25) whilst on the tag, but in the period shortly after removal had come to police attention in a number of counties outside the M25 and had been subsequently convicted for an offence in one of those areas. When asked if they thought that the tag was likely to have an impact on the tag-wearer’s longer- term behaviour, OMs were more ambivalent. Some felt it was impossible to say, others that this 23 was not the case - ‘definitely not the thinking thing’ (pointing to the tag-wearer’s subsequent conviction for the same offence). Other OMs pointed to improvements while on the tag, but which had ceased once it had been removed. Another OM felt there had been changes in the tag-wearer’s behaviour but felt that this had been due to other factors (fatherhood) rather than the tag. Only one OM suggested that there was a continuation: ‘they just continue like they’ve got it on’. It was also observed that the tag was of relatively short duration. “Hard to say because he was on it for only five or six weeks. I don’t think it impacted on long- term positive change but I think that what it did do is give him a five- to six-week period where he didn’t have to even consider going back to a risky area, where he could actually take time to focus on doing what he wanted to do … it gave him a period of time to recognize that he’s on licence, he went to prison for a reason, he can’t do back to an area where his peers might pressure him, so he might as well do something positive, which was really good for him. Since he’s come off the tag, his motivation has nosedived. He hasn’t really done much at all. He’s been rearrested. He’s almost regressed. But I acknowledge it was a really short case, really short period of time.” GPS tag wearer 24 4. Discussion The GPS knife crime pilot forms an important part of the London Knife Crime strategy, developed in the face of increased levels of knife crime in London and heightened public concern about the problem. Between the beginning of the GPS knife-crime pilot in February 2019 and the 6th of January 2020 a total of 102 GPS tags were imposed. 46 cases have completed; 27 cases ended due to recall and 19 without recall, giving an overall rate of completion without recall of 39%. Of the 27 cases ended due to recall, however, 16 cases were recalled for reasons unrelated to the GPS tag, 6 cases were recalled due to failure to comply with licence conditions directly related to the GPS tag, and 5 cases were recalled due to breach of an exclusion zone being monitored through the GPS tag. Taking these reasons for recall into account, the effective compliance rate, that is the proportion of cases completed without recall for failure to comply with the GPS tag as a licence condition, is 87% (n=40). In terms of implementing innovation, this is a relatively high level of compliance and in all likelihood should be seen as a positive – although, in itself it is not a measure of ‘impact’. However, the issue of determining impact and identifying what success might look like in terms of case outcomes is problematic. For example, a recall, while it may decrease the completion rate overall, might be judged a success in demonstrating the improved (and swifter) enforcement of licence conditions. Indeed, this was one of the benefits of the tag that was identified by OMs – the ability to tell, faster and with greater certainty, when tag-wearers had breached exclusion zones for example. It is notable that in only 11 of the 46 completed cases to date recall was due to failure to comply with the tag as a licence condition and those recalled due to a breach of an exclusion zone monitored by the tag. The qualitative research undertaken during this pilot indicates that OMs have a very positive view of the GPS tag. There was widespread support for the GPS tag as a means of monitoring the movements of tag-wearers, providing information that could be used in supervision, and the improved ability to manage risk (both to victims and tag-wearers) precisely because of this increased certainty about tag-wearers’ movements. Overall, OMs felt that availability of GPS data had an impact on the nature of supervision between them and the tag-wearer, and on tag- wearers’ behaviour whilst on the tag. A different type of discussion with the tag-wearer was feasible, because of the type of data additionally available (i.e. it was difficult for the tag-wearer to challenge the location data). Given the value seen in the data generated by the GPS tags, there was qualified support from OMs for direct access to the IT system provided by the tag contractor (Buddi), although it was clear that OMs’ understanding of the GPS data and its potential uses varied. Given this, OMs would likely benefit from ongoing training and support to promote the value of the data and enhance knowledge and confidence around the ways and formats in which data can be requested/accessed. OMs suggested that wearing the tag may have a short-term deterrent effect although were not yet convinced the tag would bring about longer-term change in offending behaviour. In the views of OMs, any change for the tag-wearer was largely dependent upon the tag-wearer’s motivation and their willingness to change – to take advantage of the ‘window’ that the tag provides. 25 There was also support from OMs for the extension of the use of the GPS tag for other offence types. However, this potential extension needs to be considered in the light of concerns around net-widening and the ethical integrity of utilising GPS tags, with the increased scrutiny (and likely detection) they provide, for some offences but not others – a point already made by one of the service users interviewed. The findings from tag-wearers were more ambivalent. Although there appeared to be support for the tag from tag-wearers at the point it was fitted, it was not possible in this report to examine how this varies over the period the tag is worn and the level of support at the point of removal. However, there does appear to be a recognition on the part of some tag-wearers of the benefits the tag provides in providing a means of avoiding risky situations/individuals, if they have the inclination to do so. This echoes the perception of OMs interviewed of the importance of motivation among tag-wearer. The tag-wearers also demonstrated a wide range of understanding to the tag, and more could be done to provide consistent information prior to and during the order. The evaluation has also identified some issues around the administration of the tag, particularly in the early days of its operation. There were initial difficulties in arranging tags to be fitted in prison on the day of release. Issues, too, relating to the charging of the tag, and the ability of OMs to establish whether the tag was faulty/not charging in the face of claims from the tag- wearer. It would also appear that there is scope to improve communication with offenders about the pilot to ensure that they are aware they are going to be tagged. Concerns, too, were expressed about the utility of some of the output routinely provided to OMs (which has been amended in the light of these concerns). Nonetheless, the OMs appeared to be very content overall with the service received from Buddi. Findings from the police crime mapping process are inconclusive at this point. Despite the large number of crimes uploaded and matches assessed, there have been relatively few “significant matches” to date and the GPS information shared with local BCUs has not led to any additional convictions so far. As discussed previously in E&I’s 2018 Interim Report on the use of GPS tagging for an IOM cohort, there are several possible explanations for the seemingly low number of genuinely ‘significant matches’ generated by the crime mapping process. The GPS-tagged population deemed appropriate for crime mapping (i.e., those with an OGRS score higher than 50) comprises a small number of individuals, who are only likely to generate a correspondingly small number of truly significant matches after the analysis process has been completed. Another complicating factor is the type of offences for which matches have been generated (theft and burglary for example). These offences, by their nature, provide the greatest latitude in terms of the reported time that they took place (and thus, the greatest risk of placing the offender at the scene without proving clear evidence that they were actually involved in the offence), so the greatest opportunity for what might be called ‘false positives’ in terms of matches (MOPAC 2018). In addition, many acquisitive offences are reported by the public over the phone or online, and as a result the crime reports for matched incidents often lack detail and may have been be closed for no further action. Another possible explanation for the low number of significant matches is that offenders subject to GPS monitoring are committing fewer offences whilst on the tag, meaning that the lack of matches reflects the deterrent effect 26 of the tag (consistent with OM feedback) rather than a failure of the crime mapping process to correctly match tag wearers with the crimes that they have committed. The pilot is focused on a cohort of knife crime offenders and violence/robbery rather than offenders with a history of burglary, and there have been very few matches to robbery offences (3%, n=3, of the 96 significant matches during the interim reporting period). In the light of the issues identified above, MOPAC is currently working closely with the MPS to consider changes that can be made to enhance effectiveness of the crime mapping process. The MPS is refining the process for identifying and triaging ‘hits’ to ensure that they focus on the cases where there appears to be the best potential for a high-quality match and more serious offences (excluding property crimes and focusing on contact crimes, e.g. violence, robberies and knife-related crime). Notably, the use of crime mapping in this pilot is being framed in terms of ‘crime reassurance’ (confirming that tag-wearers have not/or are unlikely to have committed recorded offences) rather than the crime detection impetus behind the use of crime mapping in MOPAC’s pilot using GPS tags for an IOM cohort. In addition, the process of managing the matches/offenders that emerge from the crime mapping process has been standardised across the MPS and now sits with the police IOM teams. How this management works will be examined in the final evaluation report. Our final evaluation report, to be completed within six months of the pilot’s conclusion, will continue to collect data on the number of tags fitted and levels of compliance by wearers, as well as feedback from tag wearers, OMs. Additional performance and qualitative data on crime mapping will be collected, including throughput of crimes uploaded for mapping to outcomes of matches and feedback from police staff. If the sample size (i.e., the number of tags fitted) permits, it will also examine the impact of the tag on tag-wearers’ offending behaviour against an appropriate control group. 27 5. References MOPAC (2018a) ‘GPS Tagging First Year Interim Report’ https://www.london.gov.uk/sites/default/files/mopac_gps_tagging_second_interim_report_2 018.pdf MOPAC (2018b) ‘Review of the Metropolitan Police Service Gangs Matrix’ https://www.london.gov.uk/sites/default/files/gangs_matrix_review_-_final.pdf MOPAC (2018c) ‘Alcohol Abstinence Monitoring Requirement: A review of process and performance from Year 2 https://www.london.gov.uk/sites/default/files/aamr_final_process_performance_y2_report_fi nal.pdf Office of National Statistics (2018) ‘Crime in England and Wales: year ending December 2017’ https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/crimeine nglandandwales/yearendingdecember2017 Pepper, M., and Dawson, P. (2016) ‘Alcohol Abstinence Monitoring Requirement A process review of the proof of concept pilot’. MOPAC Evidence and Insight Unit Prison Reform Trust, ‘Licence conditions and recall: Indeterminate sentences’ http://www.prisonreformtrust.org.uk/ForPrisonersFamilies/PrisonerInformationPages/Licence conditionsandrecall/Licenceconditionsandrecallindeterminatesentences (accessed 20/02/20). 1 GPS Tagging – Community Sentence Pilot Final Evaluation Report January 2020 Valerie Forrester & Tim Read MOPAC Evidence and Insight 2 Contents Executive Summary ................................................................................................................... 3 1. Introduction ........................................................................................................................... 6 Background to the pilot .......................................................................................................... 6 Location data available on pilot ............................................................................................. 7 Extension of GPS tag to knife crime offenders……………………………….……………………….…………7 GPS tagging evaluation ........................................................................................................... 8 2. Methodology ....................................................................................................................... 10 3. Results .................................................................................................................................. 11 Performance ......................................................................................................................... 11 Practitioner feedback ........................................................................................................... 16 Tag wearer feedback…….…………………………………………………………………………………………………19 Police crime mapping ........................................................................................................... 24 4. Discussion ............................................................................................................................ 25 5. References…...………………………………………………………………………………………………………………..26 3 Executive Summary In 2016, the Ministry of Justice (MOJ) agreed to work with the Mayor’s Office for Policing And Crime (MOPAC) to test the use of mandatory GPS tagging for prolific offenders through the Persistent Offender Programme (POP). The main aims of the MOPAC pilot were to test whether GPS tagging - increased compliance with the requirements of a Community Order (CO) or Suspended Sentence Order (SSO), offered sentencers an alternative to custody, and supported crime detection and/or the apprehension or prosecution of offenders by providing information on an offender’s location at a specific date and time. Consequently, GPS monitoring under the electronic monitoring requirement of community sentences was made available in the Magistrates’ and Crown Courts of the North and North East London local justice areas between March 2017-March 2019. Two dedicated pre- sentence report writers from the National Probation Service based at Thames and Highbury Corner Magistrates’ Courts had the option to recommend GPS tags as a requirement of a CO or SSO. Under the POP pilot, probation Offender Managers (OMs) were able to access and monitor location data for their cases. Where the tag wearer was the subject of an SSO and an OGRS score of 50+1, the police were also able to undertake crime mapping, matching location data with crime data where it was relevant and justified for the detection of crime. In addition, between October 2018 and March 2019, the use of GPS tagging was extended to cover offences involving the use or possession of a knife in the 8 boroughs served by Thames and Highbury Corner Magistrates Courts. Overall, there have been a total of 135 GPS tags imposed across the pilot timeframe. This splits across the POP cohort (n=117) and knife crime cohort (n=18). Taking the POP cohort first: • Over the pilot period 117 GPS tags were imposed by the judiciary (87 at Highbury Corner Magistrates’, 29 at Thames Magistrates’, and 1 at Snaresbrook Crown Court): an average of 1.1 tags imposed per week. • Of the 117 GPS tags imposed, 70 were imposed as a requirement of a Community Order and 47 as part of a Suspended Sentence Order. Alongside the electronic monitoring requirement, most orders imposed also included a Rehabilitation Activity Requirement (RAR, 108 cases, 92%). Most GPS tags imposed included a requirement of access to 24-hour whereabouts monitoring (94%, n=110),2 which meant that the tag wearer’s location data from any time or location was able to be accessed by the 1 The Offender Group Reconviction Scale (OGRS) is a risk assessment tool used to estimate likelihood of re-offending using static factors such as age, gender and criminal history. It gives a score, which shows the likelihood of someone re-offending over a 1 year or 2-year period, expressed as a percentage or on a range from 0 to 1. A lower score means a lower likelihood of re-offending. 2 Now also known as “trail monitoring”. 4 Offender Manager where this was relevant to the management of the case. Seven tags were imposed in conjunction with a restrictive requirement (e.g., GPS being used to monitor compliance with an exclusion zone). • The success rate (i.e., the % of GPS tags recommended by the PSR writers for the POP cohort which were subsequently imposed by the judiciary) in the first year of the pilot was 70%. • Of the 117 tags imposed over the POP pilot’s operation, 61 were completed successfully (‘compliant’ tag wearers), while 56 were unsuccessful (‘non-compliant’ tag wearers), either revoked for a failure to comply with the GPS requirement and/or imprisonment following further offending or were in the process of being breached: a completion rate of 52%. By comparison, the GPS knife-crime cohort (n=18) was generally more compliant, with 14 of the 18 tags completed satisfactorily (a rate of 78%). For the four cases that were breached, ‘failure to attend’ was cited in 3 cases, as well as failure to charge, and failure to comply with electronic monitoring. • The majority of tag wearers in the POP cohort were male (84%, n=98) and the average age was just over 33 years. Three-quarters of tag wearers were aged under 40. Where the ethnicity of the tag wearers was known (2 cases were blank or refused), 59% (n=39) were White, 21% (n=24) Black, 9% (n=10) Asian, 10% (n=11) Mixed, and 2% (n=2) ‘Other ethnic group’. Offence details were available in 112 of the POP cases. Of these, 39% were theft (n=44, 4 of which were motoring related), 18% burglary (n=20), 11% offences of violence (including DV) (n= 12), 9% public order/racially aggravated public order (n= 10, 5 of which were racially aggravated), 6% criminal damage (n=7), 5% drug possession supply (n=6), 4% breach (n=5), 4% possession of offensive weapon (n=4), 2% driving offences/motoring (n=2), with one case of harassment and one case involving a dangerous dog. For the Knife crime cohort: • There were 18 tags for knife crime offences imposed during the pilot period (5 from Highbury Corner, 13 from Thames Magistrates Courts). Twelve individuals had the tag imposed as part of an SSO, while the remainder were COs – almost exactly the reverse of the pattern for the POP cohort. Alongside the electronic monitoring requirement, most orders imposed included a Rehabilitation Activity Requirement (17 of the 18 cases). The average length of the GPS requirement was 3.8 months, slightly shorter than the average for the POP (4 months). • For the group tagged for knife crime offences, 16/18 were male and the mean age was just under 29 years (with a range between 18 and 61, median of 23.5 and a joint mode 5 of 19/24). In terms of ethnicity, 6 of those tagged were Black, 4 were White, 3 were Asian, and one was Mixed (the ethnicity of four was unknown). Across all individuals, there were few measurable differences between the compliant and non-compliant tag wearers; non-compliant wearers had significantly lower levels of motivation. For both groups, the most prevalent risk factor displayed was drug use, followed by mental health. Overall, findings from the research show that the GPS tagging pilot was implemented well (something never to be taken from granted when implementing innovation) and has been well received by practitioners and tag wearers alike. Offender Managers were positive and willing to use the tags. Tag wearers were confident that they would successfully complete the monitoring period and knew what they needed to do to comply. Furthermore, tag wearers generally thought that the tag would have a positive impact on their life. Although, as in E&I’s earlier interim evaluation report, concerns were expressed (by tag wearers and practitioners) about the difficulties caused for this cohort by the requirements to keep the tag charged. Consistent with E&I’s previous evaluation report, early non-compliance by tag wearers suggests there are issues around the identification of suitable individuals to be tagged that need to be addressed. Other challenges that have emerged are the repercussions arising from the restrictions placed on PSR writers making recommendations for SSOs in April 2018, which resulted in a reduction of SSOs with tagging requirements, and a reduction in eligible cases for the crime mapping strand of the pilot (which can only be undertaken on SSO tag wearers). Question marks about the effectiveness of the crime mapping strand in its current form, raised by staff in the previous interim report, remain. Nevertheless, the implementation, overall completion rate for the scheme, particularly the high completion rate for the knife crime group (albeit for a small population), and the positive views expressed by practitioners and tag wearers about their experience with the GPS tag, continue to provide grounds for optimism. 6 1. Introduction Background to the pilot A brief description of the MOPAC POP GPS tagging project is provided below, a more detailed account can be found in E&I’s First Year Interim Report3, published in July 2018. In early 2016, the Ministry of Justice (MOJ) announced its intention to pilot the use of GPS tags. The MOJ agreed to work with the Mayor’s Office for Policing And Crime (MOPAC) to test the use of mandatory GPS tagging for prolific offenders through the Persistent Offender Programme (POP) pilot, starting in March 2017. Although GPS tags had been used on a voluntary basis in several small-scale pilots across London, this was the first-time that mandatory use had been piloted with such a cohort. Initially running for 12 months, MOPAC successfully applied to the MOJ to have the statutory instrument for this pilot extended for an additional 12 months; meaning that new cases could be tagged until March 2019. Under the POP pilot, offender whereabouts and location data could be accessed by the allocated probation Offender Manager (OM) as part of an Electronic Monitoring requirement imposed as part of a Community Order (CO) or Suspended Sentence Order (SSO). New tools available to probation OMs included bespoke email notifications highlighting key locations visited by the offender in real time, heat maps and “Top 5 locations of the week” reports, so that patterns of behaviour associated with risk could be uncovered, explored and challenged. Where the offender was the subject of an SSO, the police were able to undertake crime mapping, matching location data with crime data where it was relevant and justified for the detection of crime. There was also the capacity for the police to make external requests for location data from the GPS monitoring provider (Buddi) in relation to specific crime incidents. During the pilot there were 20 External Agency Requests (EAR) made. GPS monitoring as described above was available in Thames Magistrates’ and Highbury Corner Magistrates’ Courts and the local Crown Courts between 20 March 2017 and the end of March 2019. Two dedicated pre-sentence report (PSR) writers from the National Probation Service (NPS) based at Thames and Highbury Corner Magistrates Courts were given the option to recommend GPS tags as a requirement of a CO or SSO.4 It was envisaged that the Electronic Monitoring requirement would be used in conjunction with other rehabilitative measures (e.g. Rehabilitation Activity Requirements). The main aims of the MOPAC pilot were to test whether GPS tagging: 3 https://www.london.gov.uk/sites/default/files/mopac_gps_tagging_second_interim_report_2018.pdf 4 After April 2018, PSR writers were not able to recommend SSOs. 7 1. Increased compliance with the requirements of a Community Order or Suspended Sentence Order; 2. Offered sentencers an alternative to custody; and 3. Supported crime detection and/or the apprehension or prosecution of offenders by providing information on an offender’s location at a specific date and time. Location data available on pilot The type of location data available to the various agencies involved in the pilot differed depending on the type of order made. Community Orders with Exclusion Zones or residence requirements • Probation would only get alerted if the offender breached the exclusion zone or residence requirement. • The police had no access to location data (apart from EAR). Community Orders with 24/7 whereabouts monitoring (also known as trail monitoring) • All the location data captured could be shared with the probation OMs. Alerts could be set against inclusion zones, points of interest (specific addresses) or interest zones. Requests for other data could be made via a secure email sent directly to Buddi - heat maps, top location reports etc. • Other agencies could be involved in helping to suggest monitoring locations and support compliance i.e., drug workers, police intelligence analysts, criminal behaviour order (CBO) exclusion zones. • The police had no access to location data (apart from EARs). Additional functions available for Suspended Sentence Orders • The police would carry out automatic crime mapping and may be alerted to location data that matches a crime, so they could investigate further. They could also make EARs. Extension of GPS tag to knife-crime offenders GPS tagging was extended to knife crime offences in the east London boroughs served by Thames Magistrates’ Court from the 15th October 2018, and to the north London boroughs served by Highbury Corner Magistrates’ Court from the 1st November 2018. The extension of GPS to knife crime offenders on community sentences was viewed as a test of GPS as a tool to manage those who posed a risk of serious harm to the public and to see whether GPS could be used to challenge the lifestyles of offenders who habitually carried knives. These goals 8 were seen as building on the work already done with Gripping the Offender/Persistent Offender Programme offenders5, learning early lessons about the knife crime cohort (such as their levels of compliance) with the future ambition (subsequently realised) of using GPS for knife crime on licence. To be eligible the offender had to be: • Appearing for sentence for an offence involving the use or possession of a knife; • Aged 18+; • Living in the 8 pilot boroughs; • Be appearing for sentence in the relevant LJAs; and • Have some type of accommodation – to charge the tag. In addition, it was clearly stated that this option was available at the court’s discretion for first time knife possession offenders, so as not to contradict the sentencing guidelines which directed a minimum of 6 months in custody for a second knife crime offence. N.B For the purposes of this evaluation report, the analysis of the data relating to the GPS knife crime offenders has been kept separate to that relating to the POP cohort, as it was felt that the characteristics of the two groups were too dissimilar to render the combination of the data meaningful. GPS tagging evaluation The Evidence and Insight team at MOPAC were asked to evaluate the GPS pilot, capturing learning around the implementation and design of the pilot, gaining insight into the experiences of practitioners and offenders, and measuring success. The evaluation of the GPS tagging pilot forms part of the wider POP evaluation. The overarching aims of the evaluation are to examine: • Performance - monitor the key performance indicators, including numbers receiving the tag, types of orders, compliance rates, violations, breaches, order completions. This data is vital, especially within a new innovation. • Process - understand the experiences of practitioners and offenders involved in the pilot through surveys, interviews and focus groups. 5 Details of the GtO (Gripping the Offender, later renamed the Persistent Offender (POP)) programme can be found in MOPAC’s 2018 interim report https://www.london.gov.uk/sites/default/files/mopac_persistent_offender_programme_interim_report_2018.pdf 9 • Indicative Impact - a robust impact evaluation was not possible due to the relatively small sample size. The evaluation has captured indicators of success (e.g., completion rates) and uses qualitative methods to gain learning about the success of the pilot. This report shares findings and learning from two years of the GPS tagging pilot. It consists of performance and process aspects, including feedback from offender managers and tag wearers. 10 2. Methodology Performance Performance data was gathered from a range of sources, mainly the tagging provider Buddi and MOPAC delivery staff. This data included information about the tag wearers and the types of orders that they were on, and their compliance with the tag. Offender Manager Surveys A survey was sent to offender managers (OMs) who were responsible for managing GPS cases in late February 2018 and again in late May 2019. The survey asked OMs about their experiences working with GPS cases, their perceptions of the tag, and how they have used the tag to manage their cases. 31 completed surveys were collected – nine following the February 2018 round and a further 22 following the May 2019 round (a response rate of 33%). Tag Wearer Surveys Tag wearers were asked to complete a short survey at the point the tag was fitted, and again when it was removed. The surveys aimed to collect their anticipated and actual perception of the impact of wearing the tag on their offending behaviour, their relations with friends and family, and their lifestyle more generally. Data were available from 42 individuals where a tag had been fitted, and 24 cases where the tag had been removed. It should be noted that this is a small data set and is unlikely to be representative of the GPS tagging cohort overall because it comprises compliant offenders (both at the point of fitting and removal). Surveys were administered by the tag fitter (staff members from Buddi). 11 3. Results Performance Overall, there have been a total of 135 GPS tags imposed across the pilot timeframe. This splits across the POP cohort (n=117) and knife crime cohort (n=18). In the following section analysis for the POP and knife crime cohorts is largely presented separately, because of the differences between the two groups, and to prevent confusion. POP cohort: GPS tags imposed Between the beginning of the pilot (March 2017) and its completion (March 2019), a total of 117 GPS tags were imposed by the judiciary (87 at Highbury Corner Magistrates’, 29 at Thames Magistrates’, and 1 at Snaresbrook Crown Court): an average of 1.1 tags imposed per week. In terms of the number of tags imposed across the two years of the pilot, more tags were imposed in the first year of operation (73, 62%) than during the second (44, 38%). Possible reasons for the decline during the second year are discussed below. The success rate (i.e., the % of GPS tags recommended by the PSR writers which were subsequently imposed by the judiciary) during the first year of operation of the pilot was 70%. Of the 117 GPS tags imposed, 70 were imposed as a requirement of a Community Order and 47 as part of a Suspended Sentence Order. Again, there was a markedly different pattern between the first and second years of operation. In the first-year, tags as a condition of SSOs made up 47% of all orders, whereas in year 2 they made up 30%6. Alongside the Electronic Monitoring requirement, most orders imposed included a Rehabilitation Activity Requirement (108 cases, 92%), while 28 had a Drug Rehabilitation Requirement and 8 included an Alcohol Treatment Requirement7. The average length of the tag order was just over three months (ranging from between 2 weeks to 6 months – the most popular tag lengths were 3 months (55 of the 116 cases where the tag length was known, and 6 months – 40 cases)). Most GPS tags imposed included a requirement of access to 24-hour whereabouts monitoring (94%, n=110), which meant that the tag wearer’s location data from any time or location could be accessed by the Offender Manager where this was relevant to the management of the case. Seven tags were imposed with a restrictive requirement (i.e., the data were only available in relation to a requirement not to go to a particular location). For the GPS tags 6 The explanation for the decline in the numbers of SSOs in the second year of the pilot is because, from April 2018 PSR writers were no longer able to make recommendations for SSOs 7 There were17 other conditions attached to tags. Five were for unpaid work of between 120 and 40 hours, four were curfews (of between (7 and 18 weeks), 4 were to attend a Thinking Skills programme, 2 were to attend a Building Better Relationships programme while 3 were for other accredited courses. 12 imposed, 61 Interest or Exclusion Zones had been set up by OMs. OMs receive an alert any time the tag wearer enters these zones. Offence details were available in 112 cases. Of these, 39% were theft (n=44, 4 of which were motoring related), 18% burglary (n=20), 11% offences of violence (including DV) (n= 12), 9% public order/racially aggravated public order (n= 10, 5 of which were racially aggravated), 6% criminal damage (n=7), 5% drug possession supply (n=6), 4% breach (n= 5), 4% possession of offensive weapon (n=4), 2% driving offences/motoring (n=2), with one case of harassment and one case of an offence involving a dangerous dog. Characteristics of tag wearers: POP cohort The majority of tag wearers were male (84%, n=98) and the average age of the tag wearers was just over 33 years. Three-quarters of tag wearers were aged under 40 years: 21% (n=25) were 18-24 years and 55% (n=63) were 25-39 years. Twenty-four percent (n=28) of tag wearers were aged 40-54 years, and only 1 was aged 55+. Where the ethnicity of the tag wearers was known (2 cases were blank or refused), 59% (n=39) were White, 21% (n=24) Black, 9% (n=10) Asian, 10% (n=11) Mixed, and 2% (n=2) ‘Other ethnic group’. In this respect, the characteristics of those receiving tags broadly reflected the characteristics of the POP cohort overall.8 During the first year of GPS tagging when the PSR writing service was available, there were 191 individuals in the POP cohort who received the service whose ethnicity was known: 45% (n=86) were White and 55% (n=105) Black, Asian or Minority Ethnic (BAME). Of these, 23% (n=20) of White service users and 25% (n=26) of BAME service users were recommended GPS tagging, suggesting the report writers proposed GPS tags on similar proportions of White and BAME individuals. Knife crime cohort: GPS tags imposed Between mid-October 2018 (in Thames Magistrates Court) and the beginning of November (Highbury Court Magistrates Court) and the end of the pilot (March 2019) there were 18 tags imposed on service users for knife crime offences (5 from Highbury Corner, 13 from Thames Magistrates Courts). Seventeen of the tag-wearers were being supervised by the London Community Rehabilitation Company (CRC), one from the NPS. Twelve had the tag imposed as part of an SSO, while the remainder were for COs – practically the reverse of the pattern for the POP cohort. 8 POP refresh data from January-February 2018 consisted of 91% Males and an average age of 30 years. A pan-London dataset compiled from CRC data from January 2018 and NPS data from October 2017 showed the following ethnic composition; Asian 14%, Black 26%, Mixed 8%, Other ethnic 2% and white 50%. 13 Alongside the Electronic Monitoring requirement, most orders imposed also included a Rehabilitation Activity Requirement (17 of the 18 cases), while 4 had a requirement for unpaid work, 2 had a Drug Rehabilitation Requirement, 1 included an Alcohol Treatment Requirement, 1 a curfew condition, and 1 attendance at an Attendance Centre. The average length of the tag order was 3.8 months, longer than the average for the POP cohort (the length of tags ranging from 6 weeks to 6 months, the most popular tag length being 4 months (6 cases)). For the GPS tags imposed, 11 Interest or Exclusion Zones had been set up by OMs. Characteristics of Knife Crime tag wearers Of the 18 individuals tagged for knife crime offences, 16 were male and 2 female. The average age of the tag wearers was 28.8 years, but this figure was skewed by the fact that one of those tagged was over 60 (the median age was 23.5 and a joint mode of 19/24). In terms of ethnicity, 6 of those tagged were Black, 4 were White, 3 were Asian, and one was Mixed (data were unavailable for 4 service users). Compliance with the GPS tag POP cohort Of the 117 tags imposed over the pilot’s operation, 61 were completed successfully (‘compliant’ tag wearers), while 56 were unsuccessful (‘non-compliant’ tag wearers), either revoked for a failure to comply with the GPS requirement and/or imprisonment following further offending or were in the process of being breached: a completion rate of 52%.9 Characteristics of compliant and non-compliant tag wearers are presented in Table 1 below. In terms of the demographic characteristics, both groups predominantly consisted of males. There was no significant difference between the compliant and non-compliant groups in terms of age and ethnicity. In terms of disposal, the non-compliant group had a slightly higher proportion of tag wearers on a CO (35/56 compared to 35/61); again, this difference was not statistically significant. Information was collected on the risk factors the two groups displayed (see Table 1). Specifically, individuals were measured on drug use, alcohol use, housing, physical health, mental health, and motivation levels. The non-compliant group had an average of 2.1 risk factors, out of the potential maximum of six10. The compliant group had a slightly lower average number of risk factors (1.8), but this difference was not statistically significant. 9 This is very similar to the rate of 51% (combining technical and non-technical parole violations) found in a study testing the effect of GPS tagging with a sample of high-risk gang offenders in California (Gies et al., 2013). 10 The 6 risk factors identified by probation staff were; drug use, alcohol use, housing, physical health, mental health and low motivation. 14 For both groups, the most prevalent risk factor displayed was drug use, followed by mental health. Comparing individual risk factors across the two groups, there was only one statistically significant difference: the non-compliant group displayed significantly lower motivation than the compliant group.11 The number of tag wearers assessed as having housing issues was low for both groups, but most likely reflects the fact that (relatively) stable accommodation is a prerequisite for the recommendation/ adoption of a GPS tag. Table 1. POP cohort. Characteristics of compliant and non-compliant tag wearers Non- compliant (n=56) Compliant (n=61) Gender Male 47 51 Female 9 10 Age (mean) 33.9 32.5 Ethnicity White 34 34 BAME 21 26 Order type CO 35 35 SSO 21 26 Risk factors Drug use 37 42 Alcohol use 10 18 Housing 7 6 Physical health 4 7 Mental health 25 30 Low motivation*** 31 7 Information about the reason for breach was available in 38 of the 56 non-compliant cases (it was missing in 3 cases, and n/a in 15 cases). The most popular reason for breach was ‘failure to attend/attendance’ mentioned in 28 cases, ‘failure to charge/charging’ was mentioned 19 times, and ‘non-compliance’/’non-engagement’ 11 times. Looking at the number of appointments attended with probation by individuals in the non- compliant group, these tag wearers attended an average of 4.3 appointments overall (ranging from 0 to 41 appointments). Interestingly, the subsequent pattern of attendance for those who did, and who did not, attend their first appointment with probation was markedly different. Individuals who attended their first appointment (n=29) attended an average of 6.5 appointments overall. Conversely, for the 27 individuals who failed to attend their first appointment, the average number of appointments attended overall was 1.9 (the difference 11 Fisher exact test statistic value is 0.00001, significant at p < .05 15 between the number of appointments attended by non-compliant tag wearers who either did or did not attend their first appointment was statistically significant). What this pattern appears to suggest is that there is a group of tag wearers who are largely non-compliant right from the outset of the order, compared to other tag wearers who, while ultimately being breached, engage with probation over a longer period of time. Support for this conclusion is provided by analysis undertaken by probation staff who looked at the case records for the non-compliant group and, based on their contents (levels of attendance/charging/engagement etc) assessed the extent to which the service user had tried to comply with the tag. Of the 55 cases where data were available, 36 were deemed not to have attempted to comply, 4 to have partly attempted to comply, and 15 to have tried to comply. The fact that a group of offenders was (with hindsight) not motivated to engage right from the start of their sentence, despite having superficially indicated some readiness to engage during pre-sentence assessment (hence GPS being proposed as available option), would seem to indicate the problems in reliably assessing underlying motivation at the pre- sentence stage, especially for long term persistent offenders. GPS knife crime cohort Bearing in mind the small numbers, the knife crime cohort was more compliant than the POP cohort, with 14 of the 18 tags being completed satisfactorily (a completion rate of 78%). For the four cases that were breached, failure to attend was cited in 3 cases, as well as failure to charge, and failure to comply with electronic monitoring. Again, information was collected on the risk factors the two groups displayed (see Table 2 on the next page). The non-compliant group had an average of 1.5 risk factors, out of the potential maximum of six. The compliant group had a very slightly lower average number of risk factors (1.4). 16 Table 2. Knife crime cohort; characteristics of compliant and non-compliant tag wearers Non-compliant (n=4) Compliant (n=14) Gender Male 4 12 Female 0 2 Age (mean) 31.0 28.2 Ethnicity White 2 2 BAME 1 9 Order type CO 0 6 SSO 4 8 Risk factors Drug use 1 9 Alcohol use 0 2 Housing 2 4 Physical health 0 0 Mental health 2 4 Low motivation 1 0 Practitioner Feedback Practitioner feedback was collected from offender managers (OMs)12. Surveys were sent to OMs in two rounds. Nine responses were returned (from 57 sent out) in the first round of surveys sent in February 2018 and 22 responses were returned (from 38 sent out) in the second round in May 2019, totalling 31 responses. The main themes are presented below. Use of GPS location data Twenty-four had accessed location data. Of those who had accessed location data, 14 respondents had used it to set up Interest Zones, 12 to request heat and other maps, nine to request ‘top location’ information, three to request information on a specific location, and two had used it to set up a Point of Interest. Reported data access correlated with reported confidence using GPS tagging in the management of their case(s). Eighteen respondents stated that they were confident using GPS tagging in their role and eight that they were unconfident or very unconfident, with those indicating confidence reporting on average two types of location data use compared to an average of one type reported by respondents indicating a lack of confidence. Use of location data to manage cases 12 The previous evaluation report contains details of feedback from the PSR staff 17 OMs were also asked how they used the location data to manage their case(s) resulting in many varied uses13. Fifteen replied that they had used it to inform discussions with the tag wearer, 11 to monitor attendance at a place linked to risk of reoffending, 11 to challenge the tag wearer about his or her behaviour in supervision, 11 to find out new information about the tag wearer’s lifestyle, 10 to monitor attendance at substance misuse treatment, seven to praise the tag wearer for compliance and positive behaviour, seven to monitor attendance at a place linked to risk of harm, six to try and locate a tag wearer who had gone missing, and four to protect known potential victims. The average number of types of data use reported by OMs was similar whether they stated feeling ‘confident’ or ‘unconfident’/‘very unconfident’ using GPS tagging in their role. This being the case, while stated confidence appears to impact on reported data access, confidence does not appear to impact on reported data use, which suggests that whether OMs feel confident in using GPS tagging predominately relates to how able they feel in navigating and interpreting the technical aspects of the GPS information being supplied. Respondents who had accessed location data were also asked how easy it was to do so. Of the 21 who responded, 20 felt it was ‘easy’ or ‘very easy’, only one respondent felt it was ‘difficult’. Understanding of the technology For the seven respondents who had not accessed location data, three stated that this was because they were not aware that they could set up an Interest Zone or Point of Interest, two had only been allocated their cases at the end of the GPS tagging requirement, one did not know how to set up an Interest Zone or Point of Interest, one did not think an Interest Zone or Point of Interest would be useful, and one had a service user who did not charge their GPS tag (therefore making location data unavailable). Of the 31 respondents, 26 agreed or strongly agreed that they understood when they could access and use location data, and 22 indicated they knew how to access and use location data. Unsurprisingly, OMs who had not accessed location data were less likely to respond that they understood when and how to access and use location data. Twenty-four respondents agreed or strongly agreed that they understood the restrictions and processes for sharing location data, with only one respondent (who did not access location data) indicating that they did not understand this aspect of managing the GPS tag (the remainder neither agreed nor disagreed). Twenty respondents agreed or strongly agreed that they understood the different ways GPS tagging worked for COs and SSOs. Views on GPS tagging as a probation tool OMs were asked to select their top three from several possible uses of GPS tagging. The uses that were selected most frequently were to monitor compliance with an exclusion zone or other restrictive requirement (n=20) and to deter service users from re-offending (n=17). 13 Multiple responses were possible, so the totals do not add up to 24 18 Other commonly selected uses were to monitor attendance with Rehabilitation Activity/Drug- Rehabilitation and Alcohol Treatment Requirement (RAR/DRR/ATR) appointments (n=10), to use location data to match with potential new offences (n=9), for enforcement and support of breach proceedings (n=7), and to improve risk management (n=6)14. When asked if they felt that GPS tagging was a useful way to tackle prolific offending, 28 of the 31 respondents either ‘agreed’ or ‘strongly agreed’ with the statement. Similarly, when respondents of the endpoint survey were additionally asked if they felt that GPS tagging was a useful way to tackle knife-crime offending, 19 of 22 either ‘agreed’ or ‘strongly agreed’ with the statement. When asked whether GPS location data had improved their ability to manage the case(s) and engage the service user(s) in positive change, OMs’ feelings were slightly more equivocal, although most still agreed or strongly agreed (n=20). Notably, while OMs who had not accessed location data were more likely to disagree or strongly disagree that it had improved their case management and service user engagement, most still indicated that it had done so. In the most recent survey, respondents were asked an additional question about the types of offences for which they feel GPS tagging is most useful based on their experience. Of the 22 respondents, the majority felt that GPS tagging is most useful for prolific theft (n=15), burglary (n=14), and knife crime (n=13). Several also felt that GPS tagging is useful for Serious Group Offending (SGO) cases (n=11), robbery (n=10), sex offences (n=10), and moped enabled crime (n=10). A minority of respondents felt that GPS tagging is useful for violent offences (n=6) and female offenders (n=4). Contact with Buddi Respondents were complimentary about the quality of the communication with Buddi, the GPS tagging provider, with 23 of 30 stating it as ‘good’ or ‘very good/excellent’ and comments describing Buddi as providing regular updates and a valuable service with helpful, responsive staff. A few respondents did raise some issues, however, including the volume of email received and difficulties arranging the removal or changing of tags. 14 Response option ‘to improve risk management’ was listed on the endpoint survey only. 19 Tag wearer feedback Tag-fitting Forty-two tag wearers completed the fitting survey. Positively, all but one stated that they understood why they had received the GPS tag, what they had to do to comply with the order, and how the tag operated with their other requirements. When the tag was fitted, respondents were asked the extent to which they agreed or disagreed with several statements about the anticipated impact of the GPS tag. When asked if they were worried about what their friends and family would think about the tag, most respondents disagreed (n=32). Respondents were also asked to indicate how confident they were that they would complete the tag order successfully. All tag wearers but one described themselves as being either ‘fairly confident’ (n=21) or ‘very confident’ (n=20). Respondents were asked what impact they thought wearing the GPS tag would have on specific aspects of their life (‘better’, ‘worse’, or ‘no impact’). The results are shown in Figure 1 below. Figure 1. Tag wearers’ perception of anticipated impact of wearing GPS tag Generally, respondents felt that the tag would either have a beneficial effect, or, more commonly, no impact at all; very few thought it would have a detrimental impact. Respondents were most positive about the potential impact of the tag on their relationships with family (19 tag wearers felt that the tag would make their relationships with family better). Respondents appeared to be more ambivalent in terms of their expectations of wearing the tag on their relationship with friends, their financial situation, their physical 19 15 7 12 12 13 4 10 3 17 24 30 28 27 28 33 29 34 2 2 1 0 1 0 1 2 1 3 0 2 1 1 0 3 0 3 0 5 10 15 20 25 30 35 40 45 Relationship with family Relationship with friends Money/current financial situation (n=40) Physical health Mental well-being Housing situation Employment situation Ability to go out/socialise Education situation Anticipated impact of wearing tag on.... Better No impact Worse N/A 20 health, their mental well-being, their housing situation, employment status, ability to go out or socialise, and their education situation. In every case the most frequent response was that the tag would have no impact. When asked what the anticipated impact of wearing the tag would be on their life overall, 35/42 respondents felt that it would get better, while five felt that the tag would have no impact; one thought that their life would get worse. In terms of the impact of the tag on their offending behaviour, 37/42 respondents thought it would lead to an improvement and one thought it would get worse. Respondents were also asked to provide general comments about their expectations on having the tag fitted (35 responses were received). The most popular assertion (mentioned by 21 respondents) was that the tag could help stop them from re-offending and/or make them ‘think twice’ about their behaviour in relation to potential re-offending, typified by the following statements: Twelve respondents said that the GPS tag was preferable to a curfew tag (in that it was less restrictive). Similarly, 12 respondents said that wearing the tag was a preferred alternative to going to prison, recognising the likely negative impact on their lives of serving a custodial sentence. Ten respondents said that the tag would assist them in ‘sorting their lives out’. Six respondents mentioned the positive impact of the tag in combination with engaging with support services, such as substance misuse services. Other factors mentioned were that the tag would keep them out of situations where they might cause harm to others, would assist in enforcing the restrictions of their order, would help to improve the wearer’s familial relationships, and that it would show probation they were not offending. Typical quotes reflecting these sentiments were: “I like this tag because I don't have a curfew and gives me space to try and get my life together. I don't want to go to prison but recognise that I still have a drug addiction.” Respondent “I understand that I must stay away from my partner and the tag stops me going to her address. I think it will help me sort things out by reminding me I can't go to her address. I hope this and my counselling will make life better for me.” Respondent “I hope wearing a tag for the first time will help me think about my behaviour, stay away from drink and keep using the support services that have been offered to me … I know I have a problem with drink but need interventions like the tag to make me think about the consequences.” Respondent “I am 45-years-old and sick of offending. I hope wearing this tag will help me distance myself from crime and those who are a bad influence … Knowing police and probation can see my whereabouts and link crimes to me is going to help me think about my actions.” Respondent 21 One respondent described the fitting of the tag as his ‘last chance’, but felt he was being set up to fail due to the inability of drug services to see him to address his drug dependency. Similarly, another respondent expressed concerns that the tag would not make any difference as she was wearing it for a short period of a few weeks and had issues with drug use and homelessness. Tag removal Tag wearers were also asked to complete a similar questionnaire when their tag was removed. 24 responses were received. Overall, across several questions tag wearers were generally positive on the experience. To illustrate, respondents were asked to rate (on a scale of 1-7, with 1 being ‘very negative’ and 7 ‘very positive’) their overall experience of being on the tag. The results are shown in Figure 2 below, and clearly indicate respondents’ positive perceptions (although it should be stressed that, as these individuals had completed the tag, they might be expected to be more positive/motivated). Figure 2. Tag wearers’ experience of being on a GPS tag This positive outlook is reflected in responses to several related questions. For example, when asked if wearing the tag was better than they thought it would be, 21 respondents agreed, whilst only three disagreed. The tag wearers responded similarly to a question asking if they found the tag comfortable to wear, with 19 agreeing and five disagreeing. Respondents were also asked whether they had worried about what their friends and family thought about the tag, and most disagreed (n=18). The removal survey repeated the questions from the fitting survey about the impact of the tag (‘better’, ‘worse’, or ‘no impact’) on a variety of aspects of respondents’ lives. All 24 respondents thought their offending behaviour had improved because of wearing the tag, and, in terms of life in general, 21 respondents thought the tag had made it better and only 0 2 4 6 8 10 12 14 1: Very negative 2 3 4 5 6 7: Very positive No. of respondents 22 one thought the tag had made it worse. Responses to the other questions are shown in Figure 3 below. Figure 3. Tag wearers’ perceptions of the impact of wearing the GPS tag The pattern of responses in Figure 3 echoes the ambivalence of the responses received in the fitting survey, with the most common response across all questions being that the tag had had no impact on the specific aspect of the respondents’ life, except for family relationships and mental well-being which most respondents felt had improved as a result of wearing the tag. Respondents were also asked to provide general comments about their experiences of wearing the GPS tag (23 responses were received). The comments reflect the responses given to earlier closed questions and echo comments given by respondents when the tag was fitted. Sixteen people reported that the tag had helped them stay out of trouble and away from crime. Thirteen people mentioned that the tag had helped them think about or reflect upon their actions. Thirteen also stated that they preferred the Buddi tag to a curfew tag (again, in that it was less restrictive) and, similarly, 11 specifically mentioned the advantages of wearing the tag as an alternative to serving a custodial sentence. Five people said that wearing the tag meant that others who posed a negative influence or threat of harm stayed away from them. Some respondents commented that wearing the tag had in general improved their lives or ability to engage in positive change (mentioned by five respondents), been a positive or better than expected experience (mentioned by three respondents) or left them feeling positive about the future (also three). Specific positive impacts of wearing the tag on their lives described by respondents include securing or maintaining stable accommodation (mentioned by six respondents), securing or maintaining employment (mentioned by two respondents), and improving family relationships (mentioned by three respondents). 10 respondents stated 15 11 5 10 14 6 3 7 0 8 11 19 13 9 18 19 13 23 1 2 0 0 0 0 1 4 0 0 0 0 1 1 0 1 0 1 0 5 10 15 20 25 Relationship with family Relationship with friends Money/current financial situation Physical health Mental well-being Housing situation Employment situation Ability to go out/socialise Education situation Impact of wearing the tag on... Better No impact Worse N/A 23 that they would wear the tag again or were willing to continue to wear their current tag. Six respondents said that the tag was comfortable to wear. Typical responses included: However, respondents also raised several negative aspects of wearing the tag. Nine respondents mentioned problems they had charging the tag (not the act of charging the tag itself but being organised enough in their own daily life to do so), with one respondent suggesting that having a charger at their probation or drugs service might help in this respect. Other issues raised by respondents included that the tag was uncomfortable to wear (mentioned by five respondents), that they felt embarrassed or ashamed wearing the tag (mentioned by three respondents), that wearing the tag for three months was not long enough to change long-term behaviour (mentioned by two respondents), and that wearing the tag harmed their employment prospects (mentioned by one respondent). Examples of such comments were: “I tried cutting off the strap to go and do crime, but I could not get it off. Because of this it stopped me and has helped me to think about my actions. Best tag I have worn, and I have worn a lot. If I had the choice, I would wear it again for longer.” Respondent “I prefer the Buddi tag more than being on curfew because I work night shifts. To lose work because of a curfew would have made my living situation a lot worse and probably increased my risk of re-offending.” Respondent “To be honest, wearing this tag was better than I thought. It has made me think twice about offending, helped me stay away from people who were bad influences, stopped me from getting attacked when caught by someone from a gang who had beef with me … My mum is ill in hospital and being out of prison and not on curfew means that I can spend valuable time with her. I hope the experience of wearing the tag has motivated me enough to keep it going … I feel positive about my future.” Respondent “As a female, I have not been able to wear a skirt as the tag will show. I have found it restricts my movements too much. The tag has felt uncomfortable to wear.” Respondent “I did find it hard to get into a routine of charging my tag and got fed up with probation and Buddi constantly telling me to charge the tag.” Respondent “I can see how this tag … is good for some offenders but three months seems a short period of time to see any real results with changing people’s attitudes or offending. “ Respondent 24 Police Crime Mapping Under the crime mapping intervention, which began in September 2017, a dedicated police analyst used an automated system to map the movements of tag wearers on SSOs against police crime data. Where the analyst believed that there was a sufficiently close match, details were forwarded to the relevant police IOM team for further investigation. A detailed account of the crime mapping intervention can be found in E&I’s second evaluation report (MOPAC 2018). There is no additional information to be added in this report because the funding for the police analyst post had ceased at the point that the previous interim report was written, and no additional mapping activities were subsequently undertaken by the police. The second interim evaluation report identified the low number of significant matches generated as a result of the crime mapping (46 were referred to the police IOM teams, which was described in the report as ‘disappointing, bearing in mind the large number of crimes up- loaded onto the system’ (over quarter of a million). In addition, of these 46, in only two cases was the decision taken to look at the incident further; one case had been allocated for secondary investigation, and in the other, an EAR request had been sent to Buddi. However, the report recognised the small size of the GPS-tagged population able to generate ‘hits’ (only 22 individuals in total were eligible for crime mapping15) and that the type of offences for which the matches were generated (largely theft and burglary) were likely to provide the greatest latitude in terms of the reported time of the offence and thus, the greatest difficulty in providing evidence that the individual had committed the offence. The report concluded ‘in these circumstances, it is perhaps unsurprising that the level of significant matches was so small. It is difficult to know, as a result, whether the low number of yields from the crime mapping reflects failings in the implementation of the scheme, or in the theory behind the concept of crime mapping, particularly bearing in mind the type of offences which many of the POP cohort are committing (thefts and burglaries)’. 15 Individuals who were on SSOs and had been compliant. 25 4. Discussion A total of 135 tags were imposed as part of the GPS tagging pilot (117 on the POP cohort and 18 on the knife crime cohort). The majority of Electronic Monitoring requirements (93% - 125 of 135) were imposed alongside an RAR – consistent with the intended use of the tags as part of a package of support, monitoring and punishment. The completion rate without recall for the POP cohort was 52%, that for the knife crime cohort higher at 78%, albeit on a much lower number of cases. The last figure is particularly interesting bearing in mind the subsequent agreement for MOPAC to use GPS tags with knife crime offenders released on licence from London prisons. Lack of attendance, failure to charge the tag and non-engagement were the three most common reasons for non-compliance. Overall, findings from the research show that the GPS tagging pilot was implemented well (something never to be taken from granted when implementing innovation) and has been well received by practitioners and tag wearers alike, although the findings are based on small samples. Offender managers were generally positive about the GPS tag and were willing to use interest and exclusion zones as part of supervision. Tag wearers had positive expectations of the tag. They were confident that they would successfully complete the tag order and knew what they needed to do to comply. Furthermore, tag wearers thought that the tag would have a positive impact on their life and lead to them committing less crime. At tag removal, they thought that being on the tag had improved their offending behaviour and had made their life better in general (although, again, the number of respondents was small). However, there were some concerns expressed (from tag wearers and practitioners) about the difficulties caused for this cohort by the requirements to keep the tag charged. It is clear from the data that there are still issues around the identification of suitable individuals to be given the tag, reflected in examples of early non-compliance. Question marks about the effectiveness of the crime mapping element of the pilot remain, and it is unclear whether these arise as a result of theory or implementation failure. Nevertheless, the successful implementation of the technology, the overall compliance rate for the scheme, particularly for the knife crime cohort and the positive views expressed by practitioners and service users about their experience with the GPS tag, suggests grounds for optimism. 26 5. References Gies, S V., Gainey, R., Cohen, M. I., Healy, E., Yeide, M., Bekelman, A. & and Bobnis. A. (2013) Monitoring High-Risk Gang Offenders with GPS Technology: An Evaluation of the California Supervision Program Final Report. MOPAC (2018a) GPS Tagging: First Year Interim Report https://www.london.gov.uk/sites/default/files/mopac_gps_tagging_second_interim_report _2018.pdf MOPAC (2018b) Persistent Offender Programme Interim Evaluation Report 2 https://www.london.gov.uk/sites/default/files/mopac_persistent_offender_programme_int erim_report_2018.pdf Gripping the Offender Interim Evaluation Report February 2017 MOPAC Evidence and Insight 2 Contents Executive Summary ...................................................................................................................... 3 1. Introduction .............................................................................................................................. 5 Background to the Management of Prolific Offenders ................................................................ 5 The Gripping the Offender Pilot .................................................................................................. 5 The Gripping the Offender Cohort ............................................................................................... 8 2. Methodology .......................................................................................................................... 10 3. Results ..................................................................................................................................... 12 Performance Analysis ................................................................................................................. 12 Process Learning ........................................................................................................................ 16 4. Discussion ............................................................................................................................... 23 References .................................................................................................................................. 24 3 Executive Summary The Gripping the Offender (GtO) pilot aims to test a new partnership approach to tackle the most prolific offenders across the whole Criminal Justice System (CJS). Commissioning for the two-year pilot began in April 2015 and the first phases of delivery began in February 2016 in the North and East London Local Justice Areas. The pilot aims to deliver an end-to-end approach across the CJS, including:  Courts: enhancing the pre-sentence report (PSR) capacity to speed up sentencing time  Offender Management (OM): providing additional support to resettle and rehabilitate offenders within the cohort, particularly young adult males and females  Pathways: addressing the mental health and employment needs of prolific offenders  Policing: providing the police with a greater ability to complete research and intelligence gathering and focusing on offender based tasking  Electronic monitoring: testing the use of GPS tags as an electronic monitoring requirement of a Community Order or Suspended Sentence Order The evaluation of the pilot is being conducted by the Evidence and Insight team at MOPAC. This interim report will share emerging findings from the research, including performance analysis of the enhanced services for the first 11 months of delivery and staff feedback from interviews, surveys and focus groups. This report does not include impact analysis due to the low numbers of interventions completed. The report will provide an update on the overall progress of the evaluation and next steps. Key Findings  Courts: Between 6th February and 31st December 2016, a total of 225 offenders received the PSR service, including 37 18-25 year olds; 38% of reports were delivered within the 5 day target.  OM: Referrals into the enhanced OM services for young adult males were slow to begin with, but have steadily increased over the last four months. Forty-four young males have started mentoring and three have completed the service. Groups were scheduled to begin for the Right Track and TIGER programmes in January 2017. One female has been accepted for the personalised budget service.  Pathways: Between 1st July 2016 and 31st December 2016, 102 referrals were made to the mental health service resulting in the commencement of 24 direct interventions and 305 case consultations with practitioners. Seventy-one referrals were made into the employment pathway, with two having secured employment.  Policing: Since going live in July 2016, the policing research hub has completed a total of 110 intelligence development plans and 320 offender-based taskings have been carried out on these cases.  Electronic monitoring is due to be rolled out in courts from the 13th March 2017. Training of magistrates and probation staff is currently underway. A number of positive messages emerged from stakeholder feedback, including: strong support for the rationale for the pilot; the positive impact the pilot has had on partnership working; and the additional resources available for offenders through the enhanced service offer. Practitioners had mixed views about the delivery of certain aspects of the pilot, with the PSR service, prison co-ordination service and mental health pathway rated the most positively. Overall, delivery was thought to have been compromised by: perceived challenges within probation and the poor performance of the CRC; the complexity of the pilot; and resource problems within boroughs. 4 Overall, most respondents agreed GtO has the potential to impact on a number of important outcomes, but many felt at this stage the pilot has not transformed this potential into tangible results. As a result of the phased approach to delivery and the slow commencement of some services, it has been recognised that further time is required to develop and embed the work already undertaken and to understand the impact of the pilot on reoffending and offender management. A considerably higher volume of referrals and engagement is now taking place, providing confidence that a further operational period will lead to the proposed volumes for the enhanced services being met. Consideration is currently being given to develop and extend the pilot for a further 18 months. It has also been recognised by the Implementation Team that the pilot has placed additional focus on the functionality of the IOM baseline; any future extension will continue to support and address adding consistency to IOM. Over the next six months, the evaluation will continue to gather the views of staff involved in the pilot and will also seek to capture the views of individuals receiving the enhanced services through interviews and case studies. The impact of the PSR service on court timeliness and sentencing outcomes will also be explored. The long-term aim for the evaluation is to robustly examine the impact of GtO on proven reoffending and cost to the system, but this will be dependent on having an adequate sample size, robust data being available, and the identification of an appropriate comparison group. 5 1. Introduction Background to the Management of Prolific Offenders It is accepted that a minority of offenders commit a disproportionate amount of crime. This has been the driving principle behind many offender management initiatives over the past 15 years (e.g., the Persistent Offender Programme, Prolific and Other Priority Offender Scheme, the Diamond Initiative). Prolific offenders place a disproportionate heavy demand across the entire Criminal Justice System (CJS) and therefore present the biggest opportunity for impact. Integrated Offender Management (IOM)—a strategic partnership approach to tackling prolific offenders—was launched nationally across the UK in 2004. The overarching aim of IOM is to bring a ‘cross-agency response to the crime and reoffending threats faced by local communities by managing the most persistent and problematic offenders identified jointly by partner agencies working together’.i However, it has been recognised that IOM has been adopted at different rates across London, and that the approach adopted varies significantly between London boroughs, often reflecting local circumstances and priorities. Differences include: cohort size and selection; staff resource allocation; governance arrangements; performance management; information sharing; and multi-agency activity. The North West London IOM pilot evaluation1 reported tentatively positive results in terms of proven reoffending: there was a 12 percentage point improvement comparing the predicted versus actual OGRS scores of offenders on IOM.2 Furthermore, there was a 25% reduction in offences during the first year of IOM, suggesting a lower volume of offending subsequent to IOM. The evaluation also suggested there is a narrow ‘window of opportunity’ when working with prolific offenders. Of those who reoffended, the majority did so quickly: 58% within the first 3 months and 79% within the first 6 months. Despite the impact of previous offender management initiatives, adult reoffending remains a key priority for the Mayor’s Crime Reduction Board as reoffending rates remain stubborn, driving consistently high state expenditure. Overall reoffending rates stand at 25% but the most prolific offenders have predicted reoffending rates closer to 80%. It is therefore this cohort that presents the biggest opportunity, both in terms of reducing crime and generating savings to the system. The Gripping the Offender Pilot The Gripping the Offender (GtO) pilot aims to test a new partnership approach to tackle the most prolific offenders across the whole criminal justice system with the aim of delivering significantly better outcomes and simultaneously driving down demand across the public sector. MOPAC successfully applied to the Home Office Police Innovation Fund and received a total of £1.4m over two years to design, test and evaluate an approach that creates an enhanced whole system response to those offenders causing the most harm to communities and the greatest demand on 1 The London NW IOM scheme commenced early 2012. MOPAC 'Evidence and Insight' conducted the evaluation based upon 418 offenders. The individuals were selected based on high OGRS and specific offence types. More detail available upon request. 2 Offender Group Reconviction Scale risk of reoffending in the next 12 months. 6 the broader criminal justice system. Commissioning for the pilot began in April 2015 and delivery began in February 2016. The pilot has a phased delivery approach, with different elements being launched as and when they become available. The GtO pilot includes enhancements across the whole CJS and is fully supported by the London Reducing Reoffending Board. Key partners engaged include: Metropolitan Police Service (MPS), National Probation Service (NPS), London Community Rehabilitation Company (CRC), and Local Authorities (LA). The enhanced services commissioned through GtO aim to deliver an end to end approach—from arrest to conviction and release back into the community—where all partners are working together to target and support prolific offenders throughout the whole CJS. An enhanced offer has been commissioned in the following areas: Courts A focus within the court aspect is to enhance the pre-sentence report (PSR) capacity to enable more on-the-day reports to be completed and speed up the sentencing time for individuals who are prolific in their offending. This strand of the pilot is supported by analysis which demonstrates the disproportionate demand of prolific offenders within the court system. The North West IOM evaluation showed that a total of 20,441 offences had been committed by 418 IOM offenders prior to the scheme: an average of 49 convictions each. The evaluation also reported that, although individuals on IOM proceeded through the CJS quicker than ever before (on average 12 days quicker over the first year of IOM), this was due to the reductions in the time from offence to arrest (21 to 6 days). The time taken from arrest to conviction did not significantly change subsequent to IOM (39 to 36 days), suggesting prolific offenders were not receiving a heightened response at court. Analysis was completed to identify the courts where offenders in the cohort were most likely to appear. As a result, the enhanced PSR service was rolled out in Thames and Highbury Corner Magistrates’ Courts. Two dedicated probation officers operate within these courts and prioritise GtO cases for on-the-day and other fast delivery reports. The court officers also undertake DRR and ATR assessments. Offender Management The CRC was commissioned to deliver additional support to resettle and rehabilitate offenders within the cohort. There is a universal prison co-ordination service for all GtO cases in custody, as well as further enhancements targeted at specific priority groups within the overall cohort. Young adult males (aged 18-25) are a priority focus due to the significant volume of the GtO cohort (36%) falling within this transitional age group, as well as their high reoffending rates. Young adult males will receive an increased intensity of support on release from prison alongside mentoring and interventions to address their specific needs (e.g., antisocial peers, family problems).3 These enhanced services are also available as part of a Rehabilitation Activity Requirement for young adult males in the community. 3 These services are available ‘Through the Gate’ for individuals in custody and as part of a Rehabilitation Activity Requirement. The mentoring service is being delivered by four providers across the eight pilot boroughs: Caring for Ex-Offenders, Osmani Trust, Reaching All People, and New Choices for Youth. A family intervention (Building Bridges) is being delivered by Pact and two cognitive-behavioural therapy programmes (Right Track and Transforming Inside Growing Emotional Resilience; TIGER) are being 7 Females are also a priority focus due to their bespoke needs. Research suggests that interventions with female offenders are more likely to have a positive impact when they are responsive to the specific needs of women.ii Personalised budgets will be available for 25 females within the GtO cohort. Offender managers will be able to spot purchase interventions to meet the individual needs of each woman. Policing The policing offer introduced as part of GtO aims to provide the MPS with a greater ability to complete research and intelligence gathering on select offenders within the GtO cohort. Intelligence Development Plans will be available for all GtO cases that are “Red” rated, including crime mapping, association mapping, open source research and other intelligence products. The intelligence profile should be used to influence release and resettlement plans, and sentence plans. Offender based tasking—as opposed to tasking based on hotspots or offence types—is a new approach being used by the MPS and will also form part of this strand. This tasking process will be a feature of both fast time problem solving through daily pace setter meetings and also in longer term problem solving through Tactical Tasking and Coordination Groups. Pathways NHS England has funded a pathway to address the mental health needs of the GtO cohort. Evidence suggests that individuals with mental health needs are disproportionately represented within offender populations. Data shows that 72% of prisoners face at least two mental health problems; 95% of imprisoned young offenders have a mental health disorder; and childhood and adolescent trauma is widespread amongst offenders.iii As part of the mental health pathway, a consultative function will be provided for all cases on the GtO cohort, as well as direct face-to-face interventions for offenders. The aims of the mental health pathway are: to improve offender engagement with GtO practitioners; improve confidence and competence in supporting offenders with mental health issues; and provide GtO practitioners with a range of techniques to use when working with this cohort. A second pathway has been established to pilot the delivery of employment opportunities for up to 20 identified GtO nominals. Evidence suggests that employment is one of the biggest barriers to the reintegration of ex-offenders.iv Data shows that over 60% of prolific offenders in London have an employment, education or training need.4 The employment pathway delivered as part of GtO will recruit, employ and place ex-offenders into contracts with commercial partners on fixed- term 6-month contracts. GPS tagging This element will pilot and test the impact of compulsory GPS tagging on offender behaviour with the purpose of informing the effectiveness of GPS tagging as an offender management tool. An electronic monitoring requirement will be made available to courts to use as part of a Community delivered by Rise Mutual. Young adult males managed by the CRC will also receive an increased intensity of support from their offender manager. 4 Based on OASys information for prolific offenders drawn from NPS/CRC caseloads. 8 Order or Suspended Sentence Order. This aspect is projected to go live on 1st March 2017 and will run for 12 months with a total of 75-100 offenders to be tagged during this time. The Gripping the Offender Cohort The GtO pilot targets individuals in London who are most prolific in their offending. Analysis of the cohort across London was undertaken to identify the most suitable location for the pilot. Volume was considered alongside reoffending rates, acquisitive crime performance, and IOM investment. This analysis identified the North and East London Local Justice Areas as a priority location for the pilot covering the boroughs of Camden, Enfield, Hackney, Haringey, Islington, Newham, Tower Hamlets and Waltham Forest. To be included in the pilot an individual must meet the following criteria: - Convicted of an offence - Aged 18 or over - Resident within one of the eight pilot boroughs - An OGRS 2 year5 score of either (a) 75%+ or (b) 50-74% with either a robbery or burglary offence in the preceding 12 months Individuals managed by both the NPS and CRC are in scope. Identification of GtO eligible individuals was first undertaken in July 2015 from data held by the NPS and CRC. This data showed there were 1,187 individuals residing within the eight pilot boroughs who met the GtO eligibility criteria. Consistent with their status as prolific offenders, GtO eligible individuals had been arrested an average of 53 times each. The average age at first arrest was 15 years and 62% of the cohort was arrested for the first time before the age of 16. The cohort had an average of 33 prior convictions and 48% received their first conviction before the age of 16. Within the last year, almost 60% of the cohort had been convicted of a new offence and almost 90% received a new conviction within the last 3 years. The data was pulled again in December 2015 to identify an updated cohort of GtO eligible individuals for the commencement of the enhanced services. These cases then went through a quality assurance process by local IOM teams. The final cohort consisted of 1,363 GtO cases in the eight pilot boroughs: 67% managed by the CRC and 33% managed by the NPS (see Figure 1 for a breakdown of cohort numbers by borough). Although these cases provided a starting point for the pilot, the GtO cohort is designed to be dynamic in that cases are regularly added and removed by local IOM teams.6 5 Offender Group Reconviction Scale risk of reoffending in the next two years. 6 Cases are removed from the cohort if they move outside of the pilot area, if their sentence is terminated, or if they no longer meet the OGRS criteria (e.g., the robbery or burglary offence falls outside the 12 month criteria). 9 Figure 1. Initial GtO Cohort Split by Borough The initial cohort was predominantly male (91%) and in the younger age brackets (36% were 18- 25 and 29% were 26-34). The average age of the cohort was 31 years (range 18-69 years). More offenders of all ages were managed by the CRC (73%) than the NPS (27%). The average OGRS2 score was 82% (range 52% to 98%), indicating an 82% likelihood of reconviction in the following two years. Thirty-eight percent were identified as being in custody, 38% were on a community order, and 24% were in the community on licence. 10 2. Methodology The evaluation is being carried out by Evidence and Insight (E&I): a dedicated in-house social research function based within MOPAC. The team informs strategic decision-making through robust crime and performance analysis, as well as qualitative and quantitative social research methodology. E&I are using a variety of research methods to evaluate the GtO pilot, including surveys, interviews and focus groups. Key performance measures will also be assessed throughout the pilot, and the evaluation aims to robustly explore the impact of GtO on re-offending outcomes and cost to the system. This analysis is dependent on the provision of appropriate data; without this, the evaluation will consist of only process and performance analysis. Performance Analysis A range of performance data will be gathered throughout the pilot on a monthly basis to determine what is happening ‘on the ground’. Performance data will include: cohort size, numbers eligible for the enhanced services, numbers accessing the services, numbers of completers vs. non- completers, age, gender, ethnicity, needs. Process Staff feedback will be gathered using online surveys (3 waves), interviews (with ~20-30 staff at 2 time points) and focus groups (at 2 time points across pilot delivery). These methods will enable the research team to capture learning around the design and implementation of GtO, and to capture change over the course of the pilot. Themes to be explored include: set-up and design, partnership working, training, staffing, governance, data collection, data management, ways of working, products developed, use of data, business change, multi-agency work/partnerships, new approaches in managing offenders, perceived impacts, lessons learned, and challenges. Offender insight and experience of the pilot will also be gathered through interviews and case studies. The anticipated sample would be 20-30 offenders. Impact This aspect will examine the impact of GtO on a number of primary (e.g., proven reoffending, severity, frequency, time taken to reoffend, speed through the system, disposal type) and secondary outcomes (e.g., offender attitudes, partnership involvement, data sharing between agencies, borough crime levels, offender victimisation, victim types, demand reduction, needs being met, timely referrals). The primary and secondary outcomes have different appropriate follow-up periods. Impact of the PSR service on court timeliness and sentencing outcomes will be examined first, because these outcomes do not require a long follow-up period. Early impact for the remaining services will be examined using arrest and charge rates within a six month follow-up period. Any more robust 11 measure of impact (e.g., proven reoffending) will require a longer follow-up period (18 months for proven reoffending). The aim will be to use a quasi-experimental design to statistically match offenders who receive the GtO enhancements with similar GtO eligible offenders who do not receive the enhancements. This analysis will only be possible if: there is robust data available; successful implementation of the pilot; the sample size reaches over 150 cases; and a suitable control group can be identified from the available data. Cost-benefit The evaluation may also include a financial model outlining the cost benefit of the pilot as well as outlining the cashable and non-cashable savings delivered by the model. This aspect is vital to support the longer term sustainability of interventions; external expertise may be commissioned to deliver this element. Overview of report This is the first interim evaluation report for GtO which will share progress and initial findings from the evaluation. It will consist primarily of process and performance analysis. The proposed plan for the remainder of the evaluation will be presented at the end of this report. 12 3. Results Performance Analysis7 Pre-sentence Report Service Between the 6th February and 31st December 2016, a total of 947 offenders came to the attention of the probation officers based at Highbury Corner and Thames Magistrates’ court (see Table 1). These cases were predominantly male, were 35 years old on average, and most frequently appeared in court for theft (33%), violence (16%) and burglary/robbery (11%) offences. Eighty- two percent of offenders pled guilty to their qualifying offence. Of the eligible cases that came to the attention of the court officers, 225 (or 24%) received a PSR as part of the enhanced service8; 86 reports were delivered within 5 days of the offender’s plea and 64 of those were delivered on-the-day. The desired outcome for the service is for 90% of PSRs to be delivered within 5 days; the data shows that only 38% were delivered within 5 days. The main reasons why reports were not completed within 5 days were: the court officer did not become aware of the case until more than 5 days after the plea; the case was sent to Crown Court; the offender failed to attend their appointment; or a DRR/ATR assessment was carried out.9 Table 1. Pre-sentence report service performance data Thames Magistrates’ Court Highbury Corner Magistrates’ Court10 Total All 18-25 All 18-25 All 18-25 Number of cases came to attention 588 87 359 89 947 176 Number of reports delivered 150 20 75 17 225 37 Number of reports delivered on-the-day 20 3 44 8 64 11 Number of reports delivered within 5 days 39 4 47 10 86 14 Figure 2. Sentence type for cases who received the PSR service and for cases sentenced without report 7 Full performance data is available via the GtO performance dashboard: http://data.london.gov.uk/resources/gto-dashboard. Please contact Julia.Yesberg@mopac.london.gov.uk if you require the password. 8 Reasons why offenders did not receive a report included: they were sentenced without a report (41%); the case was adjourned (11%); the case was committed to crown (7%); and the trial was in another court (5%). 9 The court probation officers have carried out a total of 88 DRR and ATR assessments: 78 in Thames and 10 in Highbury Corner. 10 There was a gap in service provision in Highbury Corner Magistrates’ Court for most of the latter half of 2016, which is reflected in the lower number of individuals receiving the service at this court. 35% 28% 30% 6% 1% Sentence Type for cases who received PSR service Community Order Suspended Sentence Order Custody Fine Deferred Sentence 8% 11% 56% 21% 4% Sentence Type for cases sentenced without a report Community Order Suspended Sentence Order Custody Fine Conditional Discharge 13 As the above charts show, offenders who received the PSR service were more likely to be sentenced to a community order (CO) or suspended sentence order (SSO) compared to offenders who were sentenced without a report (63% vs. 19%). Offenders sentenced without a report were more likely to receive a custodial sentence (56% vs. 30%). Of the 37 18-25 year olds who received a report and were eligible to receive an enhanced Rehabilitation Activity Requirement (RAR) as part of the enhanced offender management services, 12 (32%) received it. The remaining 18-25 year olds were sentenced to a CO or SSO without the enhanced RAR (n=19), custody (n=9) or a fine (n=1). Enhanced Offender Management Services Prison co-ordination service Between 25th April and 31st December 2016, the prison co-ordination role has impacted on a total of 293 custody cases located in a variety of prisons, including Pentonville (22%), Thameside (15%), Isis (11%), Brixton (5%), Bronzefield (5%), Feltham (4%), and Wormwood Scrubs (4%). The prison co-ordinators have also carried out tasks with 79 cases on a community sentence and 433 offenders on licence. The specific tasks undertaken by the prison co-ordinators include: notifying prison and/or offender manager (OM) or senior probation officer (SPO) of new cases; flagging cases as GtO on Delius; identifying cases as eligible for the enhanced services and sending in referrals; case discussions with OM or SPO; referrals to local IOM team; and providing feedback to IOM panels about progress, resettlement plans and release dates. Enhanced services for young adult males Between 25th April and 31st December 2016, 98 referrals were made to the mentoring service, 40 to the Building Bridges programme, 72 to the Right Track programme, and 48 to TIGER (see Figure 3 for a chart tracking referrals by month and a comparison by borough). Referral rates were low over the first few months of delivery but picked up in September 2016. Referral numbers vary by borough, with Tower Hamlets having the most referrals across all services. Figure 3. Number of referrals made to the enhanced services for young adult males by month and borough 0 5 10 15 20 25 30AprilMayJuneJulyAugustSeptemberOctoberNovemberDecember Mentoring Building Bridges Right Track TIGER 0 5 10 15 20 25 30 35CamdenEnfieldHackneyHaringeyIslingtonNewhamTower HamletsWaltham Forest 14 Eighty-two referrals have been accepted for mentoring (see Table 2). To date, the mentoring providers have engaged with 48 cases and 4 individuals have completed the mentoring service. Twelve young adults are due to commence the Right Track programme in January, and a group is also scheduled to commence the TIGER programme. No programmes have been scheduled for Building Bridges at the time of writing. Table 2. Mentoring performance data In addition to the above services, young adult males have received increased contact with their offender manager. Cases released between 1st October and 31st December 2016 received an average of 1.18 appointments per week, meeting the target of weekly appointments. Furthermore, 90% of cases released prior to 30th September 2016 received fortnightly appointments for the remainder of their supervision period. Female Personalised Budgets Four females have been referred to the personalised budget service since it launched on 15th November 2016: two from Tower Hamlets, one from Enfield and one from Islington. Of these referrals, one has currently been accepted. This budget will be used to spot purchase the following interventions: Moving On Resettlement Coaching, Female Wellbeing and Development Programme, and a Gender Specific Mental Health Intervention. Pathways Since the Mental Health pathway went live at the beginning of July 2016, a total of 102 referrals were made to the service. From these referrals, the forensic mental health practitioners (FMHPs) have undertaken 272 consultations with offender managers, 33 consultations with other professionals, and 45 assessments with service users. Twenty-four service users have commenced a care plan with a FMHP, which include the following direct interventions: motivational interviewing (11), psycho-education (9), supportive programmes (6), mood management (4), anger management (3), brief intervention on coping strategies (3), crisis management (3), problem solving (3), resilience (2) and anxiety management (2).11 The Employment pathway has received 70 referrals since going live in July 2016. Two referrals have been accepted. Reasons why referrals have not been accepted include: the referral does not reside in a pilot borough; the referral is in custody; the referral is an identified drug user; and no full referral form returned. 11 An individual may receive more than one type of intervention. Mentoring Provider Referrals made Referrals accepted Number of clients seen Number completed Caring for Ex-offenders 6 3 0 0 Osmani Trust 38 30 21 0 Reaching All People 30 29 25 4 New Choices for Youth 24 24 3 0 15 Policing Since the policing intelligence support function went live in the first week of July 2016, a total of 101 intelligence development plans have been requested by the eight boroughs and 110 plans have been actioned by the research hub and sent back to the pilot boroughs. A total of 320 offender based taskings have been carried out by police on these cases. Table 3 shows there is large borough variation in the use of this service, with Haringey, Hackney and Waltham Forest making the most requests for intelligence development plans and carrying out the most subsequent taskings. Table 3. Policing intelligence support function borough comparison Camden Enfield Hackney Haringey Islington Newham Tower Hamlets Waltham Forest Total Intel Dev Plans Requested 7 14 16 33 5 0 9 17 101 Intel Dev Plans Received 6 14 16 36 6 4 10 18 110 Offender Based Taskings 0 0 110 182 0 2 0 26 320 Enhanced service referral overlap GtO is designed to be a whole systems approach to offender management where prolific offenders are given enhancements across the CJS. As such, there has been significant overlap in referrals to the GtO enhanced services. Of the 225 offenders who received the PSR service, 125 also received the prison co-ordination service, 13 were referred to an enhanced OM service for young adult males, 2 to the female personalised budget service, and 10 to the mental health service. Of the 98 cases referred to mentoring, 64 were also referred to Right Track, 36 to TIGER and 35 to Building Bridges. Forty offenders were referred to both Right Track and TIGER and 28 to Right Track, TIGER and Building Bridges. Of the 102 referrals to mental health, 3 had also been referred to an enhanced OM intervention.12 12 At this stage, we are unable to determine the overlap across the Blue Sky and Policing referrals. Individual level data has been requested. Key learning The pre-sentence report service has impacted on 225 individuals over 11 months (an average of 20 reports per month). Offenders who received the service were more likely to receive a CO or SSO compared to those sentenced without a report. Timeliness of reports has not reached the target of 90% being delivered within 5 days; only 38% of reports were delivered within 5 days. Nine young adult males have been sentenced to an enhanced RAR out of the 37 eligible young people to receive the service. In the first few months of delivery, referral numbers to the enhanced offender management services for young adult males were much lower than expected. There was a large spike in referrals in the month of September (mentoring referrals increased to 22 in September from an average of 4.5 per month over the previous 5 months), reflecting improvements to referral processes. Referral numbers have been maintained over the last few months. 16 Process Learning Staff feedback was gathered through interviews with practitioners in April 2016 (n=18); an online survey with practitioners, service providers, and strategic staff in September 2016 (n=44); and a focus group with service providers in November 2016. Given the size of the research cohort, some caution should be used when considering results. Design of the pilot There was positive feedback from staff about the rationale for the pilot. Respondents supported the rationale and thought the project aims were “relevant” and “commendable”. Others noted that the pilot presents a unique opportunity to focus on particular cohorts of offenders and to make a real difference to individuals. Many respondents commented that the pilot has a great deal of potential. Although there was support for the rationale for the pilot, some staff thought the pilot was overly ambitious and contained too many elements, making the project complex and confusing to practitioners. Furthermore, some respondents thought the amount of enhanced services on offer were “starting to feel like overload.” Other feedback about the design of the pilot related to a perceived disconnection between those designing the pilot and the realities of the frontline. As one respondent commented, “there could be more join up of commissioning vision with operational knowledge of what works/is needed on the ground.” Finally, some staff criticised the services available for not being tailored to the needs of all service users and for not taking into account local variations in existing service provision. Some respondents questioned why there was nothing new or innovative for older men, while others Forty-eight individuals have engaged with the mentoring service, with four people completing it. Groups are scheduled to commence for the Right Track and TIGER programmes in January, but no interventions are currently scheduled for Building Bridges. One referral has been accepted for the female personalised budgets. A total of 102 referrals were made to the mental health pathway in the first 6 months. Forty- five service users have been assessed by the Forensic Mental Health Practitioners and 24 have commenced direct interventions. A total of 305 case consultations have been undertaken by the FMHPs. The employment pathway has received 70 referrals, with only two being accepted. A total of 110 intelligence development plans have been received by IOM teams, and 320 offender-based taskings carried out, almost exclusively in the boroughs of Hackney, Haringey and Waltham Forest. Feedback from staff on the design and delivery of the enhanced services will be reported in the next section. 17 thought the services were not accessible to service users and failed to take into account the nature and motivation levels of prolific offenders. Relating to the design of the pilot, a number of respondents expressed confusion about how GtO fits in with existing IOM schemes, how it is different to IOM, and what added value it brings over and above the services already being delivered by boroughs. Some staff questioned why GtO was created separately when it could have been implemented within IOM “which is a well-established and recognised multi-agency approach in targeting prolific offenders.��� Some staff also questioned whether the money would have been better spent by giving partner agencies more information and training on the current IOM scheme, and providing additional resources to IOM. Delivery of the pilot Practitioners were asked to give their opinion on the delivery of the enhanced services. Practitioners were most positive about the delivery of the PSR service and least positive about the delivery of the enhanced OM interventions for young adult males. The main themes about what is working well and what could be improved are presented below: - Practitioners were positive about the PSR service and found it helpful having a direct point of contact in the court. Their impression was that cases were seen quicker than usual and that the reports were of a higher quality. - The main criticism of the PSR service was the gap in service resulting from the vacancy at Highbury Corner Magistrates’ court. Other areas for improvement included ensuring the probation officers were contactable during court sessions (i.e., providing them with a mobile phone or tablet to access emails) and ensuring they were in communication with the IOM units to gather the relevant information before delivering the report. - The prison co-ordination service also received positive feedback. The co-ordinators were praised for being efficient and communicating well and staff reported that it was helpful having a direct link into the prisons. - Some respondents questioned whether the prison co-ordination service was value for money, having only used the service a couple of times. Others noted that the information provided by the service was information they would have been able to gather themselves, albeit over a longer time period. Practitioners also questioned whether the co-ordinators should be based within the prisons.13 - A number of concerns were raised by practitioners about the enhanced OM services available for young adult males. Practitioners reported difficulty progressing referrals and queried the length of time it was taking providers to act upon a referral. 13 Since this feedback was gathered, the prison-coordinators have moved to being based in prison. It is important to understand what the needs of the cohort are before putting together a range of different interventions or none of them could make an impact [CRC] It is challenging to implement a pan- London model as each borough set up is very different, so different areas need different things [Local Authority] 18 - Staff also reported that some of the enhanced services appeared non-existent and that there has been a lack of pro-activity by service providers. Practitioners also thought awareness among offender managers around what enhanced services are on offer could be improved. - The mental health pathway received positive feedback. The forensic mental health practitioners (FMHPs) were praised for their communication and availability and for offering much needed advice to offenders and partners. - There were some concerns regarding duplication of service; for example, some boroughs already having a surplus of mental health practitioners. - Suggestions for improvement to the mental health service all related to an expansion of the service, for example: having the FMHPs based in the borough for the whole week in order to reschedule missed appointments; having a FMHP for each borough rather than two across the eight boroughs; and for the FMHPs to deliver 1-to-1 interventions across all boroughs. - Only a few respondents took the opportunity to comment on the employment pathway and reported that, while the employment providers appear to be a good resource, they are currently absent from boroughs and could be more involved with referrals (e.g., by coming to meet offenders). - Some respondents also thought it would be helpful for the pathway to have a focus on training needs and providing assistance for individuals to travel in to work. - There was mixed feedback regarding the policing intelligence support function.14 Some staff reported that the process was simple, the turnaround quick and the product useful and appropriate. Others reported: the service is adding another layer of unnecessary supervision; intelligence support should be completed by IOM officers who know more about their cohort; there is risk of duplication of service; and it could be improved by providing an open source element. More generally, a number of respondents reported the effective delivery of the pilot had been affected by the split of probation following Transforming Rehabilitation. They noted the pilot was “operating in a time of unprecedented change for Probation” and cited a number of problems with the performance of the CRC, particularly regarding the “unmanageable” workloads of offender managers and high staff turnarounds. These high workloads mean offender managers do not have the capacity to deal with the added demands of the pilot, resulting in a lack of awareness of available resources and low referral numbers. There was also dissatisfaction among respondents with CRC senior level management, with some staff observing a disconnection between senior level discussions and what is being delivered on the frontline. 14 Since the 16th January 2017, the existing intelligence support function was changed to focus on enhancements to existing intelligence profiles rather than the creation of new profiles. The pilot has come at a time of great change within CRC, probation staff have unmanageable workloads and the added demands from the GtO pilot cannot be serviced adequately [Metropolitan Police] Offender managers do not know what GtO is, they feel it is another referral form they do not have time to complete [Local Authority] 19 Feedback from the service provider focus group highlighted a feeling among providers that the pilot was extremely slow to start due to a perceived lack of planning at the front end. They spoke of the frustration they felt with people (i.e., offender managers) not knowing their roles or what the pilot was, and very few referrals coming in during the first few months. They reflected that there have been recent improvements to referral processes, and praised MOPAC’s response to issues raised. Partnership working An increase in partnership working emerged as one of the main positives of the pilot so far. Almost two-thirds of survey respondents (65%; n=21/32) agreed that working on GtO has allowed them to develop relationships with new partners, and half (n=15/30) agreed that working on GtO has allowed them to improve relationships with existing partners (see Table 4). Furthermore, over half (n=17/33) of respondents agreed they understand better what other agencies do to manage prolific offenders as a result of the pilot. Comments from respondents suggested the pilot has increased contact between different partners (e.g., between prison and court services and practitioners), and has led to new partnerships being developed. Positively, almost two thirds (n=6/10) of service providers believed they were in a position to sequence interventions for offenders alongside other providers. Table 4. Survey responses to partnership working questions Strongly Disagree Disagree Neutral Agree Strongly Agree Working on GtO has allowed me to develop relationships with new partners (n=32) - 15.6% 9.4% 43.8% 21.3% Working on GtO has allowed me to improve relationships with existing partners (n=30) 3.3% 16.7% 30.0% 36.7% 13.3% I understand better what other agencies do to manage prolific offenders (n=33) 6.1% 27.3% 15.2% 39.4% 12.1% However, feedback also suggested improvements could be made to the ‘whole-systems approach’. For example, only one out of five service providers were positive about the communication between their organisation and those referring into their service and less than half of service providers (n=4/10) reported feeling involved with the other providers. A number of respondents suggested that joining up the provider and practitioner meetings could help to increase awareness of the services on offer and, as a result, increase the number of referrals. I believe that as a result of provider and other meetings we are beginning to understand how we might ensure that the approach we take compliments other referral pathways for the same service user [Service Provider] It was only through an invite to the providers’ forum did I appreciate the amount of resources available for offenders resulting in a sudden spike of referrals being made to GtO [Local Authority] 20 Training, awareness raising and communication Only 27% (n=6/22) of practitioners reported finding the training, awareness raising and communication received from MOPAC useful; 58% (n=11/19) found the operating model useful; and 59% (n=10/17) found the Local Implementation Group (LIG) meetings useful. The main feedback regarding improvements were: having a more simplified operating model highlighting key practices and removing ambiguity; linking the LIGs with the Providers’ Forum and inviting OMs and SPOs to the LIGs to increase awareness of the pilot and the services on offer; and making the information provided less repetitive and more specific to local areas. The same low proportion of practitioners (27%) found the training, awareness raising and communication received from their own organisation useful. Nevertheless, the majority of respondents (77%; n=17/22) felt they had received enough training to use GtO in their role. Positively, the majority of service providers and strategic staff surveyed found the Providers’ Forum and GtO Workshops useful, respectively. Potential for impact Of the aims and objectives of GtO, survey respondents were most confident the pilot had the potential to achieve: a reduction in reoffending (84%; n=37/44); to support offenders to tackle problems in their lives (84%; n=37/44); to improve partnership working across agencies (82%; n=36/44); and to improve the engagement of prolific offenders (80%; n=35/44). Respondents were less confident that GtO had the potential to improve public safety (50%; n=22/44), and to improve public confidence in the ability of the CJS to tackle prolific offenders (55%; n=24/44). However, despite agreeing the GtO pilot has the potential to achieve these aims and objectives, a number of respondents took the opportunity to state that the pilot has yet to transform this potential into actual results: In terms of the impact of GtO on ways of working, the majority of respondents (69%; n=23/33) agreed the pilot has the potential to impact their work in a positive way; however, only 35% (n=11/28) of respondents currently agree that GtO has had a positive impact on working with prolific offenders. Furthermore, 32% (n=9/28) of respondents reported the pilot has increased their workload beyond a comfortable level. Positive learning and challenges Overall, a number of positive aspects of the pilot were identified by survey respondents, including: having additional resources available to offenders through the enhanced service offer; developing stronger working relationships with partner agencies; the opportunity to add consistency to the GtO has a lot of ‘potential’, but I’d say at this time it is struggling to transform that potential into actual results [Metropolitan Police] Although GtO does indeed have the potential to achieve the above ticked answers, at the moment, it is definitely potential rather than actual from our perspective as a Service Provider [Service Provider] 21 IOM approach; and increasing awareness of the role of other partnership organisations. Furthermore, of the enhanced services available, the PSR service, prison co-ordination role, and mental health engagement were consistently identified as positive aspects of GtO. Despite these positives, there were a number of challenges identified by respondents, including: the pilot creating additional work without the resources to properly carry it out; disjointed implementation leading to confusion about some aspects of the pilot; too much administration and data gathering, including the continuous updating of the GtO cohort; lack of awareness about the services from OMs leading to low referral rates; lack of flexibility and some enhanced services thought to be inappropriate or impractical for this cohort; using out-of-date OGRS scores as the method of selection; and lack of adequate IT databases to accurately identify and track participants. Service providers spoke of their particular challenges in the focus group, including difficulties accessing some prisons and prisoners being moved before they have the opportunity to meet with them; problems getting service users onto courses due to locations and gang issues; poor communication between the providers; and lack of knowledge about the pilot from offender managers. Furthermore, a number of respondents were critical of the name of the pilot and thought ‘Gripping the Offender’ gave a negative perception of the pilot to partners and service users. It has been useful in bringing all parties together that work together to reduce reoffending [Mentoring provider] An opportunity to add some consistency to IOM approach across a wider area of London [Metropolitan Police] Being part of a London wide pilot programme directly linked to the Mayor’s office is a big positive for me [Local Authority] There is a feeling on occasion that some aspects are ‘box ticking’ or busy work to satisfy the terms of the bid rather than doing things we would feel might contribute to reducing reoffending [Metropolitan Police] 'Gripping The Offender' as a publicly used piece of language is inappropriate: it runs counter to the non-labelling and supportive environment that many Criminal Justice professionals support [CRC] 22 Key learning Findings from interviews and surveys with stakeholders revealed a number of key points for learning. Staff were generally positive about the rationale for the pilot, but expressed a number of concerns about how aspects of the pilot were designed. In particular, respondents questioned whether the pilot was overly ambitious and whether it should have been rolled out as an enhancement to local IOM schemes. Practitioners had mixed views about the delivery of certain aspects of the pilot, with the PSR service, prison co-ordination service and mental health pathway rated the most positively. Overall, delivery was thought to have been compromised by perceived challenges within probation and the poor performance of the CRC. The pilot has had a positive impact on partnership working, but more linking up and communication between practitioners and providers was suggested as an area for improvement. Training, awareness raising and communication was generally rated poorly by practitioners both in regards to MOPAC’s training and the training provided by respondents’ own organisations; however, most respondents agreed they had received enough training to deliver their role in GtO. Overall, most respondents agreed GtO has potential for impact but felt at this stage the pilot has not transformed this potential into actual results. 23 4. Discussion This interim report has provided an update on the progress of GtO and highlighted some key learning from stakeholders involved in the pilot. On the whole, performance analysis indicates that some enhanced services are progressing well in terms of numbers (e.g., the PSR service, mental health pathway); whereas for other services, referral numbers and numbers accessing the enhanced services are lower than anticipated. The increase in referrals into the enhanced OM services from September indicates positive steps have been taken to improve referral rates. This work needs to be sustained moving forward to meet the proposed volumes for the enhanced services and, more importantly, to make sure those individuals who would benefit from the services have the opportunity to receive them. A number of positive messages emerged from stakeholder feedback: - There was a great deal of support for the rationale for the pilot and respondents understood why the pilot was designed - Staff identified an increase in partnership working as one of the main benefits of the pilot so far, both in terms of creating new partnerships and improving relationships with existing partnerships - Staff were positive that the pilot had created additional resources for offenders through the enhanced service offer and noted the particular benefits of having a direct contact in the courts and prisons, and having ready access to forensic mental health practitioners - Most staff agreed that GtO has the potential to impact on a number of important outcomes, including reduced reoffending, helping offenders tackle problems in their lives, and improving partnership working Feedback from staff also suggested a number of challenges to the effective delivery of the pilot, including: ongoing issues with the CRC and unmanageable OM caseloads; the complexity of the pilot leading to a lack of understanding about all the services available; and resource problems in boroughs. The realities of creating a whole-systems approach involving a range of agencies was also apparent from the experiences of staff, and communication between practitioners and providers was identified as an area for improvement. Over the next six months, the evaluation will continue to gather the views of staff involved in the pilot through surveys, interviews and focus groups. The evaluation will also seek to capture the views of individuals receiving the enhanced services to get their experiences of the pilot and the perceived benefits of the enhanced services on themselves. The long-term aim for the evaluation is to robustly examine the impact of GtO on reoffending and a range of other outcomes; however, this will be dependent on the number of offenders who access the enhanced services, the quality of data received, and the identification of an appropriate comparison group. References i Integrated Offender Management Key Principles. Home Office Report. February 2015. ii Stewart, L. (2015). Effective interventions for women offenders: A Rapid Evidence Assessment. National Offender Management Service. iii PCC spotlight: Mental Health. http://www.revolving-doors.org.uk/documents/pcc-spotlight-mental-health/; http://www.youngminds.org.uk/assets/0000/9472/Barrow_Cadbury_Report.pdf; Beyond Youth Custody (2014) Developing trauma-Informed resettlement for young custody leavers, a practitioner’s guide. iv Visher, C. A., Winterfield, L., & Coggeshall, M. B. (2005). Ex-offender employment programs and recidivism: A meta-analysis. Journal of Experimental Criminology, 1, 295-316 1 Group Violence Intervention London: An Evaluation of the Shield Pilot December 2016 Tom Davies, Lynne Grossmith & Paul Dawson MOPAC Evidence and Insight 2 Executive Summary Background Gang, group and serious street orientated violence continue to be a significant problem in London, demanding innovative and collaborative solutions. In June 2014, at MOPAC’s Policing Global Cities: Gangs Summit, Professor David Kennedy delivered a keynote speech outlining the Group Violence Intervention (GVI) approach he developed in Boston during the 1990’s (known as ‘Ceasefire’). Professor Kennedy was subsequently invited by MPS Trident to deliver a two day ‘Ceasefire University’ in the GVI model. As a result, the MOPAC Evidence and Insight team, with input from MPS central intelligence, undertook analysis using crime and social demographic data to develop a comprehensive borough level picture on gang and youth violence. The analysis was used to identify potential pilot boroughs with whom MOPAC initially engaged in dialogue, and this was followed up with senior level meetings with Lambeth, Haringey, Westminster, Hackney and Newham between August and October of 2014. Agreement followed from these meetings to proceed with Lambeth, Haringey, and Westminster. In selecting the three boroughs MOPAC recognised the strength of their community safety partnerships and willingness to trial the Group Violence Intervention approach. To this end, Shield - an adaptation of the Group Violence Intervention (GVI) strategy - was developed as a pilot programme and rolled out in three boroughs (Lambeth, Westminster and Haringey). This report presents learning from the evaluation covering performance, process (i.e. implementation challenges and benefits), and impact. Performance Monitoring  A total of 19 gangs and 321 individuals were selected across the three boroughs, with one third (n=103) invited to at least one call-in.  Between June 2015 and March 2016, a total of eight gang call-ins (a key message delivery mechanism) were held, attended by a total of 27 individuals.  Over the Shield period, Collective Enforcement was triggered on three occasions, each following identification of a threshold offence meeting the stipulated criteria.  Between June 2015 and July 2016, 40 per cent of the full Shield cohort (n=129) were arrested (in total there were 300 arrests, at an average of 2.3 per arrested individual) and 115 individuals were charged with at least one offence (an average of 1.88 per charged individual, a total of 216 charges) within each borough’s live pilot period. Key Challenges  Stakeholders questioned whether a US approach is the most appropriate model for London, given the lower rates of serious violence, fluidity of gang structures and different legal mechanisms available (i.e. ability to compel call-in attendance).  Linked to this, ensuring ‘swift and certain’ enforcement action against gang cohorts created challenges in the context of ongoing (and often similar) police operational activity on each borough, particularly given police practice of not storing risk (sitting on actionable information re criminal activities) and lack of available civil sanctions. 3  Key implementation challenges included differences in interpretation of the core elements of the GVI model between some practitioners. Feedback suggested programme design would have benefited from earlier input from the National Network for Safer Communities (NNSC), and it was perceived that the operating model lacked clarity for practical application, contributing to delays in roll out and a lack of standardisation.  Initial engagement of local communities and gaining support for the pilot were perhaps the greatest challenges. The pilot highlighted general poor community relations across statutory agencies in all boroughs, as well as strong resistance from some to the pilot. Key Benefits  Over the course of the pilot, significant progress was made in engaging local communities, individuals, and voluntary groups with a shared desire to reduce violence and offer assistance and support to gang involved youths. Practitioners highlighted the benefits of using Shield as a conversation starter with previously disengaged groups, making positive links between the community, police and local authority.  Where delays in implementation did occur, the enhanced consultative process, and subsequent buy-in and involvement were thought to be positive.  Practitioners and community representatives all noted a positive shift in the levels of transparency and openness regarding the information that Local Authorities and Police held.  Shield prompted a reinvigorated and wider reaching mapping of the available local services for gang involved young people, across partner and voluntary sector agencies.  Despite challenges, all three boroughs have mainstreamed elements of Shield into their business-as-usual activity including governance structures; collective efficacy and the encouragement of gang ‘self-policing’; and more developed community engagement. Key Learning for Future Initiatives  Partnership Working: Although boroughs reported a positive impact beyond the Shield specific remit (particularly between police, probation and third sector organisations), feedback indicated that wider partnership buy-in would be necessary for any future iteration. Specifically, via enhanced civil enforcement opportunities, legal mechanisms to compel call-in attendance and fast tracking of cases through the CJS.  Communication: Improved messaging and a more sensitive approach to branding and community consultation may have been beneficial to the programme, such as more frequent sharing of information, best practice and learning between boroughs, with local communities and between and within agencies. A primary focus on specific types of group orientated violence (i.e. stabbings and shootings) rather than ‘the gang’ may have aided community buy-in and wider communication of the pilot’s core aim of violent reduction.  Programme Design: Inter-agency and community complexities have made it difficult to resource; communicate; and standardise a viable model. Implementation challenges can be anticipated and built into any future iteration. 4 Borough Specific Learning Across the three boroughs Shield was implemented on a phased basis. The following case studies present a brief narrative of the key implementation challenges and benefits specific to each borough. Full breakdowns of performance data and findings are documented in the main report. Lambeth volunteered to be the first borough to go live and initially faced significant local opposition to the implementation of Shield within parts of the community, resulting in some local resistance politically. Despite these challenges, as well as enacting change from an already established way of working, Lambeth’s partnership worked closely with organisations and individuals within the community to deliver Shield. Having volunteered to be the first live borough, Lambeth generated some key learning; for example, the adaptation from a voluntary borough wide call-in to a more focused approach utilising the compelling of probation supervised individuals. Lambeth delivered two rounds of collective enforcement. The activity immediately highlighted the differences in the law enforcement operational practice between the UK and USA. Notably for the pilot, the preference for dynamic risk management in the UK compared to the willingness to ‘store risk’ and hold onto actionable information until there is an act of serious violence in the US. This presented challenges in ensuring ‘swift and certain’ enforcement action. The biggest shift within the borough was seen within the governance structures. Although statutory partners within the borough had been working with community organisations and members, the governance structure required by Shield saw formal community representation at both strategic and operational levels within the borough. This has helped provide a more established bridge between statutory partners and the community and this set up is still in place. Westminster was the second borough to go live and took a slightly different approach. From the borough’s perspective, the distribution of in-scope gangs across the borough made a single gang call-in as stipulated by the National Network for Safer Communities (NNSC) over complicated and unnecessary. Although delaying ‘go-live’, the eventual agreement was to hold three separate call-ins in areas local to the gang. Despite the NNSC’s initial reluctance to approve of the interpretation of the GVI model in this way, their stance changed when Prof Kennedy visited London in December 2015, acknowledging the benefits of a GVI approach that adapts to the local conditions and incorporates frontline expertise. Westminster delivered one instance of collective enforcement, experiencing similar challenges to Lambeth. Positive community involvement and ongoing engagement by the statutory partners led to the borough holding a community driven call-in style event shortly after the formal end date of the pilot, attended by at risk individuals, families and community members. More so than the other boroughs, Westminster reported that Shield acted as a focal point for reviewing and refreshing the existing approach, ensuring that resources were targeted in the most effective manner. Haringey experienced the most initial community opposition to Shield but by the end had, perhaps, the most complete community engagement approach. Over the course of several months, with significant initial facilitation by Community Development Foundation, there was sustained engagement from statutory partners and MOPAC with community organisations and members, acknowledging some significant underlying tensions and issues. This led to the design of a model that focussed on support to exit the gang and a strong community voice underpinned by enforcement under the banner of Communities Against Violence (CAV). The borough held its first call-in early in March 2016. CAV continues to operate within the borough and it is hoped the positive engagement from all partners will continue. The early involvement of community members, alongside increased transparency in the development of the local model provides solid foundations for similar future initiatives and MOPAC would encourage other areas to replicate this approach. 5 Programme Impact  The pilot was set against significant MPS wide increases in violence against the person (VAP) offences between June 2015-May 2016 and the previous year; this was reflected across all three pilot boroughs.  Whilst all three boroughs recorded decreases in gang flagged violence (violence with injury only), the MPS also recorded a decrease over the comparison period.  Comparing the Shield cohort offending to that of matched control gangs over the live pilot periods, no significant difference could be detected across suspect/arrest/charge categories for serious violence or gang flagged violence. No difference could be detected across wider measures such as VAP or all offending. Overall, there was no clear indication that Shield has had any influence upon borough level violent offending.  Due to data limitations it was not possible to assess referrals and outcomes robustly. Discussion Shield was a pilot programme established to generate learning in tackling gang and related violence. It was not possible to demonstrate a significant reduction in violence across the targeted Shield groups - this does not indicate a GVI approach does not work or is not fit for London - rather that the challenges in implementation resulted in no clear test of the model. None-the-less, the pilot has generated genuine learning; notably demonstrating how communities and statutory partners can work together (both formally and informally). It is hoped that these lessons can be adopted to help reduce the risk and vulnerability of those high risk young people which will be beneficial when planning any similar future initiatives. 6 1. Introduction Context Gangi, group and serious street orientated violence continue to be significant problems in London, demanding innovative and collaborative solutions. The Metropolitan Police Service (MPS) recorded 233,928 violence against the person (VAP) offences between September 2015 and August 2016; an increase of 10 per cent on the previous year. Over the same period, there have been recent rises in proxy indicators of gang violence, such as stabbings (up four per cent) and firearms discharges (up 43 per cent).1 MPS data indicates there are currently around 182 active street gangs across the capital, with approximately 3,500 individuals identified as involved in violent gang related criminal activity, although it is recognised this is only an estimate of a complex picture. Previous research indicates these individuals are involved in a disproportionate amount of serious offending in London; accounting for 9% of all personal robbery, 16% of serious youth violence, and 40% of shootings.ii The need for innovative interventions on the pilot boroughs is supported by findings from the Public Attitude Survey (PAS).2 Concern about gangs in the local area is linked to lower levels of confidence in the police – only 49 per cent of those who consider gangs a problem report the police are doing a good job, compared to 72 per cent of those not reporting gangs as a problem. Respondents in Shield boroughs reported some of the highest levels of concern around gang problems in their local area (Lambeth, 18%; Haringey, 19%) and lower levels of confidence in policing than the MPS average (Haringey, 56%; Lambeth, 62%).iii An array of local and nationally led interventions have aimed to address the harm caused by groups/gangs. Since 2012, the Home Office Ending Gang and Youth Violence teamiv has worked with local areas to develop local strategies, whilst also driving national policy changes such as mandatory sentences for knife and gun crimes.v In London, MOPAC’s Strategic Ambitions on Gangs and Serious Youth Violence, governed by a multi-agency Gangs Panel, has generated a number of co-commissioned projects such as London Gang Exit and youth support through Redthread in major trauma centres. However, the evidence base for ‘what works’ in UK gang interventions is still lacking. The implementation of a Group Violence Intervention (GVI) model, funded by MOPAC in 2014, sought to formalise an intervention approach with previously proven success. 1 All data from MPS MetMIS; ‘Stabbing’ relates to knife crime with injury, with 160 more victims making a total of 3877 for period Sep15- Aug16, 160 more than the previous year. There were 302 firearms discharge incidents over the same period, up 92 on the previous year. 2 The Public Attitude Survey (PAS) is a pan-London continuous survey run by MOPAC. It is based on a random sample of respondents at pre- selected addresses, with a total of 3,200 Londoners interviewed face-to face each quarter, yielding an annual sample of 12,800 interviews. The survey asks a variety of questions around crime, justice and policing. 7 The GVI/Shield Strategy Pioneered in Boston in the late 1990’s as ‘Operation Ceasefire’, a problem-orientated policing approach to tackle gang violence; GVI has been developed and implemented in a number of US citiesvi as a multi-agency and community involved intervention for violence reduction.vii It is designed to focus criminal justice, social service and community attention on the small number of group or gang involved individuals who commit a disproportionate amount of serious violence. There are three core elements to the GVI strategy: 1. Community Voice – Mobilising local communities and key moral voices to re-enforce the message that the community wants gang involved individuals to be safe, alive and out of prison and that violence will not be tolerated. 2. Consequences of Violence – Future violence will be met with swift and certain action, with police and partners paying special attention to the entire group through available and proportionate legal and civil sanctions when a member commits a violent act. 3. Help for those who ask – providing individuals with a route out and the opportunity to exit from the criminal lifestyle. Fundamental to the approach is the communication of a unified message from police, partner agencies and community representatives that: the violence must stop; there will be swift and certain consequences across the entire group if it does not; and there is help available to those who wish to exit the gang. One of the mechanisms for delivering this is to ‘call-in’ influential gang members to speak to them directly. It is this element, designed to create collective accountability, and convey clear community norms against violence, that sets GVI apart from business as usual enforcement and intervention.viii GVI Evidence Base Multiple evaluations of GVI initiatives in the United States indicate significant successes, with reductions in homicide and non-fatal shootings of between 35-60%, although there is still a deficit of rigorous randomized evaluations.ix To date, there has been a limited application of the GVI strategy in the UK. In January 2005, Strathclyde Police (now Police Scotland) implemented a version of GVI in Glasgow, achieving significant success in decreasing weapon carrying and violence among gang involved youths, although reporting no significant reductions in physical violence.x The GVI approach was adapted to fit the locality; for example, call-ins (renamed self- referral sessions) were predominantly voluntarily attended following a long period of focused engagement.xi In London, the 2009 Pathways programmexii was based on GVI principles but encountered strong implementation challenges (i.e. a delayed start, one borough withdrawing and changes to the core model such as individual versus group call-ins). Aspects of the GVI/Ceasefire strategy informed the creation of Trident Gang Crime Command (TGCC), the revamped MPS response to gang violence launched in 2012. More recently, elements of the GVI approach have been implemented by several London boroughs, usually via an MPS led partnership approach, though none have received rigorous evaluation.xiii 8 About Shield Shield, the London adaptation of the Group Violence Intervention (GVI) strategyxiv began in 2014. The overarching aim of the Shield pilot was the reduction of ‘serious violence’3 by the identified gangs on the pilot boroughs; Lambeth, Westminster & Haringey. Potential secondary benefits included increases in individuals exiting the gang, uptake in interventions and greater community mobilisation and confidence. A Shield partnership agreement formalised the collective commitment of MOPAC and partners4 to the approach outlined in the Shield operating model (OM) and based on the core elements of GVI described above. Boroughs were selected through a combination of assessing local pre-existing police/local authority gang infrastructures and analysis of local crime and population data (see appendix). For the purposes of the evaluation, ‘go-live’ was determined by the first time the Shield message was delivered to the selected cohort. Ahead of any messaging and enforcement activity, boroughs were asked to secure community representatives willing to work with the Shield partners to facilitate a unified approach and have direct input in cohort selection. Boroughs were asked to notify selected individuals of their involvement in the pilot, either in person or via letters to their home addresses. The OM advocated inviting a small number of gang members from each gang to attend a call-in, receiving a clear and unified message from Local Authority, Community and Police representatives reinforcing the key elements; that the violence must stop; that there will be legal consequences if it does not and there is help available to those who wish to exit the gang. From this point, if a cohort member committed an act of serious violence from a pre-defined list (known as a ‘threshold’ or ‘trigger’ offence) ‘Collective Enforcement’5 would be initiated against the entire gang, using any available and proportionate police or civil enforcement options, for a period of two weeks. The process, from nominal selection onwards could be repeated as required over the live period. The pilot commissioned direct input and guidance from the US team who developed the GVI approach; the National Network for Safer Communities (NNSC). MOPAC also secured the services of the Community Development Foundation (CDF) to facilitate community engagement on all pilot boroughs. 3 Serious Violence, essentially ‘Most Serious Violence’; Murder and attempted murder; Grievous Bodily Harm – with or without intent; Weapon Enabled Crime (incl. stabbings, excluding knife possession); Violent Disorder, including Rioting; Possession of a Firearm offences; Kidnapping; False Imprisonment; Aggravated Burglary (all types). 4 Trident, local MPS, Probation services (National Probations Service (NPS), Community Rehabilitation Centre (CRC), Youth Justice Board (YJB), National Offender Management Service (NOMS) and the Crown Prosecution Service (CPS) 5 ‘Collective Enforcement’ is MOPAC terminology for the focussed enforcement response initiated against all members of a gang when one individual commits a threshold offence whilst Shield is live. It is designed to ensure the gang will be held ‘collectively accountable for violence committed by a fellow member, countering the group dynamic that drives violence’ (NNSC, 2013). The report discusses the possible confusions around Shield’s enforcement operational practice linked to the use of this terminology. 9 2. Evaluation Methodology The Evidence and Insight team at the Mayor’s Office for Policing And Crime (MOPAC) conducted the Shield evaluation. A holistic plan was designed, with the intention of measuring the process learning and any potential impact Shield had on its aims – foremost but not limited to, the reduction in serious violence committed by the selected cohort and gangs. Additional outcome measures (which were dependent on data capture) included intervention uptake, message delivery and call-in attendance. Whilst not quantifiable, community related benefits are discussed in process learning. In terms of process learning the evaluation utilised a mixed methodology including staff surveys (n=53, over two phases); interviews with key practitioners (n=18); ethnography at community meetings/call-ins; borough reports and assessment of performance data via the Shield Activity Tracker (SAT), an E&I developed data capture tool for practitioners. It was decided to explore potential impact via a matched control sample – that is, comparing Shield nominals to similar individuals on non-pilot boroughs on key outcomes (i.e. suspected; arrested; charged). This was selected as the most robust feasible methodology, although it limits the weight of causal statements in comparison to a Randomised Control Trial (RCT). A version of propensity score matching was used on a borough and gang level, matching offending prevalence, specialisms, gang demographics as well as some individual elements to ensure the validity of the match (see appendix for more detailed breakdown). Pre/Post analysis of more general offending rates was used to supplement findings. 10 3. Results The report is focussed upon four areas of learning from the Shield programme - these are learning from performance monitoring, the challenges of implementation, wider lessons about programme process and the impact upon offending. Performance Monitoring Performance data was collected via borough reports and the Shield Activity Tracker (SAT) in an attempt to understand who was selected to be part of the Shield cohort; what needs they had; what offers of help were provided and what enforcement activity occurred. Results from the SAT indicate:  A total of 19 gangs and 321 individuals were selected across the three boroughs: o Lambeth - 8 gangs, 132 individuals. o Westminster - 6 gangs, 90 individuals. o Haringey - 5 gangs, 99 individuals.  Individuals selected were exclusively male.  The average age of the total cohort was 21.5 years, with a minimum age of 13 (Lambeth) and a maximum age of 44 (Haringey). 6 Lambeth had the youngest cohort (average age 20.1), similar to Westminster (average age 20.4). Haringey had an older average age of 24, with a greater proportion of individuals in the 25-34 bracket (31%, n=31).  In terms of ethnicity, the overall cohort was 85% (n=272) Black African-Caribbean and 94% (n=300) Black and Minority Ethnic (BAME). Haringey had the greatest proportion of African-Caribbean individuals (96%, n=95). Westminster had the most diverse cohort with one quarter Arabic/North African (28%, n=25).7 Cohort ethnicity was broadly reflective of the demographics of specific gang affected areas within the pilot boroughs, and in line with that on the MPS Gangs Matrix.8 Figure 1 provides further breakdown of the selected cohorts. 6 Under 18’s were included in the Shield cohort (14.5%) but were not part of the same call-ins as adults. 7 Information on ethnicity should be treated with caution, as it is based on police recorded data rather than self-classified. 8 The MPS gang nominal database; the Matrix tracks and scores individual gang members based on frequency and severity of violent offending & victimisation. In November 2015, the ethnicity of Matrix nominals was 88% BAME and 77% Black African-Caribbean. 11 Shield Cohort Engagement & Offers of Help Overall, eight call-ins were held over five days during the pilot, three on Lambeth in the ten months they were live; four on Westminster in seven months; and one on Haringey in the month they were live. This resulted in eight separate call-ins and 27 attendees (of 103 invited) over the entire period. Given the low attendance levels it is not appropriate to detail the exact numbers involved.  For Lambeth’s first call-in attempt (early June 2015), gang members were invited to attend voluntarily. The event was on a large scale as advocated by NNSC, catering for around 50 invitees. The event did not attract sufficient participants and was subsequently cancelled. Of those that did attend, three quarters were compelled by Youth Offending Services (YOS), providing early signposting that compulsion may be the best route to securing attendance. Key reasons were perceived to be an unwillingness to lose face in front of rival gang members; an intimidating environment with a large numbers of service providers and observers; the time (midday) and location (Town Hall) of the meeting. Lessons learnt meant the second call-in (October 2015) was more successful, with emphasis on compulsion and a change in venue (held at both a community centre and council offices) resulting in seven individuals attending from a total of 18 invited9 across five gangs.  In late September 2015 Westminster held three call-ins in one day across the borough. A total of 22 individuals were invited; the vast majority non-compellable. Although no 9 From Youth Offending Service (YOS), Community Rehabilitation Centre (CRC) and National Probation Service (NPS). Figure 1: Shield Cohort Overview 12 one attended the first meeting, a total of seven individuals attended the latter two. Eight cohort members subsequently received a Shield message through custom notifications or other means, resulting in several individuals requesting help straight after their call-ins or highlighting willingness to engage. Westminster’s fourth call-in, held in early March 2016 was attended by a small number of members of a single gang, all compelled.  In the early stages of Shield, Haringey made a commitment to not go ahead without community approval. Implementation was subsequently postponed as the borough and local police worked with community members to design an acceptable local interpretation of GVI. This approach necessitated a substantial delay to ‘go live’; Haringey held their first call-in in early March 2016 (less than a month before the pilot end) in a school assembly hall in the north of the borough with a total of five individuals (all compelled) attended with their probation officers. This was held under the banner of ‘Communities Against Violence’ (CAV). Post cohort selection, all boroughs initiated a message delivery stage involving all members being notified of their participation in the pilot by letter detailing what this meant for them and what they could do if they wanted help exiting the gang. The methods of message delivery differed between boroughs; for example, in who delivered the message (a combination of local authority, youth workers, local police) and in what frequency. Custom Notificationsxv were also delivered to some individuals invited to the call-ins who did not attend.10 Although the SAT was provided to capture data on the message delivery, due to local resourcing issues, it was not possible to robustly document the message delivery component across the whole of Shield (at least 119 received written or verbal notification, the nearest robust metric being the 103 who were invited to attend a call-in). In addition, a key aim of engagement with the cohort was to provide support in exiting gang life via targeted interventions. However, it appeared that a number of cohort members were already involved with service providers prior to ‘go-live’ and all boroughs reported difficulties gaining new engagements. This was expected, and reflected in MOPAC’s decision, consistent with NNSC advice, not to provide additional funds for the help strand (all three pilot borough Integrated Gang Units are funded by MOPAC). All three boroughs provided a phone number for Shield nominals to call to seek advice and assistance in exiting the gang. Boroughs reported no uptake in referrals through this avenue, with some reporting abusive or threatening calls regarding the programme. Data on referral uptake and outcome was limited. From the available data, Lambeth documented 53 individuals engaging with one or more agencies/schemes, generating a total of 86 separate referrals (June 2015–December 2015). In the views of staff, generally those already being engaged with continued to do so, likewise those who did not. In April 2016, Haringey reported 27 individuals being worked with; again, it is likely these would have been engaged with regardless of Shield. 10 Custom Notifications are a form of individualised message delivery adapted to suit specific circumstances. 13 Shield Enforcement Activity Over the course of Shield, an uplift in enforcement activity was noted across pilot boroughs. However, the data received by the evaluation was not able to distinguish Shield enforcement as compared to business as usual policing. For example, Lambeth were assigned a Trident unit for the duration of the pilot, who carried out numerous operations, sometimes on Shield involved gangs and sometimes not. The backdrop of continuing rises in youth violence, knife and gun offences also prompted a proactive drive to focus on Matrix individuals, overlapping with the aims of Shield. 11 Between June 2015 and July 2016, 40 per cent of the full Shield cohort (n=129) were arrested at least once. In total there were 300 arrests (max=7, average of 2.3 per arrested individual) and 115 individuals were charged with at least one offence, totalling 216 charges (max=6, average of 1.88 per charged individual while Shield was live in each borough. The majority of charges were for drugs related offences (44%, n=94, mainly possession). 29% (n=63) of the total charges of cohort members between June 2015 and July 2016 related to violence against the person offences. Collective Enforcement (CE) was initiated three times across the pilot boroughs, twice on Lambeth (against two different gangs) and once on Westminster. On two occasions (once on Westminster and once on Lambeth) individuals who had attended the call-in went on to commit the threshold offence. When CE did occur, outcomes were often unsatisfactory; the first occurrence on Lambeth resulted in a number of arrests but all individuals were subsequently released with ‘no further action’. No threshold offences were identified following Westminster’s second call-in and Haringey’s first (both held in early March 2016). 11 Specifically, Operation Teal; a Trident led MPS operation initiated in June 2015 which used overt and covert tactics to prevent and disrupt knife and gang crime. 14 The challenges of Implementation This section documents the main issues to emerge from the implementation of the pilot. Effective implementation is an essential aspect of designing and setting up new programmes. By implementation the report is referring to the process of developing, designing and conducting the programme in question - issues such as learning from the evidence base, having a clear rationale, documenting operating models and guidance for staff, provision of training, effective involvement of stakeholders, data capture and so on. This is important as programmes that are implemented effectively are evidenced to get better results.xvi Whilst Shield has many aspects that are required (i.e. operation model, based on evidence) ultimately issues were encountered that hampered implementation. However, many of these were more keenly felt at the beginning of the pilot, with improvement over time; any future iteration should anticipate this and adopt contingency plans. Translation of GVI to London Although survey respondents agreed there is a significant gang problem to be addressed, on reflection from the pilot, there were mixed views as to whether a GVI based model can practically work in London. In particular, there was uncertainty about adapting a US model designed to tackle homicides and shootings, to focus on the lower-level violence more prevalent in London. Some staff put forward the differences in available legal mechanisms as a key barrier to successful implementation. It was not popular amongst some police officers, being viewed as either enforcement ‘watered down to nothing’ or conflicting with a key tenet of UK police practice to not store risk – holding off arresting individuals until Collective Enforcement was initiated. ‘There were issues around banking offences; police in the States are more comfortable storing risk’ The differences in structure and organisation of gangs/groups between the US and UK was also raised as a potential barrier to effective implementation, with a suggestion – to some extent supported by literature – that street gangs in London are far more fluid, less hierarchical and organised than in the states.xvii This may have impacted on message dissemination between gang members and potentially serves to illustrate early differences of opinion towards the scheme. More generally, some feedback suggested the decision to use ‘gang’ terminology rather than the NNSC advocated ‘group’ approach may have influenced negative perceptions of the pilot amongst some community members. 15 Initial Planning, Set Up & Design Operating Model Design and Communication At a senior strategic level there appeared to be a good understanding of the key principles of the GVI approach; senior leaders from MOPAC, MPS (Trident) and Partners had been aware of the overarching strategy for a number of years and elements of the three borough’s Integrated Gang Units (IGUs) were already based on GVI principles.12 However, feedback over the course of the pilot indicated some confusion as to the core driver of the model, with some believing law enforcement was the fundamental aspect (e.g. Trident), and others (e.g. Local Authorities) placing more focus on a collaborative community response, illustrating a crucial difference in approach. There was satisfaction with the (limited) general training required for Shield; although survey and interview feedback indicated the communication of the programmes strategic aims were not always clear.13 ‘There was a significant gap between the theory and the practical implementation’ Whilst it is positive an operating model (OM) was completed, many staff thought it did not provide an adequate blueprint, of either the fundamentals of the approach (e.g. a solid understanding of Collective Enforcement) or a practical, workable framework within which to implement it (e.g. building in ‘what if’ examples and carrying out dry runs of the key elements of the approach to ensure staff had clarity).14 To illustrate, due to a perceived lack of clarity in the operating model there was confusion as to what constituted a ‘threshold’ offence; it took two weeks from the first threshold offence to agree CE should be initiated, meaning the ‘certainty’ of swift consequences to the group was diminished. Opinion was split amongst practitioners as to the level of evidence required to initiate CE. More positively, wider Shield governance structures were well received, particularly at local level, with one borough explicitly stating that they would continue with the community inclusive structure implemented during Shield. Doubts were expressed as to whether the approach was adequately disseminated to front line practitioners. This included the need for more frequent sharing of information, best practice and learning between boroughs; and between and within agencies. In the absence of central Shield communication it was thought fuller updates on how the pilot was progressing would have been beneficial to maintain momentum of key practitioners. Borough Selection MOPAC underwent a process of engagement with potential Shield boroughs, meeting with all Chief Executives to establish willingness to participate. Boroughs were then selected to take part in the pilot from a combination of an assessment of local pre-existing police/local authority gang unit infrastructures15 and analysis of local crime and population data, to ensure sufficient prevalence of preselected ‘threshold offences’ to measure potential impact. However, the extent to which services were integrated between agencies or attuned to local communities 12 MOPAC Gangs Summit took place in June 2014. 13 Just over half of respondents indicated training was effective (n=17). 14 Call-in rehearsals were completed by all boroughs. 15 As carried out by the MOPAC Project Team. 16 varied, which impacted on external engagement. Some interviewees indicated these pre- existing complexities coupled with community distrust led to implementation delays. Local resourcing and capacity issues were also highlighted as risks from an early stage, particularly in terms of data collection, and remain significant barriers to long-term sustainability. Several interviewees suggested implementation may have been easier in boroughs without pre-existing integrated gang units. NNSC Input In addition, the delayed contractual agreement with the US-based National Networks for Safer Communities (NNSC), was cited by some as a barrier to planning and design, meaning much development was completed before NNSC were in a position to advise. For example, the need for lead-in analysis (such as local gang incident audits) was not identified until too near ‘go live’. This stemmed from an assumption that pre-existing mechanisms (such as the Matrix and MPS gang incident tracking) would be sufficient in terms of preparatory analysis, whereas in hindsight full gang and incident audits were required. Once involved, NNSC input was generally seen as helpful, particularly the on-site advice and guidance by the London based NNSC representative. However, difficulties were again cited around data sharing and the translation of UK specific legal, policing and social service mechanisms. ‘NNSC provided conflicting messages, changing to meet the current circumstances.’ Whilst there was some initial reluctance by boroughs to take part in teleconference calls with the NNSC, the visit of Professor Kennedy in late 2015 was viewed as extremely useful, with boroughs commenting on the benefits of his advice and reassurance. Boroughs did indicate they would have preferred a visit earlier in the pilot, likewise video conference calls; call-in examples; and US site visits could have been carried out at an earlier date to help conceptualise the approach for key practitioners. It was also suggested it may have been beneficial to hire Professor Kennedy/NNSC as full time advisors rather than merely an advisory capacity, as some practitioners observed the directives in the GVI handbook and MOPAC operating model were not as stringent as initially advised. For example, Westminster’s local adaptations, such as multiple localised call-in venues, were initially questioned by NNSC but praised during the site visits as working well operationally. Communication of Pilot Aims to Cohort Further doubts regarding the pilot design and practitioner uncertainty affected the practical application of the approach. For example, the confusion surrounding how to communicate pilot aims meant many of the cohort received multiple letters from Shield (notification, call-in) and other MPS operations with similar objectives.16 This was thought to dilute Shield specific messaging. In addition, the lack of follow up to home visits and message delivery meant front line practitioners were sceptical as to whether cohort members would even tell their peers about the pilot and specifically the consequences for not stopping the violence. Indicative feedback from youth workers suggests there were some discussions between local youths around Shield. 16 Letters for Brixton Splash street festival, Notting Hill Carnival, Operation Teal (Trident) targeting suspected habitual knife carriers. 17 Senior Policy Support Staff fed back that more top level support would have been valuable, especially in publically challenging some of the initial negativity and controversies (e.g. conflation with Joint Enterprise).17 Several respondents indicated MOPAC could have done more to challenge MPS operational actions, which directly impacted on the pilot areas and replicated many of Shield’s aims and processes. As a result there were mixed views concerning the degree of influence partners had, which some suggested made management and standardisation of approach more difficult. More clarity in the operating model would have been beneficial, to encourage a more standardised implementation approach. Standardisation across Boroughs The Shield pilot ran into delays and challenges regarding standardisation. Although a staggered roll out was expected, it was not anticipated to be over such an extended time period. Lambeth delivered the Shield message to their original cohort in early June 2015, followed by Westminster in late September 2015 and Haringey did not hold their first call-in (under the banner of ‘Communities Against Violence’ (CAV)) until early March 2016, less than a month before the official pilot end. Cohort and Gang Selection Each borough interpreted and implemented the Shield process differently, beginning with approaches to the initial selection of the Shield cohort. Cohorts were selected via a collaborative process between borough and central MPS intelligence. Involvement of the community differed heavily between boroughs, ranging from informal discussions with local youth workers (Lambeth) to official representation at intelligence meetings (Haringey). Although this provided opportunities to incorporate community intelligence and information from other agencies to inform selection, only a small number of individuals were selected who had not previously featured on the MPS gangs Matrix. Throughout the pilot there was consistent consensus from survey respondents and interviewees that the right gangs and individuals were selected by stakeholders. Haringey selected a notably older cohort (maximum age 44) and opted to include some gang members who may better be described as involved in organised crime. There were several other practical variations, including the number of gangs to include and differences in borough understanding around the need for cohort refreshes and nominal de-selection. ‘…robust enforcement has to be the lead-in the early stages of Shield to set the stall out…’ Call-ins A great deal of effort, resource and importance was placed on the gang call-ins, but in practice there were significant challenges, notably; securing attendance of invited gang nominal; the lack of appetite and inadequate mechanisms for compelling individuals to attend; and 17 Joint Enterprise is a common law doctrine that allows for several people to be charged with the same offence; even though they may have played very different roles in the crime e.g. persons who did not strike the fatal blow can still be convicted of murder. A Supreme Court decision in February 2016 found that the application of the law had taken “a wrong turn” and been misinterpreted for 30 years. 18 difficulties facilitating neutrally located multi-gang events. Although the call-ins never intended for the entire gang to be present, but instead were to feature key individuals who would relay the Shield message back to the wider gang, the low numbers of attendees was in contrast to meetings held in the US or Glasgow. This can be explained by the lack of sufficiently robust legal mechanisms to compel attendance compared with the US, and the far shorter lead-in and engagement period with cohort members compared to Glasgow. However, there were still mixed views as to the importance of call-ins to the overall success of the model – with some viewing them as fundamental and others considering them as just one option in a suite used to communicate the key Shield messages. ‘There was too much emphasis on call-ins; they shouldn’t be the focus, just a method of [message] delivery’ Although there was less cohort attendance at the call-ins than initially hoped (including less cohort members who could realistically be compelled), there was learning from the process – particularly the need to obtain the right people to deliver the message. Preparation (including rehearsals) was thought to be paramount – including understanding the audience; model; message; and nature and tone of the event, rather than reading from a script. Whilst much consideration was given to securing appropriate venues (youth centres appeared to work particularly well), learning indicated this was of lesser importance than tight message delivery. All boroughs put in a great deal of work to secure suitable speakers, either from the local community or with direct experience of gang violence or lifestyles. Practitioners felt the extended lead-in and amended local design ensured those willing to speak were both known and respected by the vast majority of young people involved in the pilot; and therefore able to better connect with the youths in question. Haringey particularly emphasised the benefits of informal mentoring as part of the pilot. ‘Success relies upon having strong community individuals who will stand united alongside agencies to set expectations and standards...identifying these individuals is difficult.’ The balance between Community Voice and enforcement was raised, with some practitioners highlighting the sheer number of visible police (and observers) at several of the call-ins as unhelpful. It was felt police input into call-ins could have at times been better planned, with some police speakers perhaps under prepared, although this varied between boroughs. The use of both Trident and local officers for message delivery at one call-in was seen as overkill. Officers with local ties to the estates and areas where the gang members lived may have been able to better engage attendees. The Extent of Preparatory Engagement The belief that gaining buy-in from the majority of the community was essential to successful implementation was queried by Professor Kennedy and the NNSC, who emphasised the approach was fundamentally law enforcement based and that community consensus is not necessary to start GVI. The NNSC advocated a non-consultative approach; gaining key representatives buy in, rather than seeking support of whole communities before launching. This interpretation of the operating model was a key marker in the differing Shield approaches across the three boroughs, with Local Authorities leading on community engagement and 19 playing down the enforcement element. Several practitioners suggested the enforcement aspect could have been handled with greater sensitivity, with one borough’s large scale community meetings creating a negative atmosphere pre-launch. Challenges of Community Engagement From the beginning of the pilot, all boroughs recognised that although the most challenging element of Shield, meaningful community collaboration and input would be crucial to achieving success. Early borough feedback suggested they would have preferred a longer lead-in time for community engagement prior to launch. It was considered by some front-line practitioners that Shield fuelled pre-existing perceptions of enforcement heavy policing amongst community groups, with existing levels of mistrust of the police, local authority and statutory bodies brought into a sharper focus. Two boroughs mentioned their well-known and entrenched issues relating to police-community relations, with some community members making it clear in meetings they believed the policing actions contributed to local gang issues rather than being part of a potential solution. Some local practitioners felt high profile enforcement would only serve to exacerbate such feelings. ‘…it is getting the communities on side that is the real challenge and would be the biggest benefit…’ Media Communications It was felt by a majority of interviewees across roles that media attention surrounding Shield was detrimental to planning and amplified tensions surrounding community engagement – in particular the Mayoral press release in January 2015.xviii At this time the press lines indicated a ‘tough’ enforcement heavy approach, where individuals would be ‘collectively punished for the criminal actions of individual members’. This was viewed by many respondents as a misrepresentation of the pilot aims and process, appearing to antagonise many of the communities Shield was trying to engage with. ‘[The Mayoral Press release] showed a lack of understanding of local dynamics; a case of five steps forward and ten steps back in gaining community support’ Practitioners thought initial messaging would have benefited from greater sensitivity and consideration of the impact it would have on a highly charged political arena, where race relations and policing are significant issues. Practitioners reported that anti-Shield sentiment spread rapidly across community networks and between pilot boroughs, in turn leading to an equally unhelpful conflation of ‘Collective Enforcement’ with Joint Enterprise. One borough described the publicity surrounding the pilot as a ‘double edged sword’ with the good work of the local gang unit highlighted, but at the same time hindered by the negative publicity. All three boroughs indicated a substantial proportion of their pre-existing community gang/street workers were unwilling to be associated with Shield in any way, fearing it could undo positive engagement due the perceived toxicity of the Shield brand. ‘Gang unit [outreach] work was tainted through association with the bad publicity surrounding Shield.’ 20 Senior Leadership Engagement In addition, some interviewees highlighted the perceived lack of top level support as contributing to community reservations. Several respondents thought there was insufficient engagement with local communities or availability from senior leaders at both MOPAC and the MPS, resulting in a failure to persuade local people the pilot was a key priority. Once live, communities reported little feedback as to how the pilot was progressing, although low attendance of Shield meetings or failure to cascade information received may have contributed to this perception. Sharing of up-to-date information on all aspects – particularly enforcement – may have avoided unhelpful rumours regarding police activity, whilst reassuring the community that individuals were not being unfairly targeted. Borough Specific Challenges to Engagement As the first borough to go live, Lambeth experienced its own unique challenges, with a difficult lead-in period and large scale public meetings resulting in some local political resistance. Feedback from interviewees and programme boards indicated engagement may have been aided by a localised focus and more informal, targeted pre-pilot discussions. Nevertheless, Lambeth reported benefits in using Shield as a conversation starter with previously disengaged groups and individuals, making fruitful links between the community and local police. Westminster viewed the pilot as an opportunity to refine and refresh their external messaging to the community, eventually reaching out to several hundred representatives across the borough, inviting feedback and suggestions.18 They found significant variations in willingness of communities to engage, with no apparent correlation to the areas most affected by gang violence. ‘[Shield] got different groups talking; many doors have been opened’ The last borough to go live, Haringey involved local communities in the redesign and implementation of the pilot to the greatest extent. This necessitated a thorough reappraisal of the original model and a re-branding to ‘Communities Against Violence’ (CAV) to quell the negativity surrounding Shield. Although the approach dictated a substantial delay to ‘go-live’, the enhanced consultative process, and subsequent buy in and involvement were thought to be positive. Process Transparency Practitioners and community representatives all noted a positive shift in the levels of transparency and openness regarding the information that Local Authorities and Police held. This included involving a small number of community members in formal processes such as the identification of gang members and assessment of harm, risk and vulnerability. Across the boroughs community representatives with appropriate clearance were invited to take part in meetings relating to the cohort; were privy to reasoning behind their inclusion; and given the opportunity to disagree with submissions or suggest alternatives. Practitioners suggested that although community intelligence of this kind was not forthcoming, the pilot encouraged greater participation from community representatives and partner agencies. 18 Total recipients, response rate and feedback collation is on-going locally. 21 ‘Allowing [designated] community members to see the Matrix enabled difficult conversations about the demographics of the cohort and wider links to organised crime; an honest approach is beneficial’ It was felt there was an opportunity to demystify what were seen as opaque processes regarding gang lists and in particular the Matrix, by involving community representatives and giving reasoning behind decision making. The enhanced communication and emphasis on two-way information flows were considered by key gang unit practitioners to be positive developments which should be built into subsequent programmes. Community Voice The pilot raised fundamental questions regarding the nature and meaning of community engagement. Whilst there was consensus engagement should be an inclusive process, reaching beyond those who had previous dialogue, there was less certainty as to whether the most suitable and/or representative community members were involved. In addition, there was a difference in the level of engagement and interaction, for example between those willing to debate at meetings and those willing to stand next to police and partners and support a consistent message at call-ins. Community speakers were generally very well received by boroughs, both in terms of delivery and perceived cohort response. ‘Boroughs went out of their way to find excellent community speakers but didn’t always do the same for the other speakers’ There were initial suggestions the pilot was a missed opportunity for the police to build bridges with local communities, by acknowledging they have not always got it right in the past - an approach found to be successful in previous GVI interventions in the United States. Despite this, as the pilot progressed, a strong shared desire to work together to reduce gang violence became evident at meetings. Practitioners felt the positive debate between diverse local community groups; service providers; and key stakeholders ultimately increased co-operation, understanding and information sharing. ‘It is crucial to engage all community members, not just those that are supportive of the pilot’ 22 Wider Process Learning This section highlights process learning over the course of the pilot. Partnership and Multi Agency Working The benefits of a collaborative approach enhanced by the pilot were acknowledged, with boroughs reporting a positive impact beyond the Shield specific remit, particularly with the police; probation; and third sector organisations. Probation services were generally seen as very supportive of the pilot, with high levels of engagement and cooperation despite the considerable resourcing issues. Police co-operation with Local Authorities and partners was strongly praised, with boroughs citing Shield as beneficial in cementing already good working relationships, particularly within gang units. Intelligence Flow At the beginning of the pilot the increased quality and availability of analytical product (such as Social Network Analysis), was highlighted with a general feeling the police developed a greater appreciation of the intelligence input from Local Authorities and partners. On a wider level, the pilot also highlighted divergence in understanding between local and central police units regarding gang tensions and individual motivations for violence - with one borough challenging the centrally produced intelligence picture. Although practitioners subsequently noted a decline in centralised resource as the pilot progressed, increased partner intelligence was indicated as learning to incorporate into future process. ‘…the tracking and analysis opened our eyes as to how fluid gang structure and membership has become’ Service Mapping Feedback suggests Shield also prompted a better attempt at mapping the available local services for gang involved young people, across partner and voluntary sector agencies. For example, Haringey provided the local community with an oversight of existing resources and gaps in service provision, to assist when planning new initiatives. CDF were instrumental in facilitating this, negotiating between interest groups, individuals and organisations with conflicting agendas. Their work identified which individuals and groups had a shared purpose of reducing violence and assisted them to better understand how their work can contribute to the bigger picture. ‘[By the end of the pilot] local partner intelligence was better reflected’ Partnership Buy-In However, working across multiple organisations also ran into difficulties, particularly around the perceived commitment from all partners. As the pilot progressed, fewer survey respondents agreed there was an equal commitment from all partners for Shield to succeed or felt confident 23 colleagues were reinforcing the Shield message and several interviewees highlighted lack of Crown Prosecution Service (CPS) involvement as a key barrier (e.g. fast-tracking offenders).19 Although it was recognised CPS participation in Shield was to some degree restricted by UK legal framework, all boroughs indicated this would be a key issue to address moving forward. All boroughs found identifying and actioning civil sanctions against cohort members was impractical, with gaining access to civil routes convoluted and time consuming (DWP, TV licensing, parking fines etc.). Other complications included the timing and scope of demonstrative enforcement operationsxix, completed before the scheduled call-ins resulting in a number of arrests of Shield cohort members ahead of the planned engagement. ‘[gang related] casework is one thing but the public visibility from CPS participation in Shield quite another’ Co-ordinating Interventions across Partners Although the ‘help’ strand of Shield was thought to be essential for long-term success, and boroughs all had strong diversionary focus, practitioners highlighted a number of practical issues which relied on strong partnership working to make the approach work. Some practitioners felt that expectations, particularly amongst the Voluntary and Community Sector (VCS), may have been unduly high. Others questioned why in contrast to the enhanced resource on the enforcement side, there was not a similar enhancement for interventions or referrals. To tackle this, partnership working was seen as paramount, with emphasis placed on how a variety of agencies can incorporate bespoke gang exit services within their standard approach. 19 48% (n=10) and 43%, (n=10) respectively. 24 Learning to Inform Future Practice All three pilot boroughs reported that they have mainstreamed elements of Shield into their activity:  Lambeth - Under the ‘Tackling Group Offending’ (TGO) banner, Lambeth indicated it will employ the governance structure implemented during Shield moving forward, seeking to re-introduce their gangs case management process (G-MAP) whilst maintaining the more dynamic identification process with greater input from community and partners. Mediated contact and targeted support for gang members through the suite of engagement options will continue, with the possibility of further call-ins but at a local and singular gang level. The emphasis on community representation and input in youth violence intervention processes is also being maintained and adapted to emphasize local area input organisation e.g. ‘Big Local’ Brixton youth violence community ‘think tank’.  Westminster - The focused messaging approach that Westminster partners – council, police, probation, housing, the community – finessed over the course of the pilot highlights the desire to have gang involved youth ‘safe, alive and out of prison’ will continue. Westminster reports the engagement processes driven by Shield has provided a strong groundwork for future development, driving plans to set up a serious youth violence advisory group. Although non-committal about future call-ins, the Shield collective self-policing approach was admired as an innovative tool and there are plans to incorporate elements moving forward. Interviewees reported Shield presented a new opportunity to approach the gang as a whole, encouraging self-regulation which fits with their underlying philosophy of ‘Your Choice’. Support offers will not change and are described as already strong.  Haringey – Haringey’s more community focused approach highlights the need for local sensitivity, with the borough emphasising the need for a holistic approach to tackling gang and group related criminality going forward. This will focus not just on violence, but on other key issues such as vulnerability and safeguarding related to CSE and County Lines. Communities Against Violence meetings are scheduled for the coming months.  Trident – Feedback suggests some members of Trident felt because Shield’s community work was led by Local Authorities, it presented local and central police units with less opportunity to build their own bridges. Therefore, Trident has taken Shield learning to implement a critical friends group on an inner London borough with significant recent increases in gang related violence. This pilot scheme involves the community lead deciding how to tackle gang related issues with police input as appropriate. 25 4. Impact on Offending In terms of impact analysis it has only been possible to review data relating to the primary aim of Shield; reducing violent offending. Analysis looked to first compare offending levels of the a) Shield boroughs to the MPS average and then b) offending of the selected cohort against a matched control (see appendix p.32 for methodology). Shield Borough Level Offending Overall levels of gang crime indicators and proxies were compared between the year before Shield (year 1) to the year in which Shield boroughs went live (year 2); (violence against the person; youth violence; gang flagged crime; gun crime; non-domestic knife crime with injury where victim is under 25; gang flagged violence with injury; firearms discharges). It reveals a mixed picture with no clear narrative; however, given the relatively small numbers of individuals involved with the pilot, it is perhaps not surprising there is no clear effect on borough level offending. Overall the MPS saw a significant increase in ‘violence against the person’ (VAP) offences20, the widest definition of violence, including all offences from harassment to murder. In line with this, all three pilot boroughs recorded significant increases on the previous year.21 22 23 Increases (though not statistically significant) were also apparent MPS wide in youth violence and firearms discharges over the comparison period. All three boroughs also reported non- significant increases against these categories. Westminster recorded significant decreases in gang flagged crime24 and non-domestic knife crime with injury25 where the victim is under 25 (a gang crime proxy measure), whereas in comparison the MPS showed a smaller (non-significant) decrease over the same period. Lambeth reported a non-significant decrease in gang flagged offences, which may be the result of local recording practice rather than reflective of real terms decline in gang related crime. Lambeth and Westminster recorded decreases in gun crime between the two comparison periods in line with MPS decreases, however the overall low prevalence means caution should be exercised in interpreting these results. Gang flagged crime (violence with injury only) was explored to control for spikes in operational activity against specific gangs (i.e. proactive operations where gang flagged weapon possession or drug offences may increase). Whilst all three boroughs recorded decreases in gang flagged violence, the MPS also recorded a decrease over the comparison period and the small numbers 20 Year 1 (M=17008.83, SD=948.959) and year 2 (M=19168.33, SD=800.557); t(11)=-14.508, p=0.000 21 Year 1 (M=832.67, SD=70.997) and year 2 (M=893.67, SD=76.245); t(11)= -3.39, p=0.006 22 Year 1 (M=775.58, SD=39.116) and year 2 (M=863.33, SD=46.682); t(11)=-6.049, p=0.000 23 Year 1 (M=575.92, SD=31.87) and Year 2 (M=686.00, SD=52.96); t(11)=-8.94, p=0.000 24 Year 1 (M=6.17, SD=4.687) and Year 2 (M=2.92, SD=2.193); t(11)=2.297, p=0.042 25 Year 1 (M=4.92, SD=2.539) and Year 2 (M=2.83, SD=1.642); t(11)=2.926, p=0.014 26 mean caution should be exercised. Overall, there was no clear indication that Shield has had any influence upon borough level violent offending. Shield Cohort Offending The entire Shield cohort consisted of 321 males; of these, 31 were excluded from the final analysis for various reasons including being removed from the cohort; being assigned membership to multiple gangs; not meeting selection criteria; or indicating changed or unknown gang allegiance. Therefore, final analysis focussed upon a total of 290 individuals across 15 gangs (five in Lambeth; six in Westminster and four in Haringey). A matched comparison group of 397 individuals was identified. See appendix for methodology, borough breakdowns and strength of matched control over a range of demographic and offending data. Recorded offending was explored using police data, to identify which individuals had appeared as either a named suspect; had been arrested; or had been charged. The analysis took into account the staggered ‘go live’ dates for each borough starting with the first from June 2015 until the end of July 2016. It is recognised that analysis using police data has a number of limitations in terms of data quality; timescales would not allow for conviction analysis, and as with all police recorded data will not be a true reflection of criminal activity. Chi square tests were used to identify statistically significant differences between groups. i. Violence and Wider Criminality Given the low prevalence of serious violence, analysis initially explored impact on wider categories of offending. When the Shield cohort (treatment) are compared to the 397 individual males in the matched group (control) it is noted the overall offending (all notifiable offences) and violence against the person offending levels across the period of analysis are very similar (see Table 4, appendix). When exploring overall effect on offending (all notifiable offences), there was no significant difference between the Shield cohort and the comparison group:  Overall, the Shield boroughs and control boroughs were broadly comparable in terms of proportion of individuals identified as a named suspect (48%, mean average per person=1.2 vs. 46%, mean average per person=1.35)  Overall, the Shield and control boroughs were similar in proportion for individuals arrested (40%, mean=0.93 vs. 38%, mean=0.99)  Overall, the Shield boroughs had a slightly larger proportion of individuals charged than the control (36%, mean=0.68 vs. 32%, mean=0.68) The similarity between Shield and control groups is further demonstrated by comparing average arrests, normalised against Shield start date, with little difference between the two (see graph 1, appendix p.35). 27 When exploring overall effect on violence against the person offending only, there was also no significant difference between the Shield cohort and the comparison group:  Overall, the Shield boroughs had a slightly larger proportion of individuals identified as a named suspect in a VAP offence than the control. (32%, mean=0.52 vs. 27%, mean=0.52)  Overall, the Shield boroughs had a slightly larger proportion of individuals arrested for a VAP offence than the control. (25%, mean=0.35 vs. 21%, mean=0.35)  Overall, there was no difference in proportion of individuals charged for a VAP offence between the Shield boroughs and the control. (15%, mean=0.19 vs. 15%, mean=0.21) ii. Borough Level Analysis In terms of exploring effect across specific boroughs, the low numbers of individuals who are suspected/arrested/charged makes analysis difficult. The above analysis was replicated for each borough; only one statistically significant difference between Shield group and control was found:  Lambeth had a statistically significant greater proportion of individuals (54%, n=65) charged for any offence than the control (40%, n=43)26 iii. Custody Analysis The research also looked at the number of the Shield cohort that had been identified as changing Matrix status from ‘live to ‘custody’ at least once during the period of analysis (a proxy for receiving a custodial sentence); there was no difference in proportion of Shield and control (30% Shield vs. 30% control) iv. Serious and Gang Flagged Violence Given Shield’s primary aim of reducing serious violence amongst the selected cohort, serious violent offending27 and gang flagged violence28 were also included in the analysis. Low prevalence of both categories meant analysis was only conducted at the aggregate level across all three boroughs. It was found that: When exploring serious violent offending there was no significant difference between the Shield cohort and the comparison group:  Overall, the Shield cohort had a slightly larger proportion of individuals identified as a named suspect for a serious violence offence than the control (19%, mean average per person=0.30 vs. 16%, mean average per person=0.19) 26 X2 (1, N=230) = 4.1693, p = .041162 27 Serious Violence was a composite of threshold offence categories namely; Murder, Attempted Murder, Aggravated Burglary, Firearms Possession or use; Weapon Enabled Crime (including stabbings, excluding knife possession); GBH and Serious Wounding, Affray, Violent Disorder and Kidnapping. 28 Due to low prevalence of gang Flagged serious violence, Gang Flagged Violence includes all VAP offences; Aggravated Burglary, Affray, Violent Disorder Firearms Possession or use, Weapons Enabled Crime; Sexual Offences and Robbery with ‘GA’ flag on CRIS. 28  Overall, the Shield cohort had a slightly larger proportion of individuals arrested for a serious violence offence than the control (15%, 0.19 vs. 13%, 0.17)  Overall, the Shield cohort had a slightly larger proportion of individuals charged (9%, mean=0.10 vs. 7%, mean=0.09) When exploring overall effect on gang flagged violence only, there was also no significant difference between the Shield cohort and the comparison group:  Overall, the Shield and control cohorts were very similar in terms of proportion of individuals identified as a named suspect (9%, mean average per person=0.10 vs. 8%, mean average per person=0.09)  Overall, the Shield and control cohorts were the same in terms of proportion of individuals arrested (7%, 0.08 vs. 7%, 0.09)  Overall, the Shield and control cohorts were the same in terms of proportion of individuals charged (4%, mean=0.06 vs. 4%, mean=0.05) As a result, there is no clear narrative in terms of an impact upon either overall offending, or on the violent offending of the specific Shield nominals in the 13 months subsequent to delivery. 29 4. Discussion Conclusion and Core Learning Shield was set up in Lambeth, Westminster and Haringey specifically as a pilot scheme based upon prior academic learning to address serious group related violence. The aim of pilot schemes is to implement for a time-limited specific period, explore viability and generate learning. It was not possible to demonstrate a significant reduction in violence across the targeted Shield groups - this does not indicate a GVI approach does not work or is not fit for London - rather that the challenges in implementation resulted in no clear test of the model. Much of the learning relates to implementation; interpretation of the GVI model; gaining sufficient support from the right partners at the right level; engaging and gaining community buy in to the pilot; and ultimately whether the model can be translated to London - given the different legal mechanisms available such as storing risk for Collective Enforcement and powers to compel call-in attendance. Positively, opinions towards the scheme changed over time, gradually becoming more favourable to core elements of the approach, notably the collective efficacy element. Despite initial strong misgivings by some practitioners and community members - all three boroughs ended the pilot in the belief it had been a positive learning experience and all are mainstreaming elements of the approach. Most importantly, Shield has opened new channels for dialogue, discussion and action, highlighting a shared determination to work together to help reduce gang related violence across London. The core learning is:  Implementation challenges can be anticipated and built in to any programme - Many of the implementation problems and delays stemmed from an incomplete understanding of how the fundamental elements of the GVI approach could be practically implemented and how much scope there was for flexibility. Standardisation and data collection issues hampered the evaluation, in terms of determining impact and potential outcomes other than the reduction of violence (i.e. referrals), although it is noted that given the different starting points and pre-existing approaches, standardisation was always likely to be a key risk to the evaluation. Implementation issues were exacerbated by a conflation of ‘business as usual policing’ (including operations with similar objectives) with Shield specific activity.  Meaningful, wide reaching and early community engagement is paramount - The pilot was thought to have highlighted and helped begin to address some significant issues across disengaged communities, individuals and voluntary groups. This is true both of statutory bodies’ relationship with local communities, and local groups’ knowledge of each other’s work towards reducing gang violence. Shield highlighted the need for an inclusive and wide ranging engagement process for future interventions, in which communities are empowered to contribute (e.g. representation at local 30 intelligence meetings), and the need to create and support local approaches which complement and collaborate with London wide or national initiatives.  Branding & media communication need to be sensitively handled – Closely linked to community engagement is the need for a strategic approach to messaging and communication. A clear, accurate and sensitive method for communications is vital, as is more nuanced consideration of the branding and terminology applied to such schemes (e.g. the focus on ‘the gang’ rather than violence per se). Shield’s external communications were widely criticised as insensitive and misleading, and were thought to contribute to the considerable barriers to moving the project forward. Haringey’s long term approach in involving the community in design (e.g. branding) and implementation (e.g. intelligence access) diverged from the standardised approach, but was thought to achieve positive outcomes in developing community relations.  Wider partnership buy-in is needed to facilitate ‘Collective Enforcement’ effectively - This is true both for civil opportunities and more serious offences; the ability to fast track cases for ‘swift and certain’ action is currently limited, although it is recognised that UK legislation may contribute to this. Additionally, obtaining sufficient legal powers to compel nominals to attend call-ins is challenging. 31 Appendix Pilot Timeline The timeline below depicts the key event dates for each of the pilot boroughs. The first call-in signifies the start of ‘go-live’. Details of trigger offences dates and subsequent Collective Enforcement activity are included. Evaluation Methodology Surveys Survey One ran for three weeks at the beginning of implementation (from 01/06/2015) and yielded 30 participants. Due to the timing, a far greater proportion of respondents were from Lambeth (57%, n=17), four were from Westminster (n=4), one from Haringey and eight pan- borough. Respondents were divided between Strategic Roles (n=11), Project Managers (n=6) and Practitioners (n=11). Half of respondents were police officers (n=15). Survey Two ran for four weeks around the middle of the programme implementation (from 16/12/2015) and yielded 23 respondents, nine from Lambeth, three from Westminster, one from Haringey and ten pan-boroughs. Overall survey response rates were considered low, but due to a ‘snowball’ email distribution method across multi-agencies it was impossible to know how many relevant practitioners were reached. Interviews Semi-structured, informal interviews were conducted with key Shield practitioners across pilot boroughs over the duration of the pilot. Interviewees were selected from as wide a range of service areas as possible and included analysts, strategic management, central project teams, police officers, youth workers and community development. A total of 18 individuals were interviewed, with all boroughs contacted in the months after the pilot to provide legacy updates. Lambeth Call In - Early June 2015(One location) Call In - 2 Locations - Early October 2015 Notifications in Person - Early June 2015 Threshold Offence - Early October 2015 Threshold Offence - Early July 2015 Collective Enforcement - Mid October 2015 Collective Enforcement - Early July 2015 Westminster Call In - 3 Locations - Late September 2015 Call In - Early March 16 Threshold Offence - Late September 2015 Collective Enforcement - Late September 2015 Haringey Call In - 1 Location - Early March 2016 W1 W2 W3 W4 W5 W6 W7 W8 W9 W10 W11 W12 W13 W14 W15 W16 W17 W18 W19 W20 W21 W22 W23 W24 W25 W26 W27 W28 W29 W30 W31 W32 W33 W34 W35 W36 W37 W38 W39 W40 Final Evaluation Report Published Delivery Planning Call In 1 Data Collection Ends Notified Threshold Collectiv e Enforce Call In Invite Call In 2 Call In 1 Collective Enforce Call In 2 Sep-16 Oct-16 Call In 1 Collective Enforce Pilot Ends Aug-15 Sep-15 Jun-16 Jul-16 Aug-16 Oct-15 Nov-15 Nov-16 May-15 Jan-15 Feb-15 Mar-15 Apr-15 Apr-16 May-16 Dec-15 Jan-16 Feb-16 Mar-16 Jun-15 Jul-15 32 Borough Feedback All three boroughs provided regular updates to the Shield programme board on all aspects of programme implementation. All three boroughs also submitted ‘lessons learnt’ reports to the December 2015 board, as well as further updates at the final board in April 2016. Cohort Churn The difficulties in identifying and tracking gang membership and allegiance have been highlighted by the pilot, particularly given the fluidity of many of London’s street gangs. This is evidenced in the data collection; individuals swap gang assignation between the Matrix and the Shield lists; are assigned multiple gangs; or are part of gangs who split or rename. For example, a Haringey cohort member assigned to a Haringey gang for Shield, featured on the Lambeth Matrix as an Islington Gang member one year previously. 18 individuals across the three cohorts did not appear on the Matrix at all between 02/06/2015 – 31/08/2016 (11 on Haringey, two on Lambeth and five on Westminster).  66 of the Haringey cohort featured on the Matrix the whole time period. 19 nominals appeared partially. The majority (17) were added during the course of the pilot. Only two nominals were removed, both amber (via Intel Management Score) but with low activity.  80 of the Westminster cohort featured on the Matrix the whole time period. Four appear partially, two were added whilst two scoring individuals were removed, one moving out of MPS jurisdiction and one for reasons unknown (amber) gang flag removed.  93 of the Lambeth Cohort featured across the whole time period. 37 of the Lambeth cohort appeared partially, with the majority of these being removed for no activity/scoring (24). One individual was murdered, whilst several scoring individuals moved to outside the borough (6). Six nominals in the final cohort were added during the time period. Matched Control Methodology Impact analysis employed a matched control group methodology. This compares the Shield outcome to statistically similar boroughs and gangs. Although this limits the weight of causal statements in comparison to a Randomised Control Trial (RCT); due to the practicalities of implementation it is the most robust design possible.29 The level of detail and resources assigned to the matching process (i.e. borough involvement, Social Network Analysis requests, individual level matching) was curtailed, as it became clear that implementation issues, standardisation of approach across boroughs and roll-out delays would affect the ability to assess impact. 29 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/220542/magenta_book_combined.pdf 33 Control boroughs were identified using a version of ‘nearest neighbour matching using a number of key socio-demographic factors, as well as prevalence of various crime types (below) with weighting given to offences viewed by the MPS as gang indicators (e.g. knife and gun crime). The process mirrored the quantitative element of analysis used to select the pilot boroughs. Boroughs were excluded from selection based on a number of rules to control for contamination (e.g. neighbouring boroughs, boroughs with frequent cross-border gang tensions). The final pool of boroughs was decided in consultation with MPS central intelligence unit gang specialists, who provided professional opinion and the latest intelligence picture. Gangs within the selected boroughs were matched using a similar process, based on prevalence and seriousness of offending within each gang across a number of crime types, whilst giving consideration to gang size, and some demographics. The variables matched on are displayed below: Gangs were matched initially on prevalence of offending across a number of crime types. The final stage involved taking into account gang demographics. The table below displays basic demographics of the gangs included in the pilot by borough and their corresponding matches, aggregated to borough level. Table 2: Matched Gang Demographics and Matrix Status Matched Cohort Breakdown Treat Control Treat Control Treat Control Treat Control Cohort Count 122 108 85 167 83 122 290 397 Age Average 20.4 20.4 20.6 20.3 23.3 20.9 21.3 20.5 Min 14 14 15 12 15 15 14 12 Max 30 35 31 38 44 29 44 38 %Under 18 18.0% 24.1% 15.3% 24.0% 8.4% 18.0% 14.5% 22.2% %18-24 68.0% 62.0% 72.9% 64.7% 60.2% 65.6% 67.2% 64.2% %Over 24 13.9% 13.9% 11.8% 11.4% 31.3% 16.4% 18.3% 13.6% Ethnic Appearance White 4.9% 7.4% 9.4% 34.1% 0.0% 2.5% 4.8% 17.1% BAME 95.1% 92.6% 90.6% 65.9% 100.0% 97.5% 95.2% 82.9% Matrix % on for >90% of Analysis* 77.0% 89.8% 94.1% 73.7% 81.9% 87.7% 83.4% 82.4% % Red (Live only, June 2015)** 3.0% 8.3% 4.8% 3.5% 5.4% 4.8% 4.1% 5.2% % Amber (Live Only, June 2015) 45.5% 47.2% 41.9% 28.7% 30.4% 34.5% 40.6% 35.4% % Green (Live only, Junne 2015) 51.5% 44.4% 53.2% 67.8% 64.3% 60.7% 55.3% 59.4% % Custody >90% of Analysis 13.9% 11.1% 8.2% 17.4% 4.8% 13.9% 9.7% 14.6% % Live >90% of Analysis 39.3% 41.7% 44.7% 35.3% 54.2% 34.4% 45.2% 36.8% Lambeth Westminster Haringey Overall Figures relate to valid cohort only, some individuals excluded from anlysis due to data complications; *Analysis period=June 2015-July 2016. **Figures are for live nominals only, excluding nominals not on Matrix on 02/06/2015 Table 1: Matching Criteria 34 Impact Analysis All data for Key Gang Indicators at borough level is sourced from MetMIS (Table 3). Table 4 compares the proportion of suspects for Shield (treatment) and control across all categories, also giving the average and maximum number of each occurrence for an individual (minimum always 0). All data extracted from CRIS using PNC ID and cross referenced with Matrix for quality assurance. Percentage changes in red denote increases, green decreases and amber no change. Statistically significant increases or decreases are highlighted with blue borders. Table 3: Key Gang Indicator Comparison Table 4: Shield Cohort Offending Offence OCU Name Jun-14 to May- 15 Jun-15 to May- 16 % Change Violence Against the Person Lambeth 9992 10724 7.3% Westminster 9307 10360 11.3% Haringey 6911 8232 19.1% MPS 204106 230020 12.7% Youth Violence Lambeth 632 705 11.6% Westminster 386 399 3.4% Haringey 616 650 5.5% MPS 15937 16518 3.6% Gang Flagged Crime Lambeth 156 99 -36.5% Westminster 74 35 -52.7% Haringey 113 135 19.5% MPS 1791 1660 -7.3% Gun Crime Lambeth 106 85 -19.8% Westminster 66 60 -9.1% Haringey 68 103 51.5% MPS 1915 1828 -4.5% Knife Crime Victim U25 (Non Domestic) Lambeth 89 102 14.6% Westminster 59 34 -42.4% Haringey 73 84 15.1% MPS 1681 1665 -1.0% Gang Flagged Violence with Injury Lambeth 38 27 -28.9% Westminster 29 17 -41.4% Haringey 42 33 -21.4% MPS 566 554 -2.1% Lethal Barrelled Gun Discharges Lambeth 16 25 56.3% Westminster 5 6 20.0% Haringey 5 11 120.0% MPS 212 254 19.8% Borough Valid Cohort 122 108 85 167 83 122 290 397 Condition Treat Control Treat Control Treat Control Treat Control Ever Suspect 77 63 41 85 21 35 139 183 % 63.1% 58.3% 48.2% 50.9% 25.3% 28.7% 47.9% 46.1% Max 10 16 8 15 3 7 10 16 Average 1.93 2.01 1.02 1.47 0.33 0.59 1.21 1.35 Ever Arrest 72 54 34 69 10 26 116 149 % 59.0% 50.0% 40.0% 41.3% 12.0% 21.3% 40.0% 37.5% Max 7 12 7 14 3 7 7 14 Average 1.56 1.51 0.76 1.05 0.18 0.45 0.93 0.99 Ever Charge 65 43 29 61 10 22 104 126 % 53.3% 39.8% 34.1% 36.5% 12.0% 18.0% 35.9% 31.7% Max 6 6 6 8 3 5 6 8 Average 1.07 0.98 0.65 0.73 0.14 0.32 0.68 0.68 Ever Suspect (VAP) 57 49 27 44 9 14 93 107 % 46.7% 45.4% 31.8% 26.3% 10.8% 11.5% 32.1% 27.0% Max 5 10 4 4 2 4 5 10 Average 0.89 0.99 0.39 0.43 0.12 0.23 0.52 0.52 Ever Arrest (VAP) 46 38 20 33 5 12 71 83 % 37.7% 35.2% 23.5% 19.8% 6.0% 9.8% 24.5% 20.9% Max 4 8 3 3 1 3 4 8 Average 0.6 0.62 0.27 0.28 0.06 0.2 0.35 0.35 Ever Charge (VAP) 29 30 12 20 3 9 44 59 % 23.8% 27.8% 14.1% 12.0% 3.6% 7.4% 15.2% 14.9% Max 3 5 3 3 1 3 3 5 Average 0.31 0.39 0.16 0.14 0.04 0.14 0.19 0.21 Date Range Overall Mar-16 to Jul-15 Lambeth Westminster Haringey Jun-15 to Jul-16 Sep-15 to Jul-16 Mar-16 to Jul-15 35 Graph 1: Average Arrests (all offences) The graph shows average arrests by date offence committed, for all notifiable offences. Post and Pre periods are equal to 30 days. Data is normalised to the start date of each Shield borough. Total period of analysis is 01/01/2015 until 31/07/2016. 36 References i Whilst the definitional anomalies in the term ‘gang’ are recognised, this report adopts the term given its ubiquity in current organisational vocabulary. The current MPS definition of a ‘gang’ is: “A relatively durable, predominantly street- based group of young people who (1) see themselves (and are seen by others) as a discernible group,(2) engage in a range of criminal activity and violence, they may also have any or all of the following features; (3) identify with or lay claim over territory,(4) have some form of identifying structural feature, and (5) are in conflict with other, similar, gangs.” As derived from: http://www.centreforsocialjustice.org.uk/core/wp-content/uploads/2016/08/DyingtoBelongFullReport.pdf ii MetMIS is an internal MPS tool for obtaining police performance data. iii All data from MOPAC’s Public Attitudes Survey (PAS) (Q1 15/16 R12) / MOPAC Confidence dashboard. iv Available from: http://www.eif.org.uk/publication/preventing-gang-and-youth-violence/ v Available from: https://www.gov.uk/government/news/repeat-knife-offenders-face-new-minimum-prison-term vi Over 50 cities or districts including Baltimore, Boston, Chicago, Cincinnati, Los Angeles (Source: www.nnscommunities.org) vii Braga, A. & Weisburd, D., (2012). The Effects of Focused Deterrence Strategies on Crime: A Systematic Review and Meta- Analysis of the Empirical Evidence. Journal of Research in Crime and Delinquency, 49(3), pp.323–358. Available from: https://campbellcollaboration.org/library.html viii National Network for Safe Communities (2013). Group Violence Intervention: An Implementation Guide. Washington, DC: U.S. Department of Justice, Office of Community Oriented Policing Services. Available from: https://nnscommunities.org/old-site-files/Group_Violence_Intervention_-_An_Implementation_Guide.pdf ix ibid x Williams D.J., et al. (2014) Addressing gang-related violence in Glasgow: A preliminary pragmatic quasi-experimental evaluation of the Community Initiative to Reduce Violence (CIRV), Aggression and Violent Behaviour, 19(6) pp.686-691. xi Burns, L., Williams, D. J., & Donnelly, P. D. (2011). A public health approach to the evaluation of the Glasgow Community Initiative to Reduce Violence. In Glasgow’s Community Initiative to Reduce Violence: Second year report. (pp. 28). Glasgow: Violence Reduction Unit. xii Metropolitan Police Service (2010) The London Pathways Initiative: Lessons Learned. MPS: Unpublished. xiii See for example; Densley, J. & Jones, D. (2016). Pulling levers on gang violence in London and St. Paul. In C. Maxson & F- A. Esbensen (Eds.), Gang transitions and transformations in an international context (pp. 291–305). New York: Springer. xiv National Network for Safe Communities (2013). Group Violence Intervention: An Implementation Guide. Washington, DC: U.S. Department of Justice, Office of Community Oriented Policing Services. Available from: https://nnscommunities.org/old-site-files/Group_Violence_Intervention_-_An_Implementation_Guide.pdf xv Kennedy, David M., and Michael A. Friedrich. 2014. Custom Notifications: Individualized Communication in the Group Violence Intervention. Washington, DC: Office of Community Oriented Policing Services. https://ric-zai- inc.com/Publications/cops-p304-pub.pdf xvi Dawson P. & Stanko E. (2013) Implementation, implementation, implementation: Insights from offender management evaluations. Policing. 7(3), 289-298. xvii There is a continuing debate regarding gang organisation in the UK; see Hallsworth, S. (2013) The Gang & Beyond: Interpreting Violent Street Worlds UK: Palgrave Macmillan and Densley, J. (2013) How Gangs Work: An Ethnography of Youth Violence UK: Palgrave Macmillan for detailed accounts of the competing perspectives. xviii Mayor launches tough new gang intervention programme, 22 January 2015 https://www.london.gov.uk/press- releases/mayoral/one-rule-for-all xix ‘Exemplary enforcement action against a local gang ahead of the first borough call-in in order to demonstrate to other gangs that the partnership is serious about ending violence’ adapted from National Network for Safe Communities (2013). Group Violence Intervention: An Implementation Guide. Washington, DC: U.S. Department of Justice, Office of Community Oriented Policing Services. Available from: https://nnscommunities.org/old-site-files/Group_Violence_Intervention_- _An_Implementation_Guide.pdf Embargoed report – not for sharing Review of the Metropolitan Police Service Gangs Violence Matrix – Second Annual Review October 2022 MOPAC Evidence and Insight 2 A Review of the MPS Gangs Violence Matrix – Second Annual Review Executive summary About the Gangs Violence Matrix Second Annual Review The Metropolitan Police Service (MPS) Gangs Violence Matrix (GVM) is a tool used to identify and risk-assess the most harmful gang members in a Borough. The Mayor of London, Sadiq Khan, made a commitment in his 2016 manifesto and his Police and Crime Plan 2017-21 to conduct a review of the GVM. This review was published in December 2018 and was the largest and most comprehensive exploration ever conducted into the Matrix. One of the recommendations was the need to systematically capture key elements of the Matrix process and report annually on outputs in terms of the Matrix population. MOPAC’s Evidence & Insight team produced the first ‘one-year update’ document published in 2021. This report present findings from the second annual review. The analysis in this paper uses Matrix data and crime data up to December 2021. Key Findings: GVM Population • Police recorded crime data shows that violent crime increased over 2021 and has exceeded pre-coronavirus pandemic levels (largely driven by an increase in violence without injury). Gang-flagged violence accounts for a relatively small proportion of overall levels of violent crime in London, but it represents a significant percentage of the most serious offences (e.g., 81% of all shootings). • Results from MOPAC’s Public Attitude Survey (PAS) show 59% of adult Londoners felt gangs were a problem in their local area during 2021. Less than half of adult Londoners believed the Metropolitan Police Service tackles gang crime effectively (47%, n=2,040). • Londoners living in the most deprived areas of the capital were far more likely to feel gangs were a problem in their local area (70%) than those living in the least deprived areas (43%). • Around one in five Londoners said they had previously heard about the Matrix (19%). While 67% felt confident the Matrix was used fairly, large inequalities by ethnicity emerged. Black (49%) and Mixed (54%) Ethnic Backgrounds were least likely to feel confident the tool was used fairly – compared with 73% of White Londoners. • A quarter of young people surveyed within MOPAC’s Youth Survey 2021-22 (11,874 responses aged 11-16) believed ‘people joining gangs’ was a big problem in their local area (25%), with higher concerns seen here amongst those from a Black Ethnic Background (32%, compared with 21% of White). • One in ten young Londoners said they knew someone who was in a gang (10%). Exposure to gangs increased with age and Black (15%) and Mixed (14%) Ethnic Backgrounds were more likely to say they knew someone in a gang than those from White (10%) or Asian (6%) ethnic backgrounds. • The 2021 GVM population has halved (-49% decrease) from the peak in August 2017. The December 2021 population of 1,933 represents the smallest GVM population to date. However, within this are a small number of boroughs that have seen increases in their Matrix population over recent years. • Despite the reduction in the Matrix population the proportion of individuals with Green RAG status has remained unchanged, with almost two thirds of the population being Green harm banded (65%). Key Findings: Violence in London and Londoners perceptions • There remains an over representation of young Black males on the overall GVM population as compared to both police recorded offending and victimisation cohorts. • Generally, disproportionality lessens when looking at more serious violent offending, with the cohort for personal robbery being closest to the cohort on the Matrix that are from Black and/or other Minority Ethnic communities, but this does not explain away all the disproportionality. • Looking at ‘Threat to Life’ data, across both suspects and victims approximately 74% were Black males. This aligns more closely to the proportion of the Matrix that is Black (80%) – so may be a useful addition exploring disproportionality. However, the small number of Threat to Life data and the over emphasis on certain boroughs make firm conclusions difficult. • Modelling indicates the strongest variables predicting inclusion onto the Matrix focus on the offending history of the individual – be it robbery, group offending, repeat offending and higher harm offending as measured by the Cambridge Harm Index. Age and Sex are also key. Ethnicity is also a significant variable but has a lesser role. • This disproportionality of the Matrix cohort (as compared to the local population) is most stark in the more affluent boroughs. • There has been a gradual shift in the age profile of the GVM with the proportion of those under 18 and under 25 decreasing (i.e., a 6 pp. reduction in the number of under-18s on the GVM since 2018). • Historically, compared to the estimated population of London, the GVM has been comprised disproportionately of Black African-Caribbean males. When examining the most recent two years, the proportion of the Matrix that were Black African-Caribbean has not changed (accounting for 80% of the cohort). • Although there have been some changes in the ethnicity of those added to the Matrix (slightly less disproportionate), this has not been enough to make any noticeable impact on overall population. The proportion of new additions that were Black African Caribbean has reduced from 83% in 2019 to 73% in 2021. Conclusion The report completes the second annual review of the MPS Gangs Violence Matrix – focussing upon the population of 2021. As a whole, there is a clear decrease in the overall numbers on the Matrix and these individuals are less likely to be young. Demographically there has been little change, the majority on the tool remain male and Black African-Caribbean - although there are signs that new Matrix additions are less disproportionate and more in line with the violent offending cohort. Understanding the drivers of this disproportionality is a difficult and complicated task and the paper seeks to advance our understanding by looking at new data such as ‘threats to life’, Social Economic factors as well as statistical modelling to determine the main individual factors predicting inclusion onto the Matrix. This analysis demonstrates disproportionality (as compared to the local population) is most stark in affluent boroughs and that offending history is the strongest predictor for inclusion (along with Sex and age), but ethnicity cannot be ruled out. Key Findings: Disproportionality Review of the Metropolitan Police Service Gangs Violence Matrix – Second Annual Review Introduction The Metropolitan Police Service (MPS) Gangs Violence Matrix (GVM) was developed in the aftermath of the 2011 London riots and is a tool used to identify and risk-assess the most harmful gang members in a Borough. From inception, the GVM has been controversial and a number of in-depth reviews have been conducted focusing on issues such as disproportionality and data protection. These include the ICO Penalty Enforcement Action (October 2017 to November 2018), Amnesty review (mid 2017 to May 2018) and the Mayoral review which took place Autumn 2017 to December 2018). The Mayor of London, Sadiq Khan, made a commitment to Londoners in his 2016 manifesto and his Police and Crime Plan 2017-21 to conduct a review of the GVM. This review was published in December 2018 and was the largest and most comprehensive exploration ever conducted into the Matrix – exploring themes of impact, partnership views and disproportionality. The Review set out nine recommendations to be completed by the 31st December 2019. Established within these recommendations was the need to systematically capture key elements of the Matrix process and report annually on outputs in terms of the Matrix population. Further to the recommendation above, MOPAC’s Evidence & Insight team produced the first ‘one-year update’ document examining data up to October 2019. Delayed due to COVID-19, this report was published in Feb 2021. This report present findings from the second annual review. The analysis in this paper uses Matrix data and crime data up to December 2021. The report first examines recent trends in violence, followed by detailed analysis of the Matrix population and a comparison with the wider London gang and violent offending profiles. This report also includes a more detailed analysis of the levels of disproportionally observed on the Matrix and an exploration of the key predictors for inclusion on the GVM. The analysis does not intend to replicate the original comprehensive analysis exploring impact and process presented in the landmark Mayoral Review of the Matrix. The focus of this paper is to explore the population on the Matrix, with an emphasis on how the Matrix population has changed - if at all - in terms of size, demographics (disproportionalities) and harm in the period since the Mayoral Review was conducted. 5 The Backdrop of Violence in London Patterns of crime over the last two years have been substantially affected by the coronavirus (COVID-19) pandemic and government restrictions. Since restrictions were lifted following the third national lockdown in early 2021, police recorded crime data shows that violent crime has increased and has exceeded pre-coronavirus pandemic levels (with violence against the person offences recording an increase of +5% as compared to the pre-coronavirus year ending 2019). The increase in violence against the person is largely driven by an increase in violence without injury (+12% as compared to 2019), whereas weapon-enabled crime has shown a consistent reduction over the last three years (with knife enabled crime 32% lower). Despite an increase over 2021, violence with injury is still below pre-coronavirus levels (-7% as compared to 2019). For a deeper exploration of the most recent MPS violence statistics – see Table 1 below. Positively, there would appear to be signs that this violence had begun to stabilise in London, even prior to COVID-19 restrictions. Yet, since the loosening of restrictions certain crimes, particularly Youth Homicide and Violence have increased, whereas weapon enabled violence continues to decrease. Table 1: Metropolitan Police Service recorded crime statistics comparing 2019, 2020 and 20211 In addition to recorded crime in London, it is important to understand what proportion of certain offences are believed to be related to gangs. Caution should be taken when assessing gang related activity through police indices, as there is potentially an underrepresentation of the true prevalence in this area given the nature of gang definitions, quality of data recording 1 Please note that crime data was extracted July 2022. Police recorded crime is liable to change due to a number of factors; for example, transfer to another force or reclassification. 2019 2020 2021 % change 2019 v 2021 % change 2020 v 2021 Violence Against the Person 222,114 220,607 233,965 5.3% 6.1% Violence with Injury 77,491 69,308 71,802 -7.3% 3.6% Homicide 152 131 133 -12.5% 1.5% Youth Homicide 32 27 33 3.1% 22.2% Serious Youth Violence 8,568 6,431 5,630 -34.3% -12.5% Non Domestic Abuse VWI 53,743 45,433 49,074 -8.7% 8.0% Violence without Injury 144,471 151,168 162,030 12.2% 7.2% Possession of Weapons 7,359 6,851 5,737 -22.0% -16.3% Knife Crime Offences 15,599 11,852 10,623 -31.9% -10.4% Knife Crime With Injury 3,987 3,202 3,185 -20.1% -0.5% Knife Crime Injury Victims under 25 (non DA) 1,661 1,208 1,207 -27.3% -0.1% Knife Crime with Injury Pers Robb Offs 553 452 367 -33.6% -18.8% Knife Possession 5,278 4,892 4,200 -20.4% -14.1% Gun Crime Offences 2,101 1,719 1,358 -35.4% -21.0% Gun Crime Lethal Discharge 270 307 211 -21.9% -31.3% Gun Crime Personal Robbery 479 328 254 -47.0% -22.6% Violence Against Person Weapons 6 and flagging and so on. None the less, examining the MPS data indicates that Gang-flagged violence represents a significant percentage of the most serious and harmful offending and victimisation as the table below illustrates. The proportion of incidences of ‘gang involvement’ across Homicide and Firearms offences would appear to have increased over the last three years, whereas the proportion of gang related knife injuries has reduced. See Table 2. Table 1: The proportions of violence associated with gangs (MPS)2 Londoners perceptions of gangs During 2021 the MOPAC’s Public Attitude Survey (PAS) interviewed a representative sample of 12,800 adult (16+) Londoners across a wide range of crime and justice issues. 59% of adult Londoners felt gangs were a problem in their local area during 2021. Relatively few differences were seen here by Ethnicity: Londoners from Mixed Ethnic Backgrounds were most likely to feel gangs are a problem in their local area - at 64% - but concerns remained relatively consistent across other ethnic groups, including those from White (57%), Black (60%) and Asian (61%) Backgrounds. Similarly, few differences were seen by age: although Londoners aged 65+ were least likely to feel gangs were a problem in their local area (44%), results for other age groups all stood at around 60% (16 to 24: 59%, 25 to 34: 60%, 35 to 64: 62%). However, Londoners living in the most deprived areas of the capital were far more likely to feel gangs were a problem in their local area (70%) than those living in the least deprived areas (43%)3. In turn, less than half of adult Londoners interviewed by the PAS between April and December 2021 believed the Metropolitan Police Service tackles gang crime effectively (47%). Around one in five Londoners said they had previously heard about the Metropolitan Police Gangs Violence Matrix (19%), while 67% felt confident the Matrix was used fairly. Particularly large 2 DA (Domestic Abuse), CT (Counter-Terrorism). 3 Areas grouped in quartiles based on the Index of Multiple Deprivation (IMD) 2019. 2019 2020 2021 All Homicides 151 132 133 Gang Related 24% 19% 25% Homicide (Non-DA & Non-CT) 135 109 107 Gang Related 27% 23% 31% Lethal Barreled Discharges 270 307 214 Gang Related 45% 59% 61% Lethal Barreled Discharges - Victim Shot 98 135 95 Gang Related 56% 77% 81% Knife Injury under 25 (Non-DA) 1,631 1,197 1,192 Gang Related 18% 15% 15% Homicide Firearms Knives Crime Type 7 inequalities by ethnicity emerged here, with those from Black (49%) and Mixed (54%) Ethnic Backgrounds least likely to feel confident the tool was used fairly – compared with 73% of White Londoners. The MOPAC Youth Survey4 2021-22 gathered the views of 11,874 young Londoners aged 11 to 16. A quarter of young people surveyed believed ‘people joining gangs’ was a big problem in their local area (25%), with higher concerns seen here amongst those from a Black Ethnic Background (32%, compared with 21% of White). 12% of young people believed ‘people joining gangs’ was a big problem in their school. One in ten young Londoners aged 11 to 16 said they knew someone who was in a gang (10%). Exposure to gangs increased with age: while 6% of those aged 11 years old said they knew someone in a gang, this reached 17% by age 16. Young people from Black (15%) and Mixed (14%) Ethnic Backgrounds were also more likely to say they knew someone in a gang than those from White (10%) or Asian (6%) Backgrounds. In total, 1% of young people aged 11 to 16 said they had personally ever been part of a gang themselves – although demographic differences seen here were far smaller. Population Change - A reducing Matrix population The overall population of the Gang Violence Matrix continues to decrease and has done so since November 2017. This decline corresponds to a number of reviews into the matrix; these include the ICO Penalty Enforcement Action (October 2017 to November 2018), Amnesty review (mid 2017 to May 2018) and the Mayoral review which took place Autumn 2017 to December 2018). As can be seen from Graph 1, the latest December 2021 population of 1,933 represents the smallest ever Gang Violence Matrix population. The GVM population has almost halved (-49% decrease) from the peak in August 2017. The move outside the dotted lines indicates a significant decrease in population compared to historical levels. 4 The Youth Survey fieldwork took place from November 2021 to January 2022. 8 Graph 1: Matrix overall population over time Initially the proportion of Matrix individuals in custody steadily increased over time (see Graph 2); with the rate accelerating between October 2018 and March 2020. Yet, during the post review period there has been a small but steady decrease. As of December 2021, 36.8% of Matrix individuals were in custody; this is less than the peak of 43.3% recorded in November 2019 and March 2020. Graph 2: Proportion of Matrix individuals in custody 9 Local variation remains evident in Matrix population change As we have seen, the size of the overall Matrix population has been reducing – however, we still see considerable variation across London Boroughs. Looking at long term change, comparing average borough Matrix population in 2021 to 2013, 27 out of 32 boroughs decreased their Matrix population with 5 showing a significant increase (increase > +40%). Of these, Islington recorded the largest notable increase, with an increase of 163% in their Matrix population since 2013, increasing from 70.4 to 184.9 average population. As Table 3 shows, the shorter-term view shows a similar picture, with 27 out of 32 boroughs experiencing a reduction in average population size between 2020 and 2021. Since 2017 the average Matrix population per borough has decreased to less than 100, falling from an average of 114.2 in 2017 to only 63.1 in 2021. Currently, only 7 boroughs have Matrix populations above 100, with all but one experiencing a reduction in population compared to the previous year. The boroughs with average Matrix populations of over 100 are Lambeth (- 19%, n=188.4), Islington (-19%, n=184.9), Waltham Forest (-7%, n=143.6), Camden (-12%, n=139.9), Newham (+2%, n=115.1), Westminster (-8%, n=109.4) and Brent (-18%, n=104.3). Table 3: Average monthly Matrix population, change and violent crime statistics Table 3 above illustrates the substantial fluctuation between years, with some boroughs experiencing significant population changes at various points in the intervening years. When 10 we consider these changes within the context of the overall violence across boroughs, we can see that the distribution of the Matrix population generally reflects the distribution of weapon enabled violence and violence with injury. That is to say generally speaking, areas with higher violence tend to have increased matrix numbers. Population Demographics A change in demographics of those on the Matrix A number of findings emerged from the original Matrix review in terms of population change and demographics – such as the almost exclusively male population, the over representation of young Black males when compared to other databases or cohorts, the substantial proportion that had remained on the Matrix for a number of years, as well as the number of individuals with a long term ‘green’ harm banding or ‘zero’ harm score. Table 4 Error! Not a valid bookmark self-reference. presents the demographics of the Matrix over time, compared with the latest additions in 2019, 2020 and 2021. This enables us to begin to explore whether the demographics of the group are changing. Table 4: Matrix demographics over time Comparison Snapshots all from October of given year apart from 2018 which relates to September snapshot presented in original review. As can be seen, in terms of gender, the proportion that is male has not changed in the two most recent years – the matrix cohort is, and has always been, made up almost exclusively of males (99.6% in 2019; 99.7% in 2020; 99.9% in 2021). 2013 2014 2015 2016 2017 2018 2019 2020 2021 Total Population 3,331 3,594 3,612 3,704 3,783 3,230 2,746 2,399 2,012 -16.1% Live 71.5% 70.2% 67.7% 67.8% 67.8% 64.8% 57.9% 60.1% 63.0% 2.9% Custody 28.5% 29.8% 32.3% 32.2% 32.2% 35.2% 42.1% 39.9% 37.0% -2.9% Gender % Male 99.1% 98.9% 99.3% 99.3% 99.3% 99.4% 99.6% 99.7% 99.9% Ethnicity % White 13.1% 13.6% 13.8% 13.6% 13.2% 11.7% 10.9% 12.7% 12.2% % BAME 86.9% 86.4% 86.2% 86.4% 86.8% 88.3% 89.0% 87.3% 87.8% % Black African-Caribbean 78% 76% 77% 77% 78% 80% 81% 80% 80% Age Average 20.8 21.5 21.5 21.9 22.2 22.3 22.8 23.1 23.3 Live 20.5 21.2 21.2 21.7 22.0 22.0 22.4 22.8 23.1 Custody 21.4 22.0 22.1 22.4 22.6 22.9 23.3 23.6 23.7 % U18 18.8% 14.7% 15.8% 14.2% 13.8% 13.8% 9.2% 6.9% 7.7% 0.7% % U25 85.7% 81.5% 80.0% 76.4% 73.1% 72.4% 67.2% 65.5% 63.2% -2.3% RAG Status Red 6.5% 4.9% 4.7% 5.4% 4.9% 4.3% 4.7% 5.0% 5.6% Amber 40.0% 33.3% 34.5% 33.5% 31.2% 31.1% 30.0% 29.0% 29.0% Green 53.5% 61.8% 60.8% 61.1% 63.8% 64.7% 65.3% 65.9% 65.4% % of GVM Green U18 7.1% 5.8% 5.9% 5.1% 4.7% 5.2% 3.2% 2.6% 2.4% -0.2% % of GVM Green 18-24 36.8% 41.5% 39.1% 38.4% 37.7% 37.3% 35.8% 36.1% 33.7% -2.4% Live Red 4.2% 3.0% 3.3% 4.7% 3.5% 3.2% 3.3% 4.9% 5.6% Amber 38.7% 31.9% 33.4% 32.6% 31.1% 31.1% 30.5% 30.0% 30.9% Green 57.0% 65.0% 63.3% 62.7% 65.4% 65.7% 66.2% 65.0% 63.5% % of Live GVM Green U18 9.0% 7.8% 8.2% 6.6% 6.3% 7.4% 5.3% 4.1% 3.5% -0.5% % of Live GVM Green 18-24 39.2% 43.0% 40.3% 38.3% 37.1% 37.4% 36.8% 36.3% 31.9% -4.4% Custody Red 12.2% 9.2% 7.7% 6.9% 7.9% 6.3% 6.6% 5.2% 5.5% Amber 43.0% 36.6% 36.7% 35.3% 31.6% 31.0% 29.2% 27.5% 25.8% Green 44.8% 54.2% 55.6% 57.8% 60.5% 62.8% 64.2% 67.3% 68.7% % of Custody GVM Green U18 2.1% 0.93% 0.9% 1.93% 1.3% 1.23% 0.3% 0.42% 0.40% 0.0% % of Custody GVM Green 18-24 30.8% 37.8% 36.7% 38.5% 39.1% 37.1% 34.5% 35.8% 37.0% 1.1% GVM Snapshot Comparison Real % Change 11 In terms of age of the cohort, the proportion and count of Under-18 individuals has decreased subsequent to the original review - from 14% in 2018 (n=445) to 8% in 2021 (n=154) – which at an individual level is a 65% reduction in this cohort. In 2020 Under-18s accounted for 7% of the cohort, the lowest recorded to date. A similar picture emerges when looking at the proportion of the Matrix that is under 25 (a reduction from 72% in 2018, to 63% in 2021). Over time, the age of individuals on the matrix has slowly increased. The average age of individuals on the matrix is now over 23 years old as compared to under 21 years old in 2013. The original Review of the Matrix identified a substantial number of individuals on the Matrix with a ‘Green’ harm banding and a ‘zero-harm’ score, and in turn recommended a ‘thorough reappraisal’ of these individuals. As a result, beginning in March 2019 the MPS undertook the review of all individuals on the Matrix with a ‘Green’ harm banding, with a focus on ‘live’ individuals (i.e., not in custody). There were 773 (of which 121 were Under 18) fewer individuals on the Matrix with Green RAG status in 2021 compared to 2018 - a 37% decrease. Yet despite the reduction in the Matrix population the proportion of individuals with Green RAG status has remained unchanged, with almost two thirds of the population being Green harm banded (65%). The proportion of Live Green RAG status remained stable (+0.7pp), while the Custody Green RAG status increased (+5.9pp) between 2018 and 2021. Over the longer term we’ve seen the proportion of individuals with Green RAG status in custody gradually increase. In 2013 only 44.8% of individuals in custody had Green RAG status, yet this has increased to 68.7% in 2021. Historically, the Matrix has been comprised disproportionately of Black African-Caribbean males. When examining the two most previous years, the proportion of the Matrix that are from Black and/or other Minority Ethnic communities has remained unchanged; 88% in September 2018 and 88% in October 2021. Similarly, the proportion of Black African Caribbean individuals has remained unchanged; 80% in September 2018 and 80% in October 2021. Analysis of how the ethnicity of the Matrix population is distributed by age shows that there is a peak between the ages of 20 and 25 (see Graph 3 below). 12 Graph 3: Gangs Violence Matrix population rate per 1,000 population across ethnicities Matrix Churn – who has been removed since the original review? One clear area of interest is a richer exploration of the churn that has been observed. As previously outlined, there has been a marked reduction in the overall size of the Matrix cohort. Over twice as many individuals have been removed from the Matrix than added since May 2018. This section explores the additions and removals to the Matrix. Additions to the Matrix There have been 1,096 additions to the Matrix after May 2018. See Table 5 below for an overview of those new additions as well as how this compares with those added to Matrix in previous years. As an overview, we can see: • Generally, the proportion that are male has remained stable; although this has reduced to 96% in the last 2 years. It remains that almost all new additions are male. • There are some changes in terms of ethnicity. The proportion of new additions that were white has fluctuated between 2018 and 2021, from a low of 10.9% (n=24) in 2019 to a peak of 27% in 2020 (n=110). In 2021 the proportion of additions that were white returned back to levels recorded in 2017.The proportion of new additions that were from Black and/or other Minority Ethnic communities has reduced from 89% in 2019 to 77% in 2021, as well as Black African Caribbean from 83% in 2019 to 73% in 2021. • The proportion of Under 18s has significantly reduced, with 40% of new additions during 2021 being this age, as compared to 56% in 2018. The average age of additions to the matrix of the last few years has remained unchanged. 13 • There were variations across boroughs on the number of additions to the Matrix. Between June 2019 and December 2021, the number of additions to the matrix have decreased. However, 58% of additions to the matrix in the 12 months between June 2019 and May 2020 were from Islington and Camden. Additions to the matrix in these two boroughs decreased by 83% in the following 12 months and is now similar to other boroughs across London. Table 5: Additions to the Matrix Removals from the Matrix A total of 2,416 removals from the matrix have occurred since May 2018. See Table 6 for an overview as how this compares to previous years and removals. Overall, findings indicate: • The overall proportion of under 18s and under 25s removed has decreased over time. In 2021, 6.0% of individuals removed from the matrix were under 18 compared to 12.7% in 2018. In comparison 53.4% of individuals removed from the matrix in 2021 were under 25 compared to 66.3% in 2018. This has resulted in the average age of individuals removed from the matrix increasing, though only marginally. • The proportion of those removed by ethnicity has remained relatively stable over time – for the BAME population this figure has remained around the 81-86% mark. Within this, the proportion of Black African Caribbean removals has fluctuated, increasing to 78% in 2021 compared to 67% in 2017. • Again, there is wide variation between boroughs in terms of numbers removed. Table 6: Removals from the Matrix Additions to Matrix Base Year One Year Two Year Three Year Four Year Five Year Six Year Seven Year Eight Year Nine 2012 Jun 13-May 14 Jun 14-May 15 Jun 15-May 16 Jun 16-May 17 Jun 17-May 18 Jun 18-May 19 Jun 19-May 20 Jun 20-May 21 Jun 21-Dec 21 Male 99% 98% 99% 98% 98% 99% 99% 100% 96% 96% White 13.5% 15.2% 17.6% 20.2% 13.0% 17.7% 10.9% 26.7% 12.2% 17.3% BAME 86.5% 84.8% 82.4% 79.8% 87.0% 82.3% 89.4% 72.1% 80.4% 77.4% Black African Caribbean 75.9% 75.2% 67.3% 69.0% 78.9% 74.6% 82.8% 61.2% 72.3% 73.2% U18 25.7% 31.9% 40.7% 48.6% 52.3% 55.6% 42.4% 29.1% 35.1% 39.9% U25 85.8% 86.2% 84.9% 90.2% 91.5% 93.8% 94.0% 85.0% 87.8% 88.1% Average Age 20.4 20.2 19.7 18.8 18.4 18.0 19.4 20.2 20.0 20.0 TOTAL ADDITIONS* 896 1,070 803 673 611 221 411 296 168 *Year Nine data is Jun-21 to Dec-21 inclusive. Removals from Matrix Base Year One Year Two Year Three Year Four Year Five Year Six Year Seven Year Eight Year Nine 2012 Jun 13-May 14 Jun 14-May 15 Jun 15-May 16 Jun 16-May 17 Jun 17-May 18 Jun 18-May 19 Jun 19-May 20 Jun 20-May 21 Jun 21-Dec 21 White 13.8% 18.4% 19.2% 16.7% 17.5% 15.0% 13.0% 14.6% 14.8% 14.8% BAME 86.2% 81.6% 80.8% 83.3% 82.5% 85.0% 86.2% 85.3% 85.2% 84.4% Black African Caribbean 70.8% 70.9% 65.6% 74.9% 67.3% 75.8% 79.3% 76.3% 76.3% 78.1% U18 21.4% 16.1% 11.5% 11.3% 13.5% 12.7% 9.8% 7.2% 4.4% 6.0% U25 77.7% 82.0% 73.5% 74.2% 69.9% 66.3% 64.3% 57.5% 59.4% 53.4% Average Age 21.6 21.5 22.4 22.6 22.6 23.0 23.2 23.8 23.8 24.5 TOTAL REMOVALS* 750 1,064 697 631 1,006 600 843 608 365 *Year Nine data is Jun-21 to Dec-21 inclusive. 14 Disproportionality The Mayoral review highlighted the heavy disproportionality on the Matrix – specifically in terms of Black African Caribbean males relative to their offending and victimisation. One of the new analytic directions within the present review was an exploration in more detail towards the levels of disproportionally observed on the Matrix. This analysis explores borough-level disproportionality as well an exploration of the key predictors for inclusion on the GVM. London-level disproportionality This analysis takes the Black under-25 cohort (as the most likely group to be included on the Matrix) and explores the proportion of individuals charged with various violent and group- related offence types; and compares these to the proportion of Black individuals on the Matrix. As shown in Graph 4 below - at a London wide level, the percentage of Black, African- Caribbean individuals on the Matrix exceeds that of the general population. Black individuals aged under 25 account for 22% of the London under 25 population while accounting for 76% of the under-25 matrix population in 2021. There remains an over representation of young Black males on the overall matrix population as compared to offending cohorts, including those most associated with serious gang - or group-related violence. Generally, disproportionality lessens when comparing with violent offending such as lethal barrel discharge, robbery and murder - but the representation of young Black males on the Matrix is still higher. The proportion of new additions to the Matrix that are from Black and/or other Minority Ethnic communities has reduced between 2019 and 2021. Disproportionality lessens when we compare the ethnicity of new additions (70% of new additions aged under 25 were Black individuals) to violent offending cohorts such as lethal barrel discharge (75%), robbery (69%), and murder (67%). 15 Graph 4: Offenders – proportion of Black individuals by offence type (under 25 cohort only) The same is true when looking at victimisation in that the percentage of Black, African- Caribbean males on the Matrix exceeds the proportion that were victims for violent crime. Graph 5: Victims – proportion of Black individuals by offence type (under 25 cohort only) 16 Threats to Life which were linked to gangs The data above shows that there is an over representation of young Black males on the overall matrix population as compared to offending and victimisations cohorts. A new aspect of this report was to explore the gang related ‘threats to life’ data provided by the MPS which includes those incidents that are most associated with serious gang- or group-related violence to explore any association with Matrix population and disproportionality. A Threat to Life (TTL) is where there is intelligence that indicates a person may be subject to a threat to take their life or a threat to cause them serious harm. This is regardless of whether they are aware or unaware of the threat and as a result of deliberate intentions or criminal acts of another. In 2021 only 27% (n=78) of threats to life incidents involved a victim or suspect with known gang links. Lambeth alone accounted for 17 (22% of total) of these incidents with the closest boroughs being Brent and Islington with 7. Due to the small sample size of Threats to Life data it is not possible to make direct comparisons to other general offending at a borough level. Of all gangs linked threats to life 32% (n=25) involved a firearm, with a further 10% (n=8) involving a knife or bladed weapon. Motives for threats to life were varied, however 28% (n=22) of gang linked threats were driven by retribution for an assault or murder, while 33% (n=26) were driven by ongoing gang rivalries. Only 5% (n=4) of threats to life related to drug supply. All known victims of threats to life were male, 74% (n=58) of which were Black, 16% (n=12) white, 6% (n=5) Asian and 4% (n=3) from other ethnicities. Overall, 68% (n=53) of victims were under the age of 25, of which just under half (48%) were Black. The oldest victim was 38 years old with the youngest being 15 years old. Similarly, the majority of suspects were male and Black; with Black males accounting for 74% (n=58) of suspects compared to 20% (n=16) being White males. 44% (n=34) of all suspects were Black males under the age of 25. Overall, 59% (n=46) of suspects were under the age of 25. The youngest suspect was 13 years old, with the oldest being 49 years old. White males aged under 25 only accounted for 11% of all suspects. Looking at Threats to life, across both suspects and victims approximately 74% were black males. This aligns more closely to the proportion of the Matrix that is Black – so may be a useful addition when exploring disproportionality. However, the small numbers of Threat to Life data and the over emphasis on certain boroughs make firm conclusion over the Threats to Life data difficult. 17 Exploring Matrix borough-level disproportionality5 Tables 7 and 8 below takes the BAME and Black under-25 populations by borough, the proportion of individuals charged with various violent and group-related offence types; and compares these to the proportion of Black/BAME individuals on the Matrix. Reading across the tables, the figures highlighted in dark grey relate to the proportions furthest from the Matrix demographic for each borough, those highlighted in yellow have the highest BAME/ Black cohort for each borough. For BAME individuals (Table 7), there is much variation between boroughs and across crime types. The proportion of BAME individuals charged rarely reflects that of the Matrix. Generally, disproportionality lessens when looking at more serious violent offending towards the right-hand side of the table, with the cohort for personal robbery being closest to the BAME cohort on the Matrix (80% of all people proceeded against for personal robbery aged under 25 are from Black and/or other Minority Ethnic communities, as compared to 84% of the Matrix cohort). However, measurement at this aggregated level of ethnicity masks key differences. 5 It should be noted that although individuals are placed on the matrix for the borough that they reside on this is not necessarily the same as the borough the gang that they are a member of is based or is criminally active. This will likely impact on the small volume boroughs such as Richmond, Kingston and Havering. 18 Table 7: For selected offence types, proportion of people proceeded against who are BAME (aged under 25 cohort)6 A higher level of disproportionality is apparent when we examine only Black individuals (Table 8). All boroughs have Black Matrix populations at least two times greater than the overall U25 population. Overall, the MPS has a Black Matrix population 3.5 times greater than its overall U25 population, and this disproportionality can be seen across all crime types. The boroughs with the largest disparity as compared to its overall Black U25 population are Richmond, Kingston, and Hounslow (although it should also be noted these boroughs have small Matrix populations). The figures highlighted in yellow in Table 8 below are the offending cohorts (aged under 25) with the highest proportion of Black individuals charged with various violent and group- related offence types on each borough. The boroughs with the highest disparity between the Matrix population and the offender cohorts for violent offending are Hounslow (56 percentage points difference between the proportion of its U25 Matrix cohort that is Black, and the drug trafficking cohort in this borough) and Bexley (52 percentage points difference from the offensive weapon cohort in this borough). As above, the cohort for personal robbery (69% of all people proceeded against aged under 25 are Black individuals) was the closest to the Black under-25 cohort on the Matrix (76%), 6 Crime types cover calendar year 2021; London Population is taken from GLA 2021 population projections; Matrix population looks at all unique individuals on the matrix in 2021 with last month on the matrix during 2021 used to determine exact profile information. Total U25 Matrix Population Dec 21 TNO Possession Of Drugs Drug Trafficking Violence Against the Person Serious Wounding Personal Property Offensive Weapon Knife Crime Knife Crime with Injury Gun Crime Islington 138 45% 54% 59% 76% 66% 55% 93% 65% 31% 48% 20% 63% 57% Camden 96 46% 54% 56% 68% 58% 57% 70% 68% 74% 60% 59% 73% Lambeth 96 56% 66% 73% 72% 80% 81% 75% 85% 86% 89% 78% 76% 94% Waltham Forest 65 62% 56% 70% 79% 64% 68% 68% 75% 68% 75% 50% 71% 79% Newham 63 80% 60% 74% 88% 74% 69% 76% 81% 74% 83% 50% 74% 91% Croydon 60 66% 63% 69% 73% 71% 70% 69% 73% 67% 67% 67% 67% 89% Barnet 58 46% 52% 64% 59% 48% 61% 93% 41% 82% 82% 60% 60% 75% Hackney 57 53% 60% 69% 88% 71% 66% 89% 82% 75% 74% 17% 76% 90% Tower Hamlets 57 69% 70% 82% 92% 72% 77% 72% 84% 81% 78% 100% 85% 83% Haringey 53 47% 50% 61% 62% 58% 42% 71% 74% 56% 43% 56% 62% 98% Westminster 49 55% 54% 70% 73% 63% 63% 83% 71% 72% 75% 50% 68% 76% Barking and Dagenham 45 64% 39% 57% 70% 55% 66% 67% 61% 60% 71% 29% 57% 86% Ealing 43 63% 61% 75% 74% 73% 70% 75% 79% 76% 64% 100% 78% 92% Enfield 41 54% 51% 47% 60% 48% 40% 74% 54% 65% 62% 73% 45% 86% Brent 38 72% 63% 78% 88% 68% 54% 92% 82% 64% 63% 82% 77% 90% Greenwich 36 51% 45% 57% 64% 48% 49% 65% 56% 48% 43% 100% 56% 93% Lewisham 32 62% 58% 71% 85% 76% 77% 82% 84% 86% 87% 100% 69% 97% Southwark 31 58% 52% 67% 65% 66% 71% 76% 75% 74% 86% 100% 66% 98% Harrow 26 72% 57% 76% 82% 68% 74% 91% 68% 73% 100% 100% 66% 88% Kensington and Chelsea 23 42% 52% 65% 61% 61% 54% 78% 71% 31% 29% 100% 66% 85% Hammersmith and Fulham 15 45% 52% 71% 79% 70% 68% 61% 66% 70% 86% 100% 67% 94% Hillingdon 15 59% 51% 60% 65% 56% 67% 79% 54% 31% 95% 100% 61% 83% Bexley 14 31% 27% 25% 21% 31% 20% 33% 46% 32% 33% 0% 25% 95% Wandsworth 14 42% 56% 76% 65% 64% 68% 100% 76% 62% 53% 17% 66% 88% Bromley 13 31% 33% 40% 50% 36% 18% 85% 44% 68% 67% 100% 30% 100% Sutton 9 37% 28% 36% 44% 38% 38% 85% 53% 56% 83% 0% 31% 90% Merton 8 46% 44% 55% 63% 49% 53% 79% 61% 56% 38% 0% 59% 100% Hounslow 7 60% 44% 64% 69% 48% 52% 63% 50% 31% 42% 63% 100% Redbridge 7 78% 64% 77% 79% 76% 58% 56% 83% 51% 71% 72% 85% Havering 4 27% 37% 35% 32% 40% 28% 54% 60% 31% 0% 25% 35% 80% Kingston upon Thames 1 41% 33% 43% 58% 38% 33% 80% 52% 0% 0% 44% 100% Richmond upon Thames 1 22% 29% 34% 29% 42% 60% 63% 56% 89% 100% 36% 100% Overall 1,215 55% 53% 65% 73% 62% 60% 80% 71% 67% 66% 61% 64% 84% Matrix BAME Borough U25 Population (% BAME) Wider Offending Violent Offending Violence - Weapons Stop and Search 19 but the representation of young Black males on the Matrix is still higher. The boroughs with no disparity between their Matrix population and the robbery cohort in their borough include Islington, Barnet, Westminster, Wandsworth, Tower Hamlets, and Sutton. However, 25 of the 32 boroughs still have a higher proportion of young Black males forming their Matrix cohort as compared to the robbery cohort. The boroughs with the highest disparity between the Matrix and the offender cohorts for robbery are Richmond (63 percentage point difference) and Bexley (62 percentage point difference). Table 8: For selected offence types, proportion of people proceeded against who are Black (aged under 25 cohort)7 Key predictors for Gangs Matrix Inclusion One of the new analytical directions within the review was to explore the drivers of disproportionality and seek to advance our understanding compared to the first annual report. Regression analysis8 was used to explore whether certain characteristics are able to predict which variables had the strongest association with being on the Gangs Violence 7 Crime types cover calendar year 2021; London Population is taken from GLA 2021 population projections; Matrix population looks at all unique individuals on the matrix in 2021 with last month on the matrix during 2021 used to determine exact profile information. 8 Regression is a technique to that seeks to estimate the relationship between an outcome variable and a series of other variables and can be used to make prediction or forecasts. Total U25 Matrix Population Dec 21 TNO Possession Of Drugs Drug Trafficking Violence Against the Person Serious Wounding Personal Property Offensive Weapon Knife Crime Knife Crime with Injury Gun Crime Islington 138 22% 44% 43% 51% 54% 48% 91% 59% 65% 48% 20% 49% 52% Camden 96 13% 32% 36% 49% 43% 49% 63% 58% 53% 40% 39% 63% Lambeth 96 39% 59% 66% 66% 74% 74% 75% 84% 83% 89% 78% 68% 94% Waltham Forest 65 24% 36% 39% 65% 49% 52% 52% 59% 56% 64% 50% 39% 70% Newham 63 23% 33% 36% 56% 50% 56% 50% 57% 57% 65% 50% 36% 81% Croydon 60 38% 49% 51% 55% 64% 63% 66% 63% 66% 67% 67% 47% 86% Barnet 58 14% 33% 36% 39% 32% 37% 86% 24% 60% 18% 40% 35% 63% Hackney 57 28% 49% 53% 65% 62% 62% 87% 76% 67% 67% 17% 57% 90% Tower Hamlets 57 11% 23% 21% 20% 27% 33% 46% 29% 27% 27% 0% 19% 38% Haringey 53 25% 45% 49% 53% 50% 33% 69% 67% 41% 26% 56% 52% 98% Westminster 49 12% 31% 36% 37% 37% 34% 59% 45% 55% 42% 25% 33% 37% Barking and Dagenham 45 34% 24% 33% 45% 46% 47% 61% 51% 38% 36% 29% 34% 82% Ealing 43 18% 28% 41% 45% 44% 44% 56% 48% 51% 52% 100% 43% 78% Enfield 41 29% 41% 38% 55% 41% 29% 71% 48% 58% 54% 73% 33% 86% Brent 38 25% 40% 50% 72% 50% 30% 73% 60% 48% 47% 82% 47% 88% Greenwich 36 32% 37% 48% 59% 44% 47% 55% 50% 48% 43% 100% 45% 91% Lewisham 32 41% 54% 62% 82% 72% 77% 78% 81% 86% 87% 100% 62% 97% Southwark 31 36% 46% 57% 60% 57% 57% 73% 65% 65% 76% 100% 56% 98% Harrow 26 15% 27% 34% 48% 34% 37% 79% 22% 50% 71% 0% 30% 85% Kensington and Chelsea 23 12% 31% 40% 34% 36% 31% 33% 53% 35% 14% 50% 36% 85% Hammersmith and Fulham 15 20% 36% 54% 55% 45% 29% 58% 58% 45% 50% 0% 50% 90% Hillingdon 15 16% 28% 30% 43% 38% 47% 45% 30% 37% 45% 67% 35% 83% Bexley 14 19% 20% 23% 14% 28% 13% 33% 43% 24% 33% 0% 20% 95% Wandsworth 14 20% 44% 63% 59% 57% 68% 96% 67% 62% 53% 17% 52% 88% Bromley 13 18% 26% 29% 46% 28% 14% 81% 42% 68% 67% 100% 25% 100% Sutton 9 12% 21% 24% 40% 32% 31% 85% 47% 50% 67% 0% 19% 80% Merton 8 15% 26% 37% 53% 33% 47% 75% 34% 38% 0% 0% 40% 100% Hounslow 7 13% 22% 28% 44% 33% 41% 42% 38% 29% 32% 33% 100% Redbridge 7 12% 23% 36% 40% 39% 31% 22% 49% 31% 57% 27% 81% Havering 4 14% 26% 25% 30% 32% 25% 49% 51% 27% 0% 25% 24% 60% Kingston upon Thames 1 6% 21% 19% 42% 22% 25% 70% 21% 0% 0% 21% 100% Richmond upon Thames 1 5% 17% 17% 16% 31% 60% 38% 50% 67% 71% 20% 100% Overall 1,215 22% 35% 41% 51% 46% 45% 69% 56% 51% 50% 53% 40% 76% Matrix Black Borough U25 Population (% Black) Wider Offending Violent Offending Violence - Weapons Stop and Search 20 Matrix9. There was a total of 24 variables applied to this model, with 14 proving significant factors. Overall, just under a third (30%) of the variation in being placed on the Matrix was explained by this model which is a reasonable proportion.10 As shown in Table 9 below, the regression model identified having been a robbery suspect (x4.8 times more likely) and offending in groups (x4.2 times more likely) were the strongest predictors of GVM inclusion followed by those aged 21-24 (x 3.9 times more likely), male (x3.8 times more likely) and a higher harm score (x3.7 times more likely). Other criteria that increase the likelihood of inclusion are repeat offenders, 18-20 age group, murder suspect and Black ethnicity. Asian ethnicity, victims of serious violence and robbery were the factors that made inclusion less likely. Table 9: Factors associated with Matrix inclusion Factor Likelihood of inclusion Previous robbery suspect x4.8 more likely Offends in groups x4.2 more likely Aged 21-24 x3.9 more likely Male x3.8 more likely Harm index score11 above the mean x3.7 more likely Repeat offenders (2+ offences) x3.0 more likely Aged 18-20 x2.9 more likely Previous murder suspect x2.3 more likely Repeat offenders (5+ offences) x1.9 more likely Black ethnicity x1.9 more likely Asian ethnicity x2.1 less likely Victim of Most Serious Violence x5.7 less likely Victim of robbery x8.3 less likely Considered from a disproportionality perspective, age and sex very much appear key predictors, and to a lesser degree Black ethnicity. However, by far the strongest variables predicting inclusion focus on aspects of the offending history of the individual – be it robbery, group offending, repeat offending and higher harm offending as measured by the Cambridge Harm Index. 9 The regression analysis covers data between January 2019 and February 2021. 10 Overall, the model has a Nagelkerke R square of 0.30. 11 Harm index score refers to the Cambridge harm index (CHI) where each offence was matched to a score on the CHI that relates to the severity of the crime – an average harm score was given to each individual within the cohort. Above average offending was measured as 1 standard deviation or more above the mean. 21 The role of Vulnerable Locations As seen above, descriptive analysis shows that young Black males are disproportionately likely to be perpetrators and victims of serious violence. However, the percentage of Black males on the matrix exceeds that of the general population, as well as those crimes most associated with serious gang or group related violence. The next section of this analysis explores some of the more structural reasons behind this disproportionality by examining the impact of social economic factors. The 2018 Review found a disproportional representation of Matrix individuals living in vulnerable locations. It found that Matrix individuals were 10.6 times more likely to live in the top 10% of vulnerable wards than the bottom 10%, and 3.5 times more Matrix individuals reside in the top 50% of vulnerable wards than the least vulnerable. When we examine levels of disproportionality within the Matrix population across the London boroughs and socio-economic measures, we can see in Table 10 below that there are that there is a negative correlation between these measures - in that, the areas with the highest disproportionality of the matrix population (as compared to the population in that borough) had lower socio-economic scores for measures of vulnerability (such as lower crime rates, lower levels of free school meals eligibility, lower levels of child deprivation, lower average income). This indicates that the disproportionality of the Matrix cohort (as compared to the local population) is most stark in the more affluent boroughs. Table 10: Correlation matrix between % Gangs Violence Matrix Black U25 and socio-economic measures Indicator Disproportionality: matrix vs population Disproportionality: matrix vs robbery cohort Disproportionality: matrix vs most similar offender cohort Crime - Average score Robbery PPAs % Black u25s Gun crime PPAs % Black u25s Offensive Weapon PPAs % Black u25s IMD - Average score Income - Average score Employment - Average score Employment rate Total Cultural Sites Free School Meals Eligible % with no Qualifications (16-64) Own Outright Rented from Local Authority or Housing Ass Disproportionality: matrix vs population 1.00 0.69 0.66 -0.47 -0.13 -0.15 -0.20 -0.54 -0.62 -0.62 0.19 -0.54 -0.62 -0.46 0.41 -0.53 Disproportionality: matrix vs robbery cohort 0.69 1.00 0.82 -0.09 -0.72 0.03 -0.03 -0.19 -0.28 -0.27 -0.11 -0.49 -0.44 -0.12 0.27 -0.25 Disproportionality: matrix vs most similar offender cohort 0.66 0.82 1.00 -0.16 -0.55 -0.31 -0.11 -0.18 -0.27 -0.27 -0.04 -0.46 -0.37 -0.17 0.27 -0.21 Crime - Average score -0.47 -0.09 -0.16 1.00 -0.01 0.35 0.62 0.90 0.87 0.85 -0.30 0.30 0.66 0.31 -0.75 0.70 Robbery PPAs % Black u25s -0.13 -0.72 -0.55 -0.01 1.00 0.17 0.21 0.10 0.13 0.14 0.36 0.17 0.18 -0.15 -0.27 0.02 Gun crime PPAs % Black u25s -0.15 0.03 -0.31 0.35 0.17 1.00 0.38 0.36 0.30 0.38 -0.11 -0.05 0.05 0.14 -0.24 0.09 Offensive Weapon PPAs % Black u25s -0.20 -0.03 -0.11 0.62 0.21 0.38 1.00 0.55 0.48 0.55 0.04 0.29 0.45 0.01 -0.55 0.38 IMD - Average score -0.54 -0.19 -0.18 0.90 0.10 0.36 0.55 1.00 0.98 0.95 -0.36 0.27 0.66 0.42 -0.78 0.74 Income - Average score -0.62 -0.28 -0.27 0.87 0.13 0.30 0.48 0.98 1.00 0.96 -0.32 0.34 0.70 0.44 -0.77 0.75 Employment - Average score -0.62 -0.27 -0.27 0.85 0.14 0.38 0.55 0.95 0.96 1.00 -0.30 0.33 0.66 0.48 -0.68 0.76 Employment rate 0.19 -0.11 -0.04 -0.30 0.36 -0.11 0.04 -0.36 -0.32 -0.30 1.00 -0.19 -0.24 -0.26 0.12 -0.43 Total Cultural Sites -0.54 -0.49 -0.46 0.30 0.17 -0.05 0.29 0.27 0.34 0.33 -0.19 1.00 0.73 0.06 -0.46 0.62 Free School Meals Eligible -0.62 -0.44 -0.37 0.66 0.18 0.05 0.45 0.66 0.70 0.66 -0.24 0.73 1.00 0.10 -0.79 0.81 % with no Qualifications (16-64) -0.46 -0.12 -0.17 0.31 -0.15 0.14 0.01 0.42 0.44 0.48 -0.26 0.06 0.10 1.00 -0.10 0.32 Own Outright 0.41 0.27 0.27 -0.75 -0.27 -0.24 -0.55 -0.78 -0.77 -0.68 0.12 -0.46 -0.79 -0.10 1.00 -0.73 Rented from Local Authority or Housing Ass -0.53 -0.25 -0.21 0.70 0.02 0.09 0.38 0.74 0.75 0.76 -0.43 0.62 0.81 0.32 -0.73 1.00 22 This is also reflected in the maps below. The boroughs with the largest disparity as compared its overall Black U25 population are Richmond, Kingston, and Hounslow (although it should also be noted these boroughs have very small Matrix populations). The boroughs with the lowest levels of deprivation are also Richmond and Kingston. The boroughs with the smallest disparity as compared to its overall Black U25 population are Tower Hamlets, Islington and Westminster. When we compare disproportionality of the matrix population as compared to the offending cohorts in each borough, there is no correlation with socio-economic measures. The only link is found with the proportion of robbery offenders that Black aged under 25 in that borough – in that those boroughs with the highest disparity between the Matrix population and the offender cohorts for violent offending (such as Hounslow and Bexley) have a smaller proportion of robbery offenders that are Black individuals aged under 25. Summary The report completes the second annual review of the MPS Gangs Violence Matrix – focussing upon the population of 2021. As a whole, there is a clear decrease in terms of overall numbers on the Matrix and these selected individuals are less likely to be young than previously found. Demographically there has been little change - the majority on the tool remain male and Black African-Caribbean - although there are signs that new Matrix additions are slightly less disproportionate and more in line with the violent offending cohort for lethal barrel discharge, personal robbery, and murder. However, given the overall reduction in volume this slight change in new additions was not enough to change the overall finding around disproportionality. Understanding the drivers of this disproportionality is a difficult and complicated task and the paper sought to advance our understanding by looking at new data such as ‘threats to life’, Disparity on the matrix (2019 - 2021 average) as compared to local population Levels of Deprivation compared to rest of England 23 social economic factors as well as statistical modelling to determine the main factors predicting inclusion onto the Matrix. The analysis demonstrates disproportionality (as compared to the local population) is most stark in more affluent boroughs and that variables relating to offending history (especially robbery) are the strongest predictors for inclusion, but the influence of ethnicity cannot at this time be ruled out. 1 | P a g e An Evaluation of Hospital Based Youth Workers Final Report March 2023 Rebecca Gurney-Read, Rebecca Barnett & Adele Harrison 2 | P a g e An Evaluation of Hospital Based Youth Workers Final Report About London’s approach to Hospital Based Youth Workers For more than five years, the Mayor’s Office for Policing And Crime (MOPAC) has contributed funding towards embedding youth services that support young victims of violence and domestic abuse within a health setting. Initially this was within London’s four Major Trauma Centres (MTCs), but it was rolled out and piloted in Accident & Emergency (A&E) departments to introduce hospital-based youth workers within this environment. A total of £2.5 million from MOPAC, the Mayor’s Young Londoners Fund, and the Violence Reduction Unit (VRU) was used to develop youth services within A&Es, situated in geographical areas experiencing high levels of knife crime. Services were delivered by three providers - across seven A&E sites: St Giles delivered within Whittington and Newham; Oasis delivered within North Middlesex and St Thomas’ and Redthread delivered within the Queen Elizabeth, Lewisham and Croydon. This is a final report from the performance and process evaluation, summarising findings from the two-year implementation period (April 2020 – end of March 2022). Learning from London’s Service Design The evaluation encountered several challenges that have limited the conclusions which can be drawn from the service and prevented a robust examination of impact • The evaluation has been impacted by poor data quality and differences in recording (i.e., gaps in outcome data, or missing data for those not engaged). • The need to increase data capture and sharing, including motivation/expectation data and repeat presentations to the A&E (and wider NHS) would be crucial. • This report has raised a need for robust evaluation - whilst the model has promising aspects, there continues to be no robust evidence base for the wider approach. Engagement rates have improved from year 1, but withdrawal from service is still high • Across the 2 years, providers offered the service to a total of 1,995 young people, with an overall uptake of 45% (n894). However, of those who initially engaged with the service, 45% (n399) subsequently withdrew service before completion1. • Service engagement differed across the providers - Redthread had the highest initial rate (62%) followed by St Giles (47%) and Oasis (30%), although the proportion of young people ‘completing’ the service was similar across all providers – with an overall completion rate of 17% (n346) across the 2 years2. This is noteworthy given the local context differs across neighbourhoods and delivery models. • Service providers engaged with more people in year 2 (n469) compared to year 1 (n324)3, thought in part to be a result of the increased physical presence of youth work teams in A&E sites, as the impact of covid-19 lessened. 1 These figures (from both years of provision) have been based on aggregate data given to us by providers. 2 The remaining 149 are a mixture of missing data (33) and cases that are still live (116). 3 The total figure (n793) engaged with for yearly comparisons is different to the overall total (n894) due to missing individual level data from providers. 3 | P a g e • Of the young people who engaged in the service, 32% were not known to statutory services prior to the intervention. • Individuals most likely to be offered the service were: Aged 15-17 (42%); Male (60%); Black (40%); Attending school, college or university (72%); Living with family, parents, carers, or guardian (81%). • Those who completed the service were most likely to be initially assessed as medium risk (49%). When comparing who did and did not engage4, 15–17-yr-olds were most likely to be offered the service, but a higher proportion of 18-22yr olds took up the provision. Service providers have addressed many of the challenges from year 1 implementation • The service encountered many of the usual implementation challenges in year one – awareness of the service; buy-in from crucial advocates (e.g., clinicians); technology and data governance challenges; and understanding how practically to deliver the service (e.g., the importance of physical proximity to A&E). The pandemic accentuated difficulties and time to rectify solutions. • Year two addressed many challenges, including better access to hospital IT systems and data sharing; dedicated spaces to work; and a greater physical presence. • Addressing challenges has reinforced support for the service – with clinicians and strategic partners agreeing the youth workers are valued and viewed as part of the hospitals extended team, as well as improving both awareness of contextual safeguarding and aiding a more multi-agency holistic approach. Some delivery challenges were not able to be addressed, for example: • A&E Staff turnover - has meant youth workers need to continually promote the service to receive buy in. • Youth worker staff shortages impacted delivery, some teams were not able to cover critical hours in A&E or due to clerical time had less time on the frontline. These staff shortages impacted data coordination, quality and resource allocation. • Getting referrals & engaging young people over 18 - youth workers reported barriers gaining consent / developing trust for this age group compared to those younger. It was also felt clinicians were more likely to refer under 18s due to safeguarding issues and a lack of service awareness in the hospital adult teams. • The continued need for training - due to the high staff turnover in A&E. • Commissioners, service providers and key stakeholder should consider if there is more that can be done regarding the key enabler of service ‘awareness’ (i.e., a dedicated resource), as reliance on youth workers to continuously promote the service directs resource away from their key role. The service and reported ‘tailoring to need’ are seen as positive • There was positivity towards the programme across service providers, stakeholders; partners; young people and their families. • In the views of youth workers, their tailored work was especially welcomed, reporting they adopted a flexible approach, personalised to the individual and their needs, considering interests, and sequencing support. • It is hoped future evaluation can collect more granular data regarding need and the services matched to them, to strengthen learning about this important aspect. 4 Just for RedThread and St. Giles. 4 | P a g e There were numerous significant improvements across perception measures including reduced risk, risk of harm and safety • The top 3 reasons for hospital presentation were: assault (53% n724), mental health (19%, n357) and substance issues (7%, n97). The top three reasons for referral were: assault (46%, n633), risk of harm (24%, n327), violence (18%, n250). • For individuals who engage with the service, perceptions of risk were assessed by the provider at the start (n582), end (n393) and during a follow up (n56). • For individuals who5 were assessed at start and end, the level of risk reduced, from 34% (n134) high-risk to only 7% High risk (n27) at final assessment. • Services appear to be helping young people to feel safe6 – for those completing the service (n141), there were improvements across self-reported feelings of safety, support network & well-being scores: o All measures significantly increased7, with the greatest change in “I could ask professionals for help if I needed it” and “I trust services could keep me safe” (increasing from 6.9 to 8.3 and 7.0 to 8.3 respectively on a 1-10 scale). o Self-reported measures of numbers of exposures to violence significantly decreased at each stage of assessment (-0.5 for witnessing violence from 2.5 to 2 and -0.4 for personal involvement in violence from 2.2 to 1.7). o Average hospital attendance numbers for violence in last 6 months significantly decreased by -0.5 from 0.7 to 0.2 respectively). o Across all measures (safety; support networks; wellbeing; and violence exposure), there were improvements for individuals initially assessed as high- risk and those assessed as medium/low risk. • Taken as a whole, the perception results are positive. However, due to a range of factors (i.e., lack of control group, sample bias, data issues) the evaluation cannot attribute any of these changes confidently to the programme. Conclusion This report completes the evaluation of the hospital-based youth workers programme, detailing what has been the largest attempt to deliver the teachable moment in England & Wales. Considerable learning has been generated that can inform future decision making. These can be grouped into: • Ensuring implementation - implementation matured over the course of the programme, any future rollout should anticipate and make contingency plans to address these routine issues. • Addressing the withdrawal challenge - commissioners and service providers should consider options to explore the retention of individuals. The ability to do this is grounded in better data collection and quality, as well as new research on attrition. • The need for robust assessment of impact - Future delivery and/or roll out should focus on data capture, access and quality, as well as striving to achieve a robust impact assessment to enable an understanding of which aspects of the service are working and for who to benefit the wider evidence base. 5 Those who complete will likely be the most engaged, the most motivated which may confound results here. 6 Those who complete will likely be the most engaged, the most motivated which may confound results here. 7 All measured through Wilcoxon test for between sample differences. 5 | P a g e Contents London’s approach to Hospital Based Youth Workers ........................................................ 6 Hospital Based Youth Workers: Service Design .................................................................. 8 Service Evaluation ............................................................................................................ 9 Methodology ......................................................................................................................... 9 Data considerations ............................................................................................................. 11 Analysis ................................................................................................................................ 12 Results from Service Delivery .......................................................................................... 13 Who is using the service? .................................................................................................... 13 What does service engagement look like? .......................................................................... 15 Service Delivery Improvements ........................................................................................... 20 Continued Service Delivery Challenges ............................................................................... 22 Perceptions of Risk & Safety ................................................................................................ 25 Discussion....................................................................................................................... 28 Appendices ..................................................................................................................... 30 Appendix A: Base sizes for individual level recorded data .................................................. 30 Appendix B: Base sizes for outcomes data .......................................................................... 31 Appendix C: Data issues and reporting considerations by service providers ..................... 32 Appendix D: Breakdown of qualitative interviews conducted by ORS ............................... 33 Appendix E: Start and End risk scores by provider .............................................................. 34 Appendix F: Start and end mean scores & mean score change .......................................... 35 Appendix G: Mean score change by Service Provider and Initial Risk Level ....................... 35 Appendix H: Borough of residence of young people offered the service (percentage of young people per borough across all service providers) .................................................... 36 Appendix I: Operating Models from service providers ...................................................... 36 References ...................................................................................................................... 38 6 | P a g e London’s approach to Hospital Based Youth Workers Since 2015, the Mayor’s Office for Policing and Crime (MOPAC) has contributed funding towards embedding youth services to support young victims of violence and domestic abuse (DA) within a health care setting. Initially this was within London’s four Major Trauma Centres (MTCs), however the mayor’s knife crime strategy published in 2017 included developing youth services within Accident and Emergency departments (A&Es), situated in geographical areas experiencing high levels of knife crime1. The decision to fund Youth Services in Major Trauma Centres and A&E departments in London was made against the backdrop of increasing knife crime and violence figures across the UK. In England & Wales, recorded knife crime increased year on year from 2013-2020. During the period April 2019 – March 2020 the total number of knife related offences was almost double (95% higher) than in April 2013 – March 2014.2 The number of homicides amongst young people (Aged 5-24) had also increased significantly prior to the start of the pandemic. From April 2015-March 2016 there was a total of 99 homicides, while in 2016-17 and 2018-19 this increased to 181 both years.3 From a London perspective, prior to the start of the Covid-19 pandemic knife offences had been increasing every year, for the previous five years. During the early part of 2020 (January- March) knife crime was 7% higher on a rolling year average basis, than the same period one year earlier in 2019, and over 60% higher than the same period 5 years earlier in 2015.4 The number of under 18s admitted to hospital with knife injuries also rose by a third between 2013-14 and 2017-185. Although knife offences decreased significantly during the pandemic, as of early 2021 they have been on an upward trend 4. The number of teenage homicides in 2021 were also the highest on record at 30 incidences6. In London, there are links between serious youth violence and group offending, with offences and victimisation specifically linked to tensions between gang members or affecting those on the periphery of gangs. MOPAC’s Serious Youth Violence Problem Profile (SYVPP) found over half of all teenage homicides in the last 3 years had links to gangs. These findings indicate there is much more to do to keep London’s young people safe6. The GLA city intelligence Unit (CIU) and MOPAC Evidence and Insight Unit profiled young victims (aged between 1-24) of serious violence in London and found rates of victimisation were highest amongst those aged 20-24, except for knife crime where rates were highest for those aged 15-19. Two thirds of victims of the ‘most serious violence’ in London were male (66%; female 34%). Young Black Londoners are disproportionately more likely to be victims for all types of serious violence7, with the MOPAC SYVPP indicating over half of weapon enabled robbery (59%) and homicide (65%) suspects were Black. Victim profiles were more diverse (i.e., 56% of weapon enabled robbery victims were white) but most youth homicide victims were male (93%) and Black (61%)6. The House of Commons Serious Youth Violence report (2019) identified strong evidence to link serious knife crime and serious youth violence with deprivation and vulnerability, exacerbated by cuts to youth services; reduced police budgets; a growing number of children being excluded from school and taken into care; and a failure of statutory agencies to keep young people safe from exploitation and violence.5 Studies have shown that violent injury is reoccurring and exposure to violence can increase the probability of an individual being both 7 | P a g e a perpetrator and victim of future violence 8 9 10 11. Various linked factors such as substance use; poor school achievement; and mental/physical health concerns have also been found to increase likelihood of recurring violent injury12 13 14. Similarly, the MOPAC SYVPP demonstrated that deprivation metrics (i.e., IMD, food insecurity); school suspensions/absence; and low youth employment were all predictors of most serious youth violence6. An A&E hospital setting provides a unique opportunity for intervening with youth injured through violence before a child or young person reaches major trauma centres or before they are known by police or local authority services. Engagement can take place at the ‘teachable moments’ which are defined as ‘naturally occurring life transitions or health events thought to motivate individuals to spontaneously adopt risk-reducing health behaviours’15. These situations have been recognised as the best opportunity to engage and where there is the greatest chance to changes lives. Interventions in a hospital setting can include mentoring; counselling services; individual or family assessment; and onward referral to services. Promoting change in health behaviour during teachable moments has been explored and evaluated in a wide range of contexts including, sexual behaviours and HIV prevention; alcohol consumption; injury prevention; general lifestyle changes; smoking cessation; suicide prevention; and cancer screening 16 17 18. In recent years there has been increasing interest in the youth violence context, particularly the role of youth work in Emergency Departments (ED) to take advantage of teachable moments and help change behaviour. Most models include contact at the initial point of entry to the hospital, but an additional longer-term effort to network young people into other kinds of support to reduce repeat presentation 19. The NHS Violence Reduction Programme produced a guide for implementation of in-hospital Violence Reduction Services in 2022. A key contributor to this guide was the service specification on A&E commissioning that MOPAC developed with clinicians. The guide supports health care professionals working in partnership with local authority and third sector organisations and provides background information on public health approaches to violence, an understanding of in Hospital Violence Reduction Programmes and key recommendations to support service implementation. 20 While current (primarily American) academic results have been inconclusive on the impact of youth workers in hospital settings, there has been some emerging evidence for crime reduction and positive responses from the young people involved in these initiatives. In the UK, hospital-based youth violence intervention programmes are gaining tractionError! B ookmark not defined. and a limited number of studies have described successful implementation of youth services in hospitals – including uptake of services; reduction in risk factors; and positive response from young people21 22 23. The most widely used model in the UK is run by Redthread, a third sector organisation that embeds Youth Workers within Major Trauma Centres (MTCs) to work with young victims of violence. MOPAC Evidence and Insight (E&I) team evaluated the Redthread Youth Violence Intervention Programme (YVIP) between April 2016-March 2017 and found tentative indications of benefit, including reduced risk scores for service users and positive response to the service from hospital staff and service users. 24 The St Giles Trust and the Oasis Programme are doing similar work in the UK with an evaluation of the Oasis Youth Support intervention 8 | P a g e service at St. Thomas’ hospital in London by Middlesex University identifying similar outcomes between service implementation and the benefits to young people’s lives 25. Nevertheless, despite attempts to expand the evidence base, there is still a lack of robust evidence regarding success of the approach19. This report details the performance and process evaluation from the two-year implementation period (April 2020 – end of March 2022) of London’s Hospital Based Youth Workers service. Whilst the crime picture across London temporarily changed during the evaluation due to the covid pandemic, knife offences have begun to increase, so the need to evaluate the effectiveness of specialist violence prevention and intervention services has never been greater. Hospital Based Youth Workers: Service Design Combined investment from the Mayor in 2020/21 and 2021/22 for Hospital based youth work sits around £4m. To extend youth provision already in place in the four Major Trauma Centres (MTCs) and in North Middlesex (Oasis), Homerton (Redthread), and St Thomas (Oasis) hospitals, a demand analysis was undertaken by the GLA city intelligence Unit (CIU) to establish which A&E sites in London would be most suited for embedded youth workers. Several demand factors were considered including: • volume of A&E incidents, caused by or classified as knife in the borough where hospital is based; • volume of A&E attendees, recorded by emergency department staff / London ambulance service; • proximity to Major Trauma Centres (MTCs) and other type 1 Emergency Departments; • social demographics (Index of Multiple Deprivation and percentage of young population); and • the presence of active gangs in the borough. Based on the results, five new priority sites were selected to receive funding for services: • The Whittington Hospital A&E; • Newham University Hospital A&E; • Croydon University Hospital A&E; • University Hospital Lewisham A&E; and • Queen Elizabeth Hospital. MOPAC already provides additional funding to three A&Es (St Thomas’ Hospital A&E, North Middlesex University Hospital A&E and Homerton University Hospital A&E). However, Homerton was excluded from the evaluation, as MOPAC only funded a small part of the service. Services were therefore delivered across a total of 7 A&E sites, by three providers. The below figure indicates the A&E location and delivery provider. The underlying service rationale is that delivering a holistic intervention, through embedding specialist youth workers within A&E’s; working with hospital staff; partners; and families will aid in delivery of ‘teachable’ moments and support victims of knife crime as outlined in the Mayor’s Knife Strategy. Figure 1: Location of the A&Es and youth work service providers 9 | P a g e Specific service aims were to: • Improve identification of young victims of violence or young people at risk. • Improve engagement of young people with available support. • Reduce risk amongst young people, (including short-term and long-term). • Improve feelings of safety & wellbeing amongst young people. • Improve networks of support & relationships amongst young people (including peers / family). • Reduce self-reported exposure to violence amongst young people. • Reduce number & frequency of young people coming to attention of authorities (e.g., hospital, police) in relation to violence. • Improve knowledge & awareness of contextual safeguarding/service amongst hospital staff. • Improve knowledge and awareness of contextual safeguarding/service amongst wider stakeholders. • Improve consistency and awareness across trauma-informed services. • Improve understanding of the impact of the youth work offer. Although services broadly follow a comparable design, they were given the ability to apply local flex to the operating model, to fit with the working practice in each NHS trust. Whilst standardisation of measures and definitions were given to service providers by MOPAC these were interpreted differently by each organisation based on their own working practices. Therefore, throughout the report, comparisons across providers should all be viewed with caution (see Appendix I -Operating Models). Service Evaluation The Evidence and Insight (E&I) Unit - MOPAC’s in-house social research and analytical team - were commissioned to undertake a multi-year evaluation of the Hospital-based youth services. The evaluation intended to look at learning from each provider, as well as overall delivery across a 2-year period (2020-2022), to understand the unique operating context; data on eligibility/throughput; demographics; onward referrals; risk assessments; training sessions provided to hospital staff; and the impact of covid on the service. Service implementation will be a focus, including challenges or barriers to delivery and how these have changed or been overcome from the first to second year. This final report focuses on the performance of the service and process learning gained from a range of sources. It was not possible to robustly evaluate impact due to data quality. Methodology The current report draws upon a mixture of quantitative data analysis and qualitative feedback from service providers. Service User Data Data collected by the three service providers (Redthread, St Giles and Oasis) between April 2020 and end of March 2022 was analysed at the end of the second year of intervention (see analysis for more information). 10 | P a g e The results are taken from aggregate and individual level performance data provided by each service. The level of data recorded across performance measures differed by provider (see data considerations), therefore analysis has been limited by data availability and not all providers could be included in analysis for all measures. Service User Outcome data Self-assessed outcome and well-being measures were recorded by individuals at the start of intervention, end of intervention and at six-month follow-up (see data consideration section below for limitations). Individuals assessed their feelings of safety in different environments and their trust and relationships with services, professionals, family and peers on a scale of 1-10 and their well-being (e.g., feeling optimistic, useful, relaxed, close to others) on a scale of 1-5. They also recorded their exposure to violence (both witnessing and involvement) on a scale of never – often. Finally, the numbers of hospital attendances by the young person for violence in the last 6 months was recorded. Service Data – Training for staff Data on hospital staff training, delivered by the service providers was collected. Most training was directed at doctors and nurses. Redthread provided individual level data on 84 training sessions but indicated that a further 256 training sessions were delivered with no details available. St Giles provided individual data for 27 training sessions in total, Oasis North Middlesex recorded that 15 training sessions were delivered in total and Oasis Waterloo indicated that 29 recorded sessions were delivered in total. A survey was distributed to clinicians to capture views on the training provided by youth workers. This survey was used to examine how familiar clinicians were with the service and how they can refer as well as the importance of the service. They were also asked whether they had received training from youth workers on a variety of different subject areas and if they had whether they thought this had improved their knowledge. This was a voluntary survey which was distributed out to clinicians across the hospital sites where contact details were known to MOPAC. A total of 29 clinicians responded to the survey. The total population of clinicians is not known, so it is not possible to pass comment on the response rate to the survey. The information provided by the clinicians was further supplemented by a range of qualitative interviews (see qualitative data). Qualitative Data The report includes qualitative data obtained from interviews with practitioners delivering the service; clinicians working in the hospitals; and young people and/or their parents using the service. Qualitative data adds crucial context to the quantitative - it aims to understand how embedded youth work services were being delivered in A&Es (including views on training), and to gather the views and experiences of youth workers; key stakeholders involved in the service; and service users (conducted by Opinion Research Services (ORS)). Fieldwork was carried out between February - April 2021, and April - June 2022 and consisted of over 40 in-depth interviews and focus groups during each wave. There was a mix of 11 | P a g e participants across the three providers and the seven A&E sites including: youth workers; hospital staff; representatives of partnership organisations; those in strategic roles who have been involved in commissioning, mobilising, and delivering the service; and five young people (or parents of young people) who have been supported through the service. Service providers recommended participants for inclusion in the interviews and these interviews were undertaken on a voluntary basis (giving a limitation of selection biased). Discussions were centred around participants views on the success of the interruption of the cycle of violence among young people presenting at the funded A&E departments. A full breakdown of interviews conducted by ORS can be found in the appendix D. Data considerations There were several data challenges for the evaluation, mainly variations between providers summarised below: Missing Data Missing individual level data: • Oasis did not provide any individual level data for service users that were contacted but did not engage in the service, or for ‘reasons for engagement’ and training data. • For some measures, Redthread, St Giles or both providers did not provide data for non- engaged individuals (notably Redthread did not provide Ethnicity data for non- engaged individuals). • Therefore, figures shown for key measures, such as demographics of the cohort and incident characteristics are not representative of all individuals who were offered the service, only those for whom data has been provided. • For Oasis, summary or aggregate data has been used in place of individual level data for some measures. • Comparisons between engaged and non-engaged groups have only been made for measures and providers where data has been recorded for both groups. Missing data across fields (data unknown / not collected): • There were missing fields at individual level in the datasets provided by all three service providers, where information was not collected or was unknown. • Reported proportions are based on all individuals with recorded data for that measure, excluding individuals for whom information was unknown or not recorded. • The sample base size therefore differs by performance measure. A full breakdown of the sample sizes used throughout this report is shown in appendix A and appendix B. Missing outcomes data: • Outcome measures were not recorded for every individual who engaged with or completed the service and not every outcome measure was recorded. • Redthread outcome measures have been excluded from analysis due to insufficient data provided. • Some analysis has not been possible or is indicative only, due to small or insufficient base sizes. Inconsistent recording practices Many reported measures were not recorded consistently across service providers, i.e., the response options differ and are not therefore comparable. Categories have been grouped 12 | P a g e where possible for the purpose of analysis, but some categories recorded are specific to each provider. For some measures we have not been able to make comparisons between providers and have reported results only at total level (all providers combined). Internal bias The evaluation explored a variety of perception metrics (i.e., risk and harm). Whilst this can provide useful results – given the nature of the programme, these measures were only taken from those who completed. This raises the issue of bias – those who completed are only a selection of the wider population and will likely be the most motivated. Results should be caveated as such. Appendix C provides a descriptive summary of the data provided, analysed and reported on for each measure. Analysis Salient findings are informed by aggregate and individual level performance data provided by each service at the end of delivery period. A full breakdown of the base sizes for each measure referred to below is shown in the appendices A & B. Testing has been used throughout to determine a significant change in or difference between two measurements. Where stated, a significant result is measured at 95% level of confidence (p<.05). Data analysed includes throughput of the service, demographics of the young people, incident characteristics (e.g., reasons for referral), service characteristics (e.g., nature of interventions delivered), risk assessments and outcome measures (conducted at start and end of interventions). Analysis was conducted to understand: • What the service looked like • Who accessed the service and why • Who was more likely to engage with the service • Whether there were changes in the second year of delivery • Whether the service impacted on perceived risk levels for young people • Whether the service impacted on young people’s feelings of safety, support networks, exposure to violence and wellbeing Key learning: Data quality, in addition to small base sizes, has affected the ability to draw conclusions and prevented an impact evaluation. 13 | P a g e To understand whether any individual demographic characteristics or incident characteristics were more or less likely to result in a young person initially engaging with the service or completing the service, logistic regression analysis was attempted. Due to small base sizes and missing data across providers the analysis did not produce significant findings of note and results have not therefore been included. Results from Service Delivery Across the seven A&E sites and three service providers, a total of 1,995 individuals were offered Embedded Youth Work services following referral by A&E staff.8 Of these, a total of 894 individuals initially chose to engage with the service, and 346 individuals completed the service. Who is using the service? Findings are based on all individuals who engaged or were offered the service. • Individuals - across all three providers, those most likely to engage or be offered the service were aged 15-17, male and Black, but to note there was only data available for 1,391 individuals which is a limitation for understanding the sample. • The profile of young people is different to those that engaged with the Redthread & St Giles Youth Violence Intervention Programme (YVIP) in MTCs from 2015-201724, A&E youth services appear to be engaging with a younger cohort, a higher proportion of females and a lower proportion of black young people.910 8 We are unable to provide figures for the total number of individuals who were referred to the service by A&E due to differing recording practices and definitions of referrals across service providers. It should be noted however that a proportion of those referred were not contactable, and a further smaller proportion were contactable but not eligible for the service. 9 Engaged figures have been used (rather than all those contacted and offered the service) in order to make comparisons with data reported in the MTC YVIP evaluation. 10 For St Giles a significantly higher proportion of those engaged were under 18 (70%, n=140) in comparison with the MTC YVIP (44%, n=67) and a significantly lower proportion were black (31%, n=60, compared with 53%, n=51). For both providers a significantly higher proportion of those engaged with the A&E youth services were female (42%, n=183 for Redthread and 49%, n=98 for St Giles) in comparison with the MTC YVIP (9%, n=89 for Redthread and 3%, n=4 for St Giles). Individual age figures were not provided for Redthread in the MTC YVIP evaluation report but the average age for MTC YVIP engagement was over 18 while for A&E youth services it was under 18. 14 | P a g e • Boroughs of residence - the top boroughs of residence for young people varied by provider due to differing hospital locations. At an overall level, over 50% of young people with recorded data resided in Newham (15%, n=213), Greenwich (14%, n=199), Croydon (13%, n=183) and Lewisham (10%, n=141) (see map appendix H). • Education & home - whilst most young people attended school, college or university (72%, n=456) and lived with family, parents, carers, or guardians (81%, n=639), there are still large number who do not – 30% not at school, college or university and 20% living elsewhere. • Prior engagement with statutory services - 68% (n=37611) of young people were either presently engaged, known to, or previously known to statutory services, with 32% (n=176) not known to statutory services. A breakdown relatively consistent by provider. The percentages of those who were known to statutory services did not significantly vary between the engaged (70%, n=318), withdrawn (74%, n=69) and completed (67%, n=185) cohorts. • Reasons for presentation & referral - Assault was the most common reason for presentation, accounting for over half of presentations (53%, n=724). Mental health, and substances (alcohol / drugs / OD) were the next most common, together accounting for 26% of presentations (n=354). Assault was also the top reason for referral across providers, accounting for 46% (n=633). Risk of harm and violence (domestic, weapon, non-weapon, sexual, honour-based, witnessing) were the next most common, together accounting for 36% of referrals (n=494).12 11 Base size = 555, 3 respondents were recorded as ‘other’. Due to missing data the base size did not equal the overall total number of respondents engaged (894). For further information please see missing data section. 12 Findings are based on all individuals with recorded data across service providers, including individuals who engaged with the service and those who were offered the service but did not engage. Hospital presentation reason Referral reason 15 | P a g e What does service engagement look like? Of the 1,995 individuals who were offered Embedded Youth Work services following referral across the two-year period13, 894 initially engaged with the service and 399 (44.6%) individuals withdrew. Within the relevant literature, withdrawal rates in similar hospital based scheme were not widely stated, those that were tend to have a lower attrition (3-11%)22 25 than this programme, but with the current data there is no way to unpick a cause. This withdrawal rate may not be surprising given the cohort services are working with, however future work should explore what best supports retention or reasons for disengagement (e.g. the amount of waiting time to receive the service). When comparing who did and did not engage14, it was found although 15–17-yr-olds were most likely to be offered the service, a higher proportion of 18-22-yr-olds took up the provision. This uptake of the service provision in the 18-22 age group reflects the age/crime curve where engagement in crime decreases with age in adulthood26. 13 We are unable to provide figures for the total number of individuals who were referred to the service by A&E due to differing recording practices and definitions of referrals across service providers. It should be noted however that a proportion of those referred were not contactable, and a further smaller proportion were contactable but not eligible for the service. 14 Just for Redthread and St. Giles. 16 | P a g e This programme data contrasts with views from staff, who felt it was a continued challenge to engage with youths aged over 18, specifically obtaining consent and gaining their trust. Youth workers anecdotally attributed this to the ingrained lack of trust in services, whereas younger individuals or those with a parent present, who may want support for the wider family, were more receptive. The perceived challenge of engaging with the older cohort was thought to be amplified by the lack of referrals for this age group from clinicians – as they were more likely to refer under 18s due to safeguarding concerns and there was still a lack of awareness of the youth service in the A&E adult teams. ‘Trying to engage adult safeguarding teams in supporting victims of violence over the age of 18 is quite often a challenge because often those people won’t meet a threshold for adult safeguarding, despite the fact it’s a clear safeguarding need. So, there can sometimes be that conflict between third sector working, acting appropriately in wanting to safeguard someone and the barriers of where safeguarding thresholds sit.’ (Strategic Stakeholder) ‘I think the children we’re very good at picking up are under 18. The harder age group is the 18 to 25… because they have more autonomy, and often they are a bit past the intervention in some ways. By then, those behaviours, those challenges in life are so deep that it’s hard for them to get out of those…’ (Clinician, St Giles Trust, Whittington Hospital)’ Challenges gaining consent Service providers have identified several challenges facing youth workers in gaining consent from the young person to initially engage or sustain engagement with the service. According to service providers, face to face engagement between the youth worker and the young person at the critical moment within A&E often results in better engagement with the service. Face to face opportunities can be missed as… • Service operating times do not always match demand for the service – referrals made to the service by clinicians outside of youth worker operating hours often result in lower engagement rates due to a lack of physical presence and timely contact with the young person in the A&E setting. • Referrals from clinicians are not always timely and sometimes referrals are missed altogether (particularly for the over 18 cohort) – missing the best window for engagement with the young person. • The window of opportunity for face-to-face engagement in hospital A&E settings is smaller than in MTCs as patients have less serious injuries and are discharged sooner. Consequently service providers have developed ways to maximise engagement with young people by… • Flexing the service operating hours based around demand analysis and recruiting more staff to improve visibility during peak times. • Providing consistent training to clinical staff to raise awareness of the service and age range, to promote referrals, and to emphasise the importance of gaining consent for the service to contact the young person post discharge,. • Screening hospital systems for missed referrals, notifying clinicians of the eligibility criteria and the missed referral. • Partnership working across safeguarding teams in hospitals and across other professionals within the child’s network to ensure any missed referrals for under 18s are still picked up by teams and the hospital gives permission to override the lack of consent to contact for children 17 | P a g e Following a referral, young people were contacted by youth workers in hospitals or remotely - 75% of contacts by Oasis (n=187) and 77% of contacts by St Giles were remote (n=328). For Redthread, a higher proportion of referrals were made in hospital (59%, n=417). Providers delivered face to face and telephone sessions with individuals who engaged with the service. The number of telephone conversations per individual ranged from 1-34 and the number of face-to-face conversations ranges from 0-26. Oasis offered more face-to-face sessions per individual on average than other providers, while St Giles delivered more telephone conversations. Redthread also recorded details of text and email conversations with young people and family members. Due to the service implementation occurring in the covid-19 pandemic there was a higher likelihood that contact was made remotely such as on the telephone during this period. For completed cases15 (cases where the young person did not withdraw or the case was not closed prematurely) the median duration of intervention was similar across providers, suggesting consistency in the length of time young people were supported for, averaging at a of 4-6 months16. However, the range of time was extremely wide, with data suggesting between 0-524 days1718. This could be data error (as there were also ‘unknown’ intervention lengths), but it also indicates the service varies substantially depending on the individual. 15 Base size <100 for St Giles (Oasis:132, Redthread:138, St Giles:71. 16 Based on individuals with completed interventions only. 17 11 young people engaged prior to April 2020 (project start date) but received the service during the evaluation period – their data increased maximum duration of support to 1030 days. 18 Telephone range: 1-34; Face to face range – 0-26. 18 | P a g e Emotional support was the most common short-term intervention19 - offered for both Redthread and St Giles (Redthread offered to 76%, n=18820, St Giles offered to 64%, n=7021). For Redthread, Emotional and Mental health was the most common long-term intervention (offered to 80%, n=148), followed by ETE support (educational, training and employment) (offered to 63%, n=116) and Family and Peer Relationships, offered to 59% (n=109).22 For St Giles, the most common long-term interventions were ETE (26%, n=6) and Positive Activity (26%, n=6).23 Oasis did not provide details of the nature of short and long-term interventions offered24. All three service providers signposted / referred individuals to other organisations or services. Where specified25, Oasis referred individuals in 86% of cases (n=184), Redthread in 39% of cases (n=169) and St Giles in 32% of cases (n=19). Referrals were most often made to ETE (25%, n=88), VCSE sector (25%, 88), Victim services (21%, n=73), Health services (18%, n=64), Accommodation (9%, n=32) and social services (9%, n=32). 19 Differences in provider labels is due to data inconsistencies. 20 Redthread recorded full details of long-term interventions offered, with some young people being offered multiple interventions. 21 St Giles recorded only one short term intervention was per person. 22 Redthread recorded full details of long-term interventions offered, with some young people being offered multiple interventions. 23 St Giles recorded only one long term intervention was per person. 24 Base size <100 for st Giles (Redthread:184, St Giles:23). 25 Base size <100 for st Giles (Oasis:213, Redthread:433, St Giles:59). 19 | P a g e Engagement differed across providers - when looking at the number of young people who were eligible and were offered the service, Oasis had the highest figure (n861) (see figure 2). However, when comparing engagement rates Oasis has the lowest at 30% and Redthread the highest rate of 62% (see figure 3). Whilst potentially interesting, as outlined earlier, services had somewhat different approaches to design, so caution should be used when comparing across in this way. Redthread had a larger criterion compared to the other two providers and therefore their engagement levels are considerably higher. Nevertheless, despite differences in engagement rates, the percentage of young people who completed the intervention is similar across providers26. Whilst the provider interpretation of ‘initial engagement’ could still play a role, the data also suggests Redthread had a higher dropout rate (Redthread had a significantly higher proportion of individuals who were recorded as having initially engaged and then withdrawn than Oasis or St Giles). When attempting to understand ‘dropouts’ or disengagement with the service, most users were recorded as ‘due to personal choice’, with young people stating they simply decided not to27. Additional reasons highlighted by providers were around losing contact with the young person28; or the intervention was not required due to multiple professional involvement29. Future work should examine the underlying reasons for non-engagement (i.e., more in-depth work with staff, youth, wider services), so providers can provide a more targeted approach to sustain service delivery. 26 completion rates based on closed cases only, excluding live cases. 27 Redthread: Did not want to engage 48%, n=129, St Giles: Didn’t need the service 52%, n=120, Oasis: Declined/dropped 26%, n=29. 28 Redthread: 21%, n=56 and Oasis: 20%, n=22. 29 St Giles: 15%, n=35. Delivery period Redthread St Giles Oasis Total Year 1 186 75 63 324 Qtr2-20 6 12 30 Qtr3-20 35 10 17 79 Qtr4-20 67 19 20 126 Qtr1-21 84 40 14 152 Year 2 247 97 125 469 Qtr2-21 89 37 37 200 Qtr3-21 41 16 38 133 Qtr4-21 65 18 23 129 Qtr1-22 52 26 27 132 Figure 3: Young people who engaged with the service, and completed the service by provider, April 2020 – March 2022 *Completion = completed the intervention / contacted and offered the service As of data collection in June 2022, Redthread had 28 interventions that were still live, St Giles had 24 and Oasis had 64. Table 1: Number of engaged young people by quarter Figure 2: Young people contactable and offered the service by provider, April 2020 – March 2022 As of data collection in June 2022, Redthread had 28 interventions that were still live, St Giles had 24 and Oasis had 64 62% 47% 30% 45% 20% 17% 16% 17% 0% 20% 40% 60% 80% Redthread St Giles Oasis Total Engagement rate Completion rate* 704 430 861 0 200 400 600 800 1000 Redthread St Giles Oasis 20 | P a g e When assessing delivery across the years, service providers engaged with more people in year two compared to year one30. Reasons for this could be the unblocking of multiple service implementation challenges (see service delivery challenges), including the increased physical presence of youth work teams in A&E sites, as the impact of covid-19 lessened. Analysis explored whether there were differences between those who engaged in the first year of service delivery (April 2020 – March 2021) and the second year of delivery (April 2021 – March 2022), with a key difference found to be a significant increase in the proportion of 11–14-year-olds engaging in the second year of delivery across all providers. This could indicate a focussed improved of targeting of this younger group over time31. Service Delivery Improvements Most programmes will likely encounter implementation challenges when embedding a new service and the A&E youth workers project was no exception. The initial set-up was disrupted in year one by covid-19, although year two saw many improvements and staff were generally positive about the scheme. Although some long-standing challenges were not able to be remedied. One of the largest blockers in the first year was a lack of dedicated spaces for youth work teams to talk within the A&E departments. With a more prominent presence and a greater buy-in at a trust level (compared to this point last year), it was felt this issue has largely been resolved, with processes, contracts and dedicated spaces now in place to ensure the youth teams were fully embedded in the hospital environment. The linked process barriers around accessing hospital systems; data/information sharing and governance, thought to effect referrals to the service have also been resolved, leading to a freer exchange of information and more joined up approach between youth workers and hospital staff. In year 2, youth workers were able to access the hospital systems to identify young people that may be eligible for their service. 30 The total figure (n793) engaged with for yearly comparisons is different to the overall total (n894) due to missing individual level data from providers. 31 11–14-year-olds: 22% of those who engaged in Year 1 (n=72), 29% in year 2 (n=134). Key learning: Engagement rates have improved from year 1, but withdrawal from the service is still high ‘We had some real issues with the data protection side of things and barriers originally, and we managed to successfully overcome that.’ (Clinician, St Giles, Whittington) ‘I think we’ve done really well, especially launching it in the middle of the pandemic, having a backlog of referrals, having to adapt to the stresses of COVID, the limitations being created, having to work with our young people initially.... able to maintain really good numbers, we’ve had some really good success stories as well.’ (Youth worker, Redthread, Lewisham Hospital) 21 | P a g e A key finding from the first-year report was the struggle to implement the service due to the covid-19 pandemic. To illustrate, delivery models had to change and ‘hot and cold zones’ set up in response to the pandemic limited the youth workers access to hospital space. However, the relaxation of COVID restrictions allowed for an increased physical presence - therefore more opportunities to engage with young people & hospital staff. The increased physical presence of youth work teams has allowed coverage across a range of hours, to see young people when they are most likely to present in A&E departments. As a result, most survey participants felt the youth work teams had successfully embedded in each site and commented on their ability to ‘catch’ the young person at the teachable moment, due to being in the hospital setting. In year one, there was a lower than predicted referral rate, assigned in part to the lack of physical presence of youth workers. In year 2, youth workers have been able to take advantage of both formal and informal interactions with clinical staff (e.g., delivering more face-to-face training), to promote awareness of the service. Examples of training clinicians was identified as: speaking in safeguarding sessions; presenting at ward rounds; and ensuring new staff are made aware of how to facilitate the referral process. Youth workers felt training was still crucial for buy in among clinical and hospital staff. Clinicians and youth workers now feel referral numbers have increased or remained consistently high in most hospitals. Hospital staff feel youth work teams are more visible and have embedded better since last year – something reflected in the feedback that the ‘champions’ identified in the year one report, required to promote the service, have been needed less. In addition to a greater awareness and buy-in of the services, an array of flexible referral pathways were developed, to maximise accessibility for hospital staff. Most interviewed staff/stakeholder said the referral process works best using several different options to receive and seek out referrals, such as a blend of email, phone call, face to face and paper options. Results from a clinician’s survey revealed most think the service is very important; that they ‘It’s very good for visibility of the service. We rely on clinicians to refer into our service a lot, especially whenever we’re not here or out of hours…’ (Youth Worker, Redthread, Queen Elizabeth Hospital) ‘…because you’re there for them in that moment you become someone they can trust, and you can really build that relationship really quickly and just create better rapport and better engagement across the board.’ (Youth worker, Redthread, Croydon Hospital) ‘They’ll take referrals any which way; they can be referred on our computer system, they can be referred in person, they can be referred by a phone call. They were very much open.’ (Clinician, Redthread, Croydon University Hospital) ‘It’s a vital service which can change the course of young peoples lives for the better.’ ‘They are an excellent resource that we are patients are very lucky to have access to.’ ‘This has been an invaluable service to the department and local area.’ (Comments from clinicians from the survey) ‘The staff are more engaging with them, and they are part of the team. We refer to them, we speak to them, they are part of our safeguarding meetings, our handovers and our huddles.’ (Clinician, Redthread, Lewisham) 22 | P a g e understand the criteria for referring a young person; that most had made a referral; and many disclosed they were very satisfied with the process. As a result of improvements and changes to covid practices, during year two, many interviewed staff and stakeholders were satisfied with delivery progress and commented the service is in line with what was originally planned. There was praise for the flexibility of the teams, particularly during the pandemic, and that youth workers were friendly, had a strong work ethic and added value. Continued Service Delivery Challenges Whilst awareness and buy-in have improved, the greatest challenge continues to be around high A&E staff turnover. Interviewees still expressed concern over the continued need to build awareness of their work, with rotating staff effectively ‘taking their connections with them’. The fast staff churn means youth worker resource is being diverted from their front-line delivery role. Many youth workers said that clinician training is an on-going, regular process to secure buy-in, as the make-up of ward staff is ever-changing. This has led to occasions when there are gaps in knowledge, with some staff concerned that training can be piecemeal and dependent on their own identification of opportunities. Some hospital staff still report an unawareness of the importance of contextual safeguarding and some staff did not recognise the need for contextual safeguarding in the 18+ age group. Nevertheless, for those who have received it, training is perceived as useful by hospital staff. A clinician survey found that 19 of the 29 clinicians said they were very familiar with the service; and 17 agreed/strongly agreed the service had improved their knowledge of identifying vulnerable people. Redthread delivered the most training sessions and 'Safeguarding YP exposed to violence' training most frequently (67 times). St Giles provided awareness of service/introduction/promotion to service training most frequently (15 times). Suggestions for new processes to increase awareness were considered, including shadowing and joint working between providers to share learning; youth workers attend council briefings; increasing the circulation of their newsletter; getting MOPAC to promote the service; improved hospital communication, including via posters, newsletters and publicity on the intranet. ‘That’s something we have to do quite regularly and hand over meetings for new staff’ (Youth Worker, Oasis, St. Thomas’ Hospital) ‘It had embedded brilliantly, until we lost all the staff. I think now we’ve got to start again and really try and work on it.’ (Clinician, Redthread, Croydon)’ ‘… We learn something new every day. To have a different perspective than a medical perspective… it gives us an understanding of what to signpost, what to see and what to look out for’ (Clinician, Redthread, Croydon University Hospital) Key learning: Service providers have addressed many of the challenges from year 1 implementation. 23 | P a g e Although the increased presence of youth teams in A&E was identified as an enabler of engagement with young people across a range of hours, the impact of youth worker staff shortages and lack of staff to cover peak times (when young people are likely to present at A&E), has still affected service delivery. There was concern some eligible young people were ‘falling through the gaps’, with fewer referrals occurring when youth workers were ‘off duty’ (i.e., nights or weekends) and those times covered by A&E bank staff (supply staff). This was also picked up by interviewed parents and young people who said, despite the positive impact the youth workers have had, there were challenges to provide support due to working practices and resource restrictions. Another ‘A&E environment’ challenge was the short amount of time the youth worker initially spends with the young person, which does not facilitate building relationships, disseminate information and promote the service. In addition, some youth workers also mentioned the amount of administration they must do (including data collection for partners working in violence reduction), takes their time away from ‘on the ground’ work. The youth workers suggested more investment in administration to reduce the burden on frontline workers. Perceptions of the Service Overall, there was positivity towards the programme across service providers, stakeholders and partners, which was mirrored in feedback from the young people and their families. In the views of youth workers, their tailored work was especially welcomed, reporting that regardless of who engaged they adopted a flexible approach, personalised to the individual and their needs, considering interests, and sequencing support (as outlined by the case studies provided below). This would involve a developed plan put in place at the beginning with specific goals related to immediate support, focusing on precise requirements, such as housing or employment. Staff also drew upon both formal and informal methods, focusing on what the young person or the wider family needs. Youth workers were reported to be ‘safe adults’ in an A&E setting, able to act as role models, build trusted relationships, and be there to handhold; guide; listen to; advocate for; and support the young person. A key role was ‘our day to day team youth workers are also covering quite a lot of the admin that they wouldn’t usually be involved with’ (Youth worker, oasis north Middlesex hospital) Key learning: Some delivery challenges are yet to be addressed. ‘We can challenge professionals…the multi-agency working has really changed…It’s a lot more advocacy work, watching the advocacy work and trying to build a picture of youth service that should be involved or who needs to be communicating more…’ (Youth worker, Redthread, Queen Elizabeth Hospital) ‘I think [name] really listened… and adapted her responses to what she was hearing from me and what she was hearing from [name]. It’s just an absolute godsend. It was just having someone else to help us at a time when we’d run out of ideas, and we just felt very very lucky’ (Parent) 24 | P a g e identified to be advocating with professionals and attending important meetings, which helped to build relationships and put young people at ease. Interviews with young people and their parents who have accessed the support similarly highlighted the positives and benefits of having a tailored individually led support programme and how diverse the support can be. This approach has been made possible by the youth workers building networks with agencies within the community to refer young people to. In the views of staff, this kind of ‘bridge’ has been missing for many young people, so this support can help break down the distrust they have with professionals. It is hoped future evaluation can collect more granular data regarding need and the services matched to them, to strengthen learning about this important aspect of service delivery. ‘She was looking at property with me, she referred me to a social worker…she chased the council to see if they could offer me a place. She tried to get him medication, she tried to get him a job. She did a lot of things for our family. Throughout coronavirus she provided food… Whatever I need, I call and she does it…’ (Parent) ‘it could be going into school with them, having the meeting in school with them, having support networks, relationships with their parents even.’ (Youth Worker, Oasis, North Middlesex Hospital) Key learning: The service and reported ‘tailoring to need’ are seen as positive. Case study A: Person A presented at A&E after experiencing bullying, anger issues and fights at her school. Initially sessions with her youth worker were focused on her self- esteem and friendships however as these sessions progressed, Person A was able open- up about her home life and the complicated relationship she has with her mother. The youth worker was able to refer Person A and her mother to Early Help to stabilise and rebuild their relationship. During Person A’s support she also disclosed sexual abuse, a toxic relationship with a boyfriend which resulted in a pregnancy and was facing a lot of allegations from her family including drinking alcohol, stealing and believing she has a personality disorder. As a result Person A’s relationship with her family completely broke down and she was at a real risk of homelessness. Her youth worker helped with accommodation, working with services including CAMHS and social care and supported her through her termination. Throughout the pandemic Oasis has maintained virtual contact supporting Person A to deal with her trauma. 25 | P a g e Perceptions of Risk & Safety Over time service providers reported that young peoples’ risk reduced 32 - for measures of overall risk, risk of harm from others, risk of harm to others and risk of harm to self, there was a significant decrease in the proportion of individuals assessed as being 'High' risk and a significant increase in the proportion of individuals assessed as ‘Low’ risk between initial and final assessment (see figure 4). It is noteworthy to highlight the split of risk levels, with over half of youth initially categorised as medium/low risk (66%) – this maybe unexpected given they are attending A&E with injuries and involvement in violence and again may relate to the potential bias as these assessments were only conducted on those youths engaging. Figure 4: Overall risk assessment scores at initial assessment and end assessment for service providers combined (1) and providers individually (2 & 3) – Engaged individuals only, includes only those that completed initial and end risk assessments Based on ‘Overall risk at initial assessment (H/M/L)’ and ‘Overall risk at final assessment (H/M/L)’ from individual level data. Excludes individuals for whom data was not recorded or was unknown 32 Most risk assessments conducted (n=582) were with young people who chose to engage in the service and analysis has been completed based on engaged young people only. When comparing start and end risk measures, analysis has been completed only on those who have completed both start and end assessment. See Appendix A for base sizes used. 39% 9% 40% 42% 78% 42% 19% 13% 18% Redthread St Giles Oasis Overall risk at initial assessment by provider High Medium Low 10% 7% 39% 10% 35% 51% 90% 58% Redthread St Giles Oasis Overall risk at final assessment by provider High Medium Low 34% 7% 48% 32% 18% 61% Initial Final Overall risk and initial & final assessment - all providers High Medium Low Case study B: Person B is a 21-year-old man who is well known to hospital services, with Emotionally Unstable Personality Disorder (EUPD), Attention Deficit Hyperactivity Disorder (ADHD) and a mild learning disability and has a history of significant risk to self. Treatment teams who work with him have been concerned about his escalating level of risk. He was referred to the service after being raped and assaulted by an older man who would sometimes supply him with drugs. Although the crime had already been reported to the police, the youth worker made a safety plan and a referral to victim support and rape crisis services to provide further support around the court proceedings. A good rapport was built with the youth worker and Person B is being helped long-term including mending relationships with his mother as well as learning to live independently and helping to improve his self-worth and confidence. Person B has said the following about his youth worker: “He is a very good and polite person, and he always makes time to speak to me and go through my problems and gives me some really good advice. I don’t really get on with men, but I have this good vibe and I feel I can trust him when we speak. He is really good at his job and I’m happy that I am working with him, he makes me feel safe and comfortable talking to him about my problems without being paranoid he will judge me. I’m happy that he is working with me, he’s great at his job.” 26 | P a g e For overall risk, these changes were also significant at service provider level33, suggesting for all three service providers, individuals were perceived to demonstrate a lower risk of harm to themselves, to others and from others after completing the service (see Appendix E for breakdowns by individual risk measure by provider). Young people who withdrew from the service post initial engagement were more likely to present as low risk at initial assessment (34%, n=45) vs those who remained engaged with the service (19%, n=65), suggesting they may have been less ‘in-need’ of the support. Services also appear to be helping young people to feel safer34 - as when comparing self- assessed outcome measures at start and end of interventions, there was a significant improvement in the average score across all safety and support network measures and most wellbeing measures. For safety measures, the greatest increase was for 'I trust that services can keep me safe' - a mean score increase of 1.3. For support networks the greatest increase was seen for ‘I could ask professionals for help if I needed it’ - a mean score increase of 1.4. For wellbeing, the greatest increase was seen for 'I've been thinking clearly' - a mean score increase of 0.8. There was also a significant decrease in the mean score for both violence exposure measures (witnessing violence and involvement in violence) and for the number of hospital attendances for violence3536. Figure 5: Mean outcome and well-being scores grouped by category at start and end assessment. Includes only individuals who completed both start and end assessments). Figure 6: Mean start-end score change for grouped outcome and well-being measures by service provider. Includes only individuals who completed both start and end assessments. Hospital / A&E attendance in last 6 months could not be compared due to insufficient sample at provider level Start-end score changes were calculated for everyone who participated in both start and end assessments. Mean score changes based on all individuals were calculated for each outcome and wellbeing measure. 33 95% confidence. 34 When comparing start and end outcome measures, analysis has been completed only on those who have completed both start and end assessment. See Appendix B for base sizes used. 35 Base size<100: 72 for hospital attendance. 36 It was not possible to break down measures to assess differences across groups (e.g. age, gender, ethnicity). 7.0 7.4 2.2 0.7 3.2 7.9 8.2 1.7 0.2 3.8 Support networks and relationships mean Safety measures mean Violence exposure measures mean How many times has YP attended hospital/A&E for violence in the last 6 months Wellbeing measures mean Start End 0.7 0.5 -0.2 0.6 1.1 1.1 -0.7 0.6 Support networks and relationships mean Safety measures mean Violence exposure measures mean Wellbeing measures mean St Giles Oasis 27 | P a g e At a provider level - improvements in mean scores from start-end assessment for all outcome measures were significant37 for Oasis and St Giles. The greatest improvement for Oasis was seen in support networks and safety measures which saw an increase score of 1.1. For St Giles the greatest score increase was in support networks, which saw an improvement of 0.7 (see figure 6, see appendix G for full breakdown by individual measure). When digging deeper into any differences across those service users initially assessed as being high-risk or those as medium/low risk, there was an improvement in score for all safety and support network measures, most wellbeing measures38 and both violence exposure measures, suggesting the service is delivering consistently regardless of risk level (see figure 9, see appendix G for full breakdown by individual measure), although these findings are indicative as only based upon very small sample sizes39. Figure 9: Mean start-end increase for outcome and well-being scores grouped by category according to initial risk. Includes only individuals who completed both start and end assessments. Start-end mean change in score were calculated for everyone who participated in both start and final assessments. Mean scores based on all individuals were calculated for each outcome and wellbeing measure. Medium and low risk have been combined due to insufficient sample sizes for individuals initially assessed as low risk. Hospital / A&E attendance in last 6 months could not be compared due to insufficient sample at provider level. Taken as a whole, the perception results are positive. However, due to a range of factors (i.e., lack of control group, sample bias, data issues) the evaluation cannot attribute any of these changes confidently to the programme. These results could be an outcome of service delivery, or a perception of reduction by the youth worker or driven by sample bias (i.e., only incorporating those youth to complete). It is hoped that future opportunities to explore more robust impact are taken up to strengthen the evidence base. 37 95% confidence. 38 For medium/low risk individuals there was a significant improvement in mean score for all wellbeing measures, for high risk individuals there was a significant improvement in mean score for all wellbeing measures except ‘I’ve been feeling close to other people’ which increased but not significantly. 39 28 individuals assessed a high risk, 86 individuals assessed as low/medium risk. Key learning: There were numerous significant improvements across perception measures including reduced risk, risk of harm and safety. 1.2 1.2 -0.5 0.6 0.8 0.7 -0.3 0.7 Support networks and relationships mean Safety measures mean Violence exposure measures mean Wellbeing measures mean HIGH MEDIUM/LOW 28 | P a g e Discussion A total of £2.5 million from MOPAC, the Mayor’s Young Londoners Fund, and the Violence Reduction Unit (VRU) was used to develop youth services within A&Es, situated in geographical areas experiencing high levels of knife crime. Running alongside the delivery of embedding youth workers, London has seen a sustained backdrop of rising youth violence – with the number of teenage homicides in 2021 the highest on record at 30 incidences27; knife offences increasing from 2021 and returning to higher than pre-2018 figures28; and links between serious youth violence and group offending, with over half of all teenage homicides in the last 3 years had links to gangs6. This restates the continued need to address immediate and future risk for London’s young people. A wealth of findings have been documented in the evaluation. Overall, it is worth reflecting upon several areas that emerged in more detail. These are ensuring implementation, addressing the withdrawal challenge and the need for robust assessment of impact. Ensuring implementation Almost all schemes will encounter implementation issues of one type of another. The Hospital Based Youth Worker programme was no exception. This is an imperative to bear in mind, as the evidence is clear - those schemes with better implementation are more likely to achieve the desired outcomes29. The importance of implementation regarding the current scheme was of course complicated by covid. It is therefore no surprise the first year did not achieve the desired programme fidelity. However, to the credit of the staff, results demonstrated that year two saw a wealth of improvements (i.e., improved awareness, buy in from staff, information sharing and data access), which by and large has resulted in staff believing the programme is now being delivered as originally intended. As outlined, implementation challenges are commonplace within programmes, and any attempt to refine or extend the current scheme should seek to pre-empt similar issues from the outset to ensure effective programme integrity. The reported effect of staff churn and the knock-on consequence to service awareness and therefore referrals rates are something for future consideration. Directing youth workers resource from their ‘key role’ is not efficient and warrants thought over a dedicated resource. Addressing the withdrawal challenge A key finding was 45% of young people withdrew from the service after initial engagement, largely attributed to the young persons ‘personal choice’. It is difficult to place this into a wider context, and within the literature withdrawal rates are not generally stated22 25. Staff were mixed on the reasons and spoke of older groups being harder to engage, presenting trust issues, whereas formal programme data outlined it was the younger groups that were less likely to engage. High withdrawal maybe somewhat expected, given the target youths are appearing within a hospital setting and likely to present a range of complex needs, for example: • Research indicates any engagement may be complicated by the overlaps between victimisation and offending (i.e., SYVPP majority of teenage victims (63%) and suspects (87%) accused of homicide had been previously arrested by police)6; • It has been found young people who are victims/offenders are likely to be socially disadvantaged across many domains (individual, family, peer, school, community); and 29 | P a g e • Potentially these young people hold negative attitudes towards police, are anti- authority and/or have an aversion towards engaging with formal services30. Ultimately, there could be a variety of reasons why certain groups may be less ‘ready’ for the teachable moment than others and it would be hoped more could be done to explore the motivations and needs of those who do and do not engage within any future programme. The ability to do this is grounded in better data collection and quality - documenting specific needs that could assist tailoring the offer and support as well as new research on attrition. The need for robust assessment of impact One theme to emerge from the wider academic literature, as well as the findings from the current research, is the lack of robust outcome analysis. Generally, research in this area documents findings from the programme implementation rather than robust ‘impact’. Likewise, the current evaluation was not able to explore impact due to issues with data quality (i.e., notable inconsistencies in recording practice; missing data for individuals who did not engage; and data in relation to outcomes for those who did engage; the impact of COVID upon implementation). Whilst the subjective measures of success reported are to be welcomed – they cannot be confidently ascribed to the service. The need for a robust impact assessment is still very much needed and future delivery / roll- out should focus on data quality to enable robust assessment of which aspects of the service are working and for who. A Pan London approach or wider roll out would provide opportunities to build such an evaluation into the design - ensuring policy and practitioner support throughout; resulting in clear outcomes; effective data capture; and the development of a counterfactual. On a related point, the need for increased data capture and access would be an essential step forward for the programme. This would not only assist future evaluation, but aid youth workers within their practice. To illustrate, capturing need; motivation; expectations of service data; and repeat presentations to the A&E; would all be crucial in understanding where a young person is in their unique ‘journey’, assisting the initial tailored pitch and the understanding if those who are refusing the service or dropping out are presenting multiple times, have higher risk or different support needs. Data would be best serviced by a bespoke case management system harmonised across service providers, or at least by a process bolstered with data quality checks and staffing. Such information could then feedback on a routine manner for accurate performance data and contract management. Conclusion The report details the largest implementation and assessment of the ‘teachable moments’ approach in a UK health setting. The research has identified numerous promising aspects - a generally positive response to the programme; implementation issues being addressed over maturation of the programme - albeit it was not possible to fully address some issues. The service was valued by providers; stakeholders; and the young people and their families who engaged. Significant positive perception findings on harm, risk and safety were reported (although we are not able to ascribe these changes confidently to the programme). It is hoped the key lessons documented will be able to inform future iterations of the programme. 30 | P a g e Appendices Appendix A: Base sizes for individual level recorded data Base sizes only shown for measures used for analysis and findings in this report. *For Oasis aggregate summary figures provided by the service provider were used to report on total contacted and referred (861) and total engaged (261) as these were understood to be more reliable for calculating engagement rates. For all other analysis included in this report figures have been based on individual level data. **Oasis Waterloo provided reasons for declined as aggregate summary data. *** Calculated based on duration between date of referral and date of closure. Base sizes for each measure reported on: excludes blank, unknown and not recorded fields Did not engage Engaged Total Oasis Redthread St Giles Total Oasis Redthread St Giles Total Oasis Redthread St Giles Total TOTAL N/A 271 230 517 233* 433 200 866 258* 704 430 1392 DEMOGAPHICS Age N/A 271 230 501 N/A 432 200 632 258 703 430 1391 Gender N/A 271 230 501 N/A 433 200 633 258 704 430 1392 Ethnicity N/A N/A 213 213 N/A N/A 191 191 245 173 404 822 Borough of residence N/A 271 227 498 N/A 432 196 628 258 703 423 1384 ETE status N/A N/A 109 109 N/A N/A 88 88 251 185 197 633 Known to or engaged with statutory / non-statutory services N/A N/A 86 86 185 182 89 456 197 183 175 555 Living arrangements N/A N/A 202 202 214 182 177 573 228 183 379 790 SERVICE CHARACTERISTICS Reason declined (non-engaged only) N/A* * 269 206 475 N/A N/A N/A N/A N/A N/A N/A N/A Time from referral to closure (completed only)*** N/A N/A N/A N/A 132 138 71 341 N/A N/A N/A N/A Reason for closure (complete, withdrawn etc.) N/A N/A N/A N/A 231 432 198 861 N/A N/A N/A N/A No. face to face sessions (completed only) N/A N/A N/A N/A 133 112 70 315 N/A N/A N/A N/A No. telephone conversations (completed only) N/A N/A N/A N/A 95 124 71 290 N/A N/A N/A N/A Short term intervention offered (Y/N) N/A N/A N/A N/A 139 433 146 718 N/A N/A N/A N/A Long term interventions offered (Y/N) N/A N/A N/A N/A 188 433 53 674 N/A N/A N/A N/A Please specify which short term interventions N/A N/A N/A N/A 36 229 109 374 N/A N/A N/A N/A Please specify which long term interventions N/A N/A N/A N/A 153 184 23 360 N/A N/A N/A N/A Referred / signposted to other organisations (Y/N) N/A N/A N/A N/A 213 433 59 705 N/A N/A N/A N/A Organisations referred to N/A N/A N/A N/A 184 150 16 350 N/A N/A N/A N/A Did the young person engage with the onward support? (Y/N) N/A N/A N/A N/A 177 33 22 232 N/A N/A N/A N/A INCIDENT CHARACTERISTICS Whether initial contact was made in hospital or remotely N/A 270 228 498 N/A 432 199 631 248 702 427 1377 Reason for hospital presentation N/A 271 228 499 N/A 433 199 632 241 704 427 1372 Reason for referral N/A 270 228 498 N/A 433 199 632 256 703 427 1386 Whether attended A&E for assault / exploitation / self-harm in last 5 years N/A 18 228 246 N/A 98 199 297 258 704 430 1392 Who referred from N/A 271 228 499 N/A 433 199 632 258 704 430 1392 Who referred from (other) N/A 166 127 293 N/A 219 100 319 255 704 427 1386 RISK ASSESSMENTS Overall risk - Start & End completed N/A N/A N/A N/A 175 150 68 393 N/A N/A N/A N/A Harm from others - Start & End completed N/A N/A N/A N/A 127 150 73 350 N/A N/A N/A N/A Harm to others - Start & End completed N/A N/A N/A N/A 128 150 74 352 N/A N/A N/A N/A Harm to self - Start & End completed N/A N/A N/A N/A 128 150 75 353 N/A N/A N/A N/A 31 | P a g e Appendix B: Base sizes for outcomes data Start & End completed Start & End completed - Oasis Start & End completed - St Giles Start & End completed - Medium / Low Risk Start & End completed - High Risk I feel supported by my family/ parents/ carers 140 71 69 86 28 I like most young people that I spend my time with 140 71 69 86 28 Young people I spend my time with have a positive influence on my life 140 71 69 86 28 I could ask my family or friends for help if I needed it 140 71 69 86 28 I could ask professionals for help if I needed it 139 70 69 86 28 I feel safe when I am at home 140 71 69 86 28 I feel safe when I am out in my local area 140 71 69 86 28 I feel safe at school/ college/ university/ where I work 123 65 58 77 26 I am aware of the spaces/ places that may not be safe for me 140 71 69 86 28 I trust that services can keep me safe 141 72 69 86 29 How often have you witnessed any form of violence 140 71 69 86 28 How often have you personally been involved in or experienced violence yourself 140 71 69 86 28 How many times has YP attended hospital/A&E for violence in the last 6 months 72 N/A N/A N/A N/A I've been feeling optimistic about the future 138 71 67 86 27 I've been feeling useful 138 71 67 86 27 I've been feeling relaxed 137 71 66 85 27 I've been dealing with problems well 136 69 67 86 26 I've been thinking clearly 138 71 67 86 27 I've been feeling close to other people 138 71 67 86 27 I've been able to make up my own mind about things 134 71 63 83 26 Base sizes only shown for measures used for analysis and findings in this report. Excludes blank, unknown and not recorded fields. 32 | P a g e Appendix C: Data issues and reporting considerations by service providers Performance measure REDTHREAD ST GILES OASIS Referred to service Figures not reported due to differing recording practices and definitions of referrals across service providers Contacted and offered the service Individual level data provided and used in reporting figures. Individual level data provided for only 16 individuals who did not engage with the service. Aggregate summary data used in reporting figures. Engaged Individual level provided and used in reporting figures Individual level data provided. Aggregate summary data used in reporting figures for consistency with above. Status (complete, withdrew, live) Individual level data provided and used in reporting figures Reasons for non- engagement Individual level data provided and used in reporting figures. Oasis Waterloo provided summary figures for reasons for non- engagement. Oasis North Middlesex did not provide reasons for non- engagement. Reasons for non-engagement were not recorded consistently across providers and lacked detail. Reasons for closure Individual level data provided and used in reporting figures. Reasons for closure were not recorded consistently across providers. Categories have been grouped for the purpose of analysis, but some categories recorded are specific to each provider. We have not therefore made comparisons between providers. Demographics Individual level data provided and used in reporting figures. Engaged and non-engaged comparisons made. Redthread did not provide Ethnicity, ETE Status, Living Arrangements or Statutory Services data for non-engaged individuals so engaged vs non- engaged comparisons for these measures are based on St Giles only. Individual level data provided and used in reporting figures. Engaged and non-engaged comparisons not possible due to insufficient data for non- engaged individuals. ETE status and living arrangements were not recorded consistently across service providers. Categories have been grouped where possible for the purpose of analysis, but some categories recorded are specific to each provider. We have not therefore made comparisons between providers. Incident characteristics Individual level data provided and used in reporting figures. Engaged and non-engaged comparisons made. Individual level data provided and used in reporting figures. Engaged and non-engaged comparisons not possible due to insufficient data for non- engaged individuals. Reasons for referral, reasons for hospital presentation and source of referral were not recorded consistently across service providers. Categories have been grouped where possible for the purpose of analysis, but some categories recorded are specific to each provider. We have not therefore made comparisons between providers. Training Individual level training data provided and used in reporting figures (one row per training session). Redthread provided individual level data on 84 training sessions but indicated that a further 256 training sessions were delivered with no details available. Individual level data not provided. Aggregate summary data used for reporting figures. Service characteristics Individual level data provided and used in reporting figures. Data mainly provided for engaged individuals across all providers. Analysis based on engaged individuals only. Details of the nature of short-term and long-term interventions offered were not recorded consistently across providers, so we have not made comparisons between providers. Redthread was the only provider to record full details of both short-term and long-term interventions for all individuals offered, with some individuals being offered multiple interventions. Organisations referred to were not recorded consistently across service providers. For St Giles, referral information was recorded for only 16 individuals. Risk measures Individual level data provided and used in reporting figures. Data mainly provided for engaged individuals across all providers. Analysis based on engaged individuals only. The number of individuals who completed risk assessments at six-month follow-up was small across all providers (56 in total). Findings relating to improvements in risk level of young people from start-end-follow-up are therefore indicative. Outcome measures Insufficient data provided. Excluded from analysis. Individual level data provided and used in reporting figures. Analysis based on individuals who completed the intervention only. The number of individuals who completed outcome assessments at six-month follow-up was too small across providers for analysis to be reliable, therefore change in outcome score from start-end- follow-up has not been included in this report. Improvements in outcome measure scores from start-end have been compared at service provider level (Oasis and St Giles) and risk level (Individuals initially assessed at High Risk and individuals initially assessed as Low Risk). Due to small base sizes for each group findings relating to improvements in outcome scores are indicative. The number of individuals with recorded data for hospital attendance (‘How many times has YP attended hospital/A&E for violence in the last 6 months’) was sufficient only at total level for analysis on change in score start-end. Improvements in this measure have not therefore been analysed at service provider or risk level 33 | P a g e Appendix D: Breakdown of qualitative interviews conducted by ORS Breakdown of youth worker teams by service provider and A&E site: Provider A&E Site No. of interviews and focus groups per site Redthread Croydon University Hospital 1 x interview with youth worker The Queen Elizabeth 2 x interviews with youth workers 1 x interview with youth team leader University Hospital Lewisham 2 x interviews with youth workers All Redthread Sites 2 x interviews with programme managers St Giles Both Newham and Whittington Sites 1 x focus group with youth work team and leam leader (6 participants in total) Oasis North Middlesex 4 x interviews with youth workers 1 x interview with project coordinator St Thomas 1 x interview with youth team leader Breakdown of frontline hospital staff by service provider and A&E site: Provider A&E Site No. of interviews Total Redthread Croydon University Hospital 1 x interview 6 x interviews The Queen Elizabeth 2 x interviews University Hospital Lewisham 3 x interviews St Giles Newham Not able to secure any interviews, one strategic stakeholder had overview of every site as a clinically based VR lead 3 x interviews The Whittington 3 x interviews Oasis North Middlesex 1 x interview 1 x interview St Thomas Not able to secure any interviews, one strategic stakeholder had overview of every site as a clinically based VR lead Breakdown of young people and parents by service provider and A&E site: Provider A&E Site No. of interviews Total Redthread Croydon University Hospital Not able to secure any interviews 0 x interviews The Queen Elizabeth University Hospital Lewisham St Giles Newham Not able to secure any interviews 2 x interviews The Whittington 2 x interviews (1 parent, 1 young person) Oasis North Middlesex 2 x interviews (2 parents) 3 x interviews St Thomas 1 x interview (1 young person) 34 | P a g e Appendix E: Start and End risk scores by provider Start and End risk scores for ‘harm from others’, ‘harm to others’ and ‘harm to self’ by provider For Oasis, there was a significant increase in the proportion of ‘Low’ risk and a significant decrease in the proportion of ‘High’ risk cases between start and end for ‘harm from others, ‘harm to others’ and ‘harm to self’. For St Giles, there was a significant increase in the proportion of ‘Low’ risk cases for ‘harm from others’ and ‘harm to self’, there was not a significant decrease in proportion of ‘high risk’ cases across any risk measure (note – St Giles had a low proportion of high-risk cases at start of intervention). For Redthread there was a significant increase in the proportion of ‘Low’ risk and a significant decrease in the proportion of ‘High’ risk cases between start and end for ‘harm from others’ and ‘harm to self’. 35% 19% 37% 32% 43% 28% 67% 39% Redthread St Giles Oasis Harm from others - initial risk High Medium Low 9% 2% 29% 4% 29% 62% 96% 69% Redthread St Giles Oasis Harm from others - final risk High Medium Low 5% 16% 19% 4% 41% 75% 95% 44% Redthread St Giles Oasis Harm to others - initial risk High Medium Low 3% 2% 11% 4% 24% 86% 96% 74% Redthread St Giles Oasis Harm to others - final risk High Medium Low 15% 28% 25% 36% 37% 60% 57% 35% Redthread St Giles Oasis Harm to self - initial risk High Medium Low 4% 5% 23% 5% 34% 73% 95% 60% Redthread St Giles Oasis Harm to self - final risk High Medium Low 35 | P a g e Appendix F: Start and end mean scores & mean score change Start and end mean scores and mean score change (start – end) for outcome measures across all service providers Outcome measure Start mean score End mean score Mean score change I feel supported by my family/ parents/ carers 6.8 7.6 0.8 I like most young people that I spend my time with 7.0 7.9 0.9 Young people I spend my time with have a positive influence on my life 7.0 7.9 0.9 I could ask my family or friends for help if I needed it 7.0 7.7 0.8 I could ask professionals for help if I needed it 6.9 8.3 1.4 Support networks and relationships mean 7.0 7.9 0.9 I feel safe when I am at home 7.8 8.4 0.6 I feel safe when I am out in my local area 6.9 7.7 0.8 I feel safe at school/ college/ university/ where I work 7.6 8.4 0.8 I am aware of the spaces/ places that may not be safe for me 7.6 8.5 0.8 I trust that services can keep me safe 7.0 8.3 1.3 Safety measures mean 7.4 8.2 0.9 How often have you witnessed any form of violence 2.5 2.0 -0.5 How often have you personally been involved in or experienced violence 1.9 1.5 -0.4 Violence exposure measures mean 2.2 1.7 -0.4 No. times has YP attended hospital/A&E for violence in last 6 months 0.7 0.2 -0.5 I've been feeling optimistic about the future 3.1 3.8 0.7 I've been feeling useful 3.1 3.7 0.6 I've been feeling relaxed 3.1 3.7 0.7 I've been dealing with problems well 3.1 3.8 0.7 I've been thinking clearly 3.0 3.8 0.8 I've been feeling close to other people 3.3 3.7 0.4 I've been able to make up my own mind about things 3.5 4.0 0.5 Wellbeing measures mean 3.2 3.8 0.6 Table 3: Start and end mean scores and mean score change by individual outcome measure. Includes only individuals who completed both start and end assessments Support networks and relationship scores are recorded on a scale of 1-10 (1 being lowest, 10 being highest). Violence exposure is based on frequency of exposure to violence and is based on a 4-point scale which has been converted into numerical scores (Never = 1, Rarely = 2, Sometimes = 3, Often = 4). Hospital / A&E attendance in last 6 months is a numerical figure for number of attendances. Wellbeing scores are recorded on a scale of 1-5 (1 being lowest, 5 being highest). Appendix G: Mean score change by Service Provider and Initial Risk Level Mean score change (start – end) for outcome measures by Service Provider and Initial Risk Level Initial Risk Level Service Provider HIGH RISK MEDIUM /LOW RISK St Giles Oasis I feel supported by my family/ parents/ carers 1.2 0.6 0.8 0.8 I like most young people that I spend my time with 0.9 0.8 0.6 1.1 Young people I spend my time with have a positive influence on my life 1.3 0.8 0.5 1.3 I could ask my family or friends for help if I needed it 1.2 0.7 0.7 0.9 I could ask professionals for help if I needed it 1.7 1.3 1.0 1.7 Support networks and relationships mean 1.2 0.8 0.7 1.1 I feel safe when I am at home 1.1 0.5 0.3 0.9 I feel safe when I am out in my local area 0.8 0.7 0.6 1.0 I feel safe at school/ college/ university/ where I work 1.0 0.7 0.3 1.1 I am aware of the spaces/ places that may not be safe for me 1.3 0.7 0.6 1.0 I trust that services can keep me safe 1.8 1.2 0.9 1.7 Safety measures mean 1.2 0.7 0.5 1.1 How often have you witnessed any form of violence -0.6 -0.3 -0.1 -0.8 How often have you personally been involved in or experienced violence yourself -0.5 -0.2 -0.2 -0.6 Violence exposure measures mean -0.5 -0.3 -0.2 -0.7 I've been feeling optimistic about the future 0.9 0.7 0.7 0.8 I've been feeling useful 0.7 0.5 0.6 0.6 I've been feeling relaxed 0.5 0.7 0.7 0.7 I've been dealing with problems well 0.6 0.7 0.7 0.6 I've been thinking clearly 0.8 0.9 0.7 0.9 I've been feeling close to other people 0.2 0.5 0.6 0.3 I've been able to make up my own mind about things 0.6 0.6 0.6 0.4 Wellbeing measures mean 0.6 0.7 0.6 0.6 Support networks and relationship scores are recorded on a scale of 1-10 (1 being lowest, 10 being highest). Violence exposure is based on frequency of exposure to violence and is based on a 4-point scale which has been converted into numerical scores (Never = 1, Rarely = 2, Sometimes = 3, Often = 4). Hospital / A&E attendance in last 6 months is a numerical figure for number of attendances. Wellbeing scores are recorded on a scale of 1-5 (1 being lowest, 5 being highest). 36 | P a g e Appendix H: Borough of residence of young people offered the service (percentage of young people per borough across all service providers) Appendix I: Operating Models from service providers Redthread 37 | P a g e Oasis St Giles 38 | P a g e References 1 MOPAC, (2017). Mayor’s Knife Crime Strategy mopac_knife_crime_strategy_june_2017.pdf (london.gov.uk). 2 Office for National Statistics (ONS), (2022). Offences involving the use of weapons, year ending March 2022 https://www.ons.gov.uk/file?uri=/peoplepopulationandcommunity/crimeandjustice/datasets/offencesinvolvin gtheuseofweaponsdatatables/yearendingmarch2022/offencesinvolvingtheuseofweaponsfinalv2.xlsx. 3 Office for National Statistics (ONS), (2022). Homicide in England and Wales: year ending March 2021 Homicide in England and Wales - Office for National Statistics (ons.gov.uk). 4 Violence is prevented and reduced dashboard, London Datastore (2022). Violence is prevented and reduced dashboard – London Datastore 5 House of Commons, Home Affairs Committee, (2019). Serious Youth Violence https://publications.parliament.uk/pa/cm201719/cmselect/cmhaff/1016/1016.pdf. 6 MOPAC Evidence and Insight Team, (2022). A Problem Profile of Violence, Gangs And Young People PowerPoint Presentation (london.gov.uk). 7 GLA Strategic Crime Analysis Team & MOPAC Evidence and Insight Team, (2021). Understanding serious violence among young people in London https://data.london.gov.uk/download/serious-youth- violence/b9b4258a-5d02-4677-9dd6-374f55fd167f/Crime%20speech%20report%20December%202021.pdf. 8 Dowd D., et al, (1996). Hospitalizations for Injury in New Zealand: Prior Injury as a Risk Factor for Assaultive Injury. American Journal of Public Health, (86), 929-934. 9 Redeker N., et al, (1995). Risk factors of adolescent and young adult trauma victims. Am J Crit Care, (4), 370– 378. 10 Cheng TL., et al, (2003). Adolescent assault injury: risk and protective factors and locations of contact for intervention. Pediatrics, (112), 931–938. 11 Cunningham R., et al, (2009). Before and after the trauma bay: the prevention of violent injury among youth. Ann Emerg Med, (53), 490-500. 12 Cunningham R.M., et al, (2015). Violent re-injury and mortality among youth seeking emergency department care for assault-related injury: a 2-year prospective cohort study, JAMA Pediatrics, (169), 63-70. 13 Sims D.W., et al, (1989). Urban trauma: a chronic recurrent disease, J Trauma, (29), 940–946. 14 Borowsky I.W. and Ireland M, (2004). Predictors of future fight-related injury among adolescents, Pediatrics, (113), 530-6. 15 McBride C. M., et al, (2003). Understanding the potential of teachable moments: the case of smoking cessation. Health Education Research, (18), 156-170. 16 Lawson P. J. and Flocke S. A., (2009). Teachable moments for health behavior change: a concept analysis, Patient education and counseling, (76(1)), 25–30. 17 Johnson S. B., et al, (2007). Characterizing the teachable moment: is an emergency department visit a teachable moment for intervention among assault-injured youth and their parents? Pediatric Emergency Care, (23), 553-559. 18 Cohen D. J., et al, (2011). Identifying teachable moments for health behaviour counselling in primary care, Patient Education & Counselling, (85), 8-15. 19 Wortley E. and Hagnell A., (2020). Young victims of violence: using youth workers in the emergency department to facilitate ‘teachable moments’ and to improve access to services, Arch Dis Child Educ Pract Ed, (106), 53-59. 20 Royal College of Nursing, (2022). In-Hospital Violence Reduction Services: A Guide to Effective Implementation, Violence Reduction Programme London, March 2022 In-Hospital-Violence-Reduction-Services- A-Guide-to-Effective-Implementation-FINAL.pdf (england.nhs.uk). 21 Travers C. and Hann G., (2018). The impact of a youth violence intervention programme on reattendance rates and young people’s wellbeing, Archives of Disease in Childhood, (103), A136. 22 DeMarco J., et al, (2016). Improving mental health and lifestyle outcomes in a hospital emergency department based youth violence intervention, Journal of Public Mental Health, (15(3)), 119-133. 23 Potter S., et al, (2016). The impact of a dedicated youth worker in a paediatric accident and emergency, Archives of Disease in Childhood, (101), A133-A134. 24 MOPAC Evidence and Insight Unit, (Unpublished). Redthread Youth Violence Intervention program, year 2 report, August 2017. 39 | P a g e 25 Middlesex University, (2016). Evaluation of Oasis Youth Support violence intervention at St. Thomas’ hospital in London, UK, Final Report 2010-2016 http://www.oasiswaterloo.org/wp-content/uploads/2019/11/Final- report-15-Nov-2016_Evaluation-of-St-Thomas-OYS-intervention-1.pdf. 26 Farrington D.P., (1986) Age and crime. In: Tonry M, Morris N (eds) Crime and justice: an annual review of research, vol 7. Chicago University Press, Chicago, pp 189–250. 27 MOPAC Evidence and Insight, (2022). A Problem Profile of Violence, Gangs And Young People PowerPoint Presentation (london.gov.uk). 28 London Datastore (2022), Violence is prevented and reduced dashboard Violence is prevented and reduced dashboard – London Datastore 29 Dawson P. and Stanko B., (2013). Implementation, implementation, implementation: Insights from offender management evaluations, Policing: A Journal of Policy and Practice, (7(3)), 289-298. 30 Ramshaw N., et al, (2018). Youth Voice Survey 2018 https://www.london.gov.uk/sites/default/files/youth_voice_survey_report_2018_final.pdf. London Gang Exit Interim Evaluation Report October 2017 Anna Romaniuk & Tom Davies MOPAC Evidence and Insight April 2016 Julia Yesberg, Liam Fenn & Paul Dawson MOPAC Evidence and Insight Contents Executive Summary ............................................................................................................................... 1 1. Introduction ...................................................................................................................................... 3 Background .......................................................................................................................................... 3 London Gang Exit Overview ................................................................................................................. 3 Process model ...................................................................................................................................... 4 Evaluation overview ............................................................................................................................. 5 2. Performance Monitoring ................................................................................................................. 6 Cohort demographics ........................................................................................................................... 6 MPS Gangs Matrix ................................................................................................................................ 6 Referrals and interventions ................................................................................................................... 7 3. Process Findings ............................................................................................................................... 8 LGE model and process ........................................................................................................................ 8 Partnership working and data sharing .................................................................................................. 9 Staff training ........................................................................................................................................ 9 Staff views on client impact ................................................................................................................. 9 Client perspectives of LGE .................................................................................................................. 11 4. Impact on offending: Preliminary findings ................................................................................. 12 Offending history ............................................................................................................................... 12 LGE offending and victimisation (before and after referral comparison) ........................................... 12 Distance Travelled Assessment Tool ................................................................................................... 15 5. Discussion ........................................................................................................................................ 16 Appendices .......................................................................................................................................... 17 References ........................................................................................................................................... 22 1 Executive Summary London Gang Exit (LGE) is a £1.5M multi-agency intervention, jointly commissioned by the Mayor’s Office for Policing And Crime (MOPAC) and the London Community Rehabilitation Company (CRC). LGE commenced in February 2016 and was initially funded until September 2017. A DMPC decision in July 2017 committed to sustaining the MOPAC funding for this programme until September 2019. The programme is being delivered by a consortia led by Safer London, and includes Only Connect and Redthread. The pan London service was designed to complement and enhance existing local services, filling gaps in provision of support services for young Londoners who are involved or affected by group violence. Young people are not compelled to work with LGE, rather engage on a voluntary basis. The evaluation of London Gang Exit is being conducted by MOPAC’s Evidence & Insight Team (E&I). This interim report provides an update on progress from the preliminary progress report1 and Quality Assurance report commissioned by CRC, highlighting ongoing and emerging risks. It presents key performance data to date and process learning from staff survey and client interviews. The report also explores the early indicative impact of LGE on offending and victimisation. Given the low throughput of clients at this stage2, caution should be applied when interpreting these findings. Key findings: A total 224 cases were referred to LGE between the launch of the programme (17th February 2016) and 15th March 2017. Of those 46 individuals completed3, 67 cases were still open, 21 were marked as ‘on hold’4 and 90 cases were closed meaning they were not completed or the client disengaged. LGE clients were predominately male (91%, n=204) and the average age was 20 years old. Process evaluation • Practitioners believed LGE had a positive impact on clients both in the short and long term, highlighting client motivation as key in achieving results. They also had largely positive views on all key aspects of the programme including partnership working, data sharing, LGE model and processes. • Interviewed clients reported positive experiences with LGE. They described their case workers as being very responsive to any emerging issues and easy to contact. Clients emphasised how well LGE staff engaged clients from their first meeting, which helped to build trust early on. • Centralisation of a Referral Centre has been well received and the staff believe it creates more appropriate referrals by facilitating better communication between Referral and Assessment Teams. Offending and victimisation analysis From the 113 in-scope cases, indicative impact of LGE on offending and victimisation was explored for 70 cases where at least nine months has elapsed post-referral. • A volume of offending seen in re-offending analysis was found to be driven by less than a half of LGE clients who had previous history of offending. The same pattern was seen for victims of crime, who represented only around 30% of the analysed cohort. 1 Made available to partners in December 2016. 2 Period between 17th February 2016 (project launch) and 15th March 2017. 3 Meaning they had received a minimum of six months intervention and had exited the programme. 4 A case can be on hold for multiple reasons including client moving, temporary loss of contact or medical treatment. 2 • Early offending analysis shows that LGE clients are very diverse in terms of their involvement in group violence, risk posed and victimisation levels, which is a reflection of broad eligibility criteria. Differences in offending before and after referral were seen between those who appear on the MPS Gangs Matrix and those who do not. • Preliminary analysis of non-proven offending and victimisation indicates a decrease in an average count of being a suspect/accused and a victim after receiving interventions from LGE. However, this is not statistically significant and the lack of a counterfactual combined with small sample size mean results should be treated with caution. 3 1. Introduction Background Youth violence in London continues to be a significant concern with 16% of Londoners believing that gangs are a problem in their local area5. Gangs range from organised criminal networks involved in Class A drugs supply and firearms, through to street-based gangs involved in violence and personal robbery6. According to recent MPS intelligence there are around 225 recognised gangs in London, with around 3,600 identified as being involved in gang violence on the MPS Gangs Matrix of the most harmful gang members identified on the MPS Gangs Matrix7. It is estimated that this relatively small number of people is responsible for approximately 17% of serious violence and stabbings, 7% of personal robbery, 40% of shootings and 12% of aggravated burglary8. However, it should be noted that the Matrix does not reflect the full extent of youth violence in London. The 2017-21Police and Crime Plani sets out strategies to comprehensively address the problem of gang related violence, offending and exploitation in London in response to recent rises in proxy indicators of gang related violence, such as knife and gun crime offences. The Mayor of London’s Knife Crime Strategyii further seeks to address the wider but overlapping problem of knife-related violence in the capital. Whilst recognising the continuing need for tough enforcement, both approaches also emphasise prevention and intervention, highlighting the importance of joined-up, diversionary approaches in helping vulnerable young people out of serious offending and violent victimisation. London Gang Exit Overview London Gang Exit (LGE) is a £1.5M multi-agency intervention, jointly commissioned by MOPAC and the London Community Rehabilitation Company (CRC). LGE commenced in February 2016 and will run until September 2019. The programme is being delivered by a consortia led by Safer London, and including Only Connect and Redthread, who fund secondments with expertise in secure estates and health. The pan London service was designed to complement and enhance existing local services aiming to support exit from gangs, filling gaps in provision. It seeks to create a consistent service across London, integrating with local delivery and providing additionality to local services with the aims specified in the table below (Table 1). Young people are not compelled to work with LGE, rather engage on a voluntary basis. 5 2016/17 Public Attitudes Survey. Percentage agreeing gangs were a problem in their local area (major/minor problem) at an MPS level. 6 These estimates are dictated by a number of factors, particularly the definition of ‘gangs’ applied, thus it should be noted that the figures can vary across sources. Particularly there are gangs who are involved in less serious crimes, whose activity will not be reflected by these figures. 7 The MPS intelligence tool for identifying and monitoring the most violent gang members on each borough. 8 Metropolitan Police, Trident Gang Crime Command FAQs [accessed 07 July 2017] 4 Table 1. List of LGE outcomes as set out by LGE Theory of Changeiii. Outcome 1 Reduced/ceased involvement in gangs Outcome 2 Reduced/ceased harm from gangs Outcome 3 Reduced seriousness and frequency of offending Outcome 4 Increased ability to access stable and secure housing Outcome 5 Improved health and well-being Outcome 6 Improved relationships (peer, sexual, partners, etc.) Outcome 7 Improved family dynamics Outcome 8 Improved work-related skills, knowledge and employment9 Process Model To be eligible for referral to LGE, the young person (male or female) is required: • To be aged between 16 and 24 and be involved in or at risk of becoming involved in group violence. • To be at significant risk of harm from or exploited by those involved ingroup violence, (such as through child sexual exploitation), or a risk to themselves, or posing a risk of harm to others. • To show some motivation to want to move away from violence and a willingness to engage in the service. Anyone can make a referral including the young person themselves, but they are typically received from statutory organisations and processed by a centralised LGE Referral Unit10. Cases are assessed using information on the referral forms, and through dialogue with the young person to inform risk and motivational assessments. Once accepted a client is usually referred for either: • ‘Component 1’ – one to one support / mentoring - they are allocated a support worker to act as their case manager; • ‘Component 2’ – specialist input - they are allocated a relevant specialist worker to act as case manager. (Note: In many cases clients are submitted for both component 1 and 2). In the initial bid made by the delivery consortium the proposed time for intervention was 6 months per client. However, it was recognised by both sides that this was an average and that there would be a good proportion of clients that would need more than 6 months engagement, and a small proportion that would need less, thus the standard duration of the programme was extended to 9 months. 9 Not every Aim and Outcome is relevant to every young person engaged in the programme. Either Outcome 1 or 2 (the overall goals of the LGE programme) will apply to every young person engaged, depending on whether they are involved in gang activity or affected by gang activity. Outcome 3 relates primarily to service users who are involved in gang activity. Outcomes 4 to 8 will apply to young people depending on the assessment of their needs and the Specialist Services to which they are assigned (if any). The Specialist Services provided in the LGE programme in general map one-to-one to one of these outcomes. 10 Initially referrals were processed by CRC, which caused issues with communication. Therefore, a decision was made to move the referral system into LGE’s offices. 5 Evaluation Overview Aims of the evaluation This interim evaluation report of the LGE programme follows the initial progress report, which was made available to key LGE stakeholders in December 2016. The objectives of the 2017 interim evaluation are: • To assess early indicators of LGE’s impact on the client’s reoffending. • To critically assess the implementation process and track progress in addressing key recommendations from the initial progress report and Quality Assurance reportiv (see Appendix 4 for a list of recommendations). Methodology The evaluation employed a diverse range of research methods to gain an in-depth understanding of the LGE projects impact and implementation process. • Impact evaluation11 – a simple pre and post intervention offending analysis was conducted using CRIS12. This was supplemented with semi-structured client interviews, which offered them a chance to reflect on the perceived impact of participating in LGE on their lives. The final report will seek to employ a Matched Control Sample design to assess impact more robustly with the longer timeframe and larger throughput available (see Appendix 3 for timeline). • Process evaluation – surveys were used to explore practitioner and client views on the implementation process and clients were interviewed about their experience of LGE. This was supplemented with additional information such as Quality Assurance and evaluation reports and contract meeting minutes. The final report will seek to provide a robust assessment of LGE’s impact on client’s reoffending frequency and severity, alongside documenting learning from the whole duration of the programme, in combination with value for money analysis13. Staff survey LGE staff were asked to complete an online survey aimed at understanding their views of the programme. The survey was anonymous to encourage employees to express their views freely. The survey was open for completion from 26th April 2017 until 30th May 2017 and included a mix of multiple choice and open-ended questions. 15 responses were received, a response rate of 83%. Client interviews Five semi-structured interviews were conducted with LGE clients (two conducted via phone and three face-to-face) between 30th May and 7th June 2017. Questions covered their motivation to join LGE, perceived impact on them, and general views on the programme. All interviewees were males under 25. 11 Impact evaluations aim to determine whether an initiative or intervention has had an effect. Process evaluations explore how an initiative or intervention has been implemented. 12 Crime Information Recording System used by Metropolitan Police Service (MPS). Police National Computer (PNC) analysis of proven reoffending was not possible to due restricted time periods. 13 Dependent on throughput, data provision and its quality. 6 2. Performance Monitoring Cohort Demographics Performance data was extracted from the LGE case management system. MOPAC Evidence & Insight were provided with information on 224 clients referred between the launch of the programme (17th February 2016) and 15th March 201714. Analysis showed that of these: • 46 cases were completed, meaning that they had received a minimum of six months interventions and had exited the programme. • 67 cases were still open and 21 were marked as ‘on hold’15. • 90 cases were closed meaning they were not completed or the client disengaged. This data suggests that 40% of clients had not completed LGE. However, it should be noted that this includes cases, which were referred but were never open as they did not meet eligibility criteria. Referrals have been received from most London boroughs16; no referrals have been received from Kingston upon Thames, Richmond and Sutton boroughs during this period. Southwark and Lambeth made the most referrals (n=21 and 18 respectively). Westminster, Greenwich and Hounslow all made 12 referrals each. 91% (n=204) of clients were male, which is only slightly lower than figures for perpetrators of gang- flagged offences across London; only 2017 (9%) females have participated in LGE18. The average participant age was 20; the youngest client was 12 years old and the oldest 29. 86% of the cohort (n=194) were of Black and Minority Ethnic (BAME) background, which is higher than London average for gang flagged offences (82%). A total of 23% clients (n=52) had an identified disability, with learning difficulties being the most common category (n=17). MPS Gang Matrix Just under half (44%, n=50) of clients in ongoing or completed cases had appeared at some point on the MPS Gangs Matrix. Of these, the majority (n=42) featured on the Matrix prior to their referral to LGE. A total of 35 were on the Matrix at the time of their referral19. The Matrix scores individuals based on violent offending, police intelligence and victimisation, separating them into three harm bands indicating relative harm (Red, Amber, Green with Red signifying greatest harm), which means it does not reflect the full extent of gang-related crimes (e.g. drug supply) or those on the peripheries of gangs (i.e. gang associates). A total of 37 appeared on the Matrix at the time of analysis.20 Of these, 11% were graded red (n=4); 35% Amber (n=13) and 54% Green (n=20), which is broadly in line with the RAG status distribution of the full Matrix. The proportion of clients not on the Matrix (54%, n=63) highlights the broad eligibility criteria for LGE; the cohort varies considerably from those with recent, frequent violent offending histories who are well known to police through to individuals who have not come to police notice or are on the periphery of the group violence and vulnerable to exploitation or victimisation. It should also be noted that the eligibility 14 Data was received on 13th June 2017. 15 A case can be on hold for multiple reasons including client moving, temporary loss of contact or medical treatment. 16 Hammersmith & Fulham, Bexley, Harrow and Hillingdon made only one referral. 17 For one client demographic information was not provided. 18 MetStats – Gang Flagged offences from 1st August 2016 to 31stJuly 2017. 19 Seven appeared and were removed prior to LGE referral and eight clients were added to the Matrix after being referred. Seven individuals were removed from the Matrix post-LGE. 20 As per 3rd June 2017. 7 criteria for Matrix inclusion varies between boroughs. It may be those individuals appearing on the Matrix already have access to local or statutory services, in which case there is a greater need to fill the gaps in service provision for a cohort who are affected by gangs but may not be highlighted as high risk for violent offending21. Data permitting, the final report will explore the relationship between LGE and the Matrix in greater detail. Referrals and Interventions The table below indicates the number of LGE clients referred to specific intervention/support strands22. One client can be referred to multiple strands depending on identified needs. 99% (n=210) of the cohort were referred to more than one intervention, 89% (n=189) to more than 3 and 58% (n=124) to more than 4 strands. The most commonly used intervention aims to reduce harm from gangs (n=187). Housing (n=155) also proved popular, which can be reflective of shortage of affordable housing in London (see Table 2). Table 2. Volume of clients per intervention. Strand name N % of full cohort Harm from gangs 187 88% Housing 155 73% Health and wellbeing 125 59% Gang exit 123 58% Family support 97 46% ETE 91 43% Offending behaviour 84 40% Child sexual exploitation 6 3% Harmful sexual behaviour 5 2% In terms of actual outcomes from those interventions (e.g. supported house moves), MOPAC receives this information on a quarterly basis as a part of performance reporting. In the 12 months prior to June 2017 (no data available since programme launch), 43 clients were supported with housing moves, 36 developed an Independent Living Plan and 78 registered with a GP as a part of health and wellbeing strand23. Anecdotal evidence suggests needs assessment and matching needs to interventions is done as a part of the first meeting with the case worker. However, it would be beneficial to have the process of matching needs to interventions documented for future reference and final evaluation. 21 At this stage, exploration of change in harm scores or RAG status is not feasible given the limited sample size and limited of data on closed cases. 22A total of 212 cases had relevant intervention data. 10 remaining cases were referred recently so it is thought they did not undertake needs assessment before data was shared. 1 case was on hold. 23 London Gang Exit Quarterly Performance Report April-June 2017. 8 3. Process Findings Process learning aims to understand the experiences of individuals to explore how a programme is working enabling challenges to be identified and addressed. Findings in this section are drawn from client interviews and staff surveys and focus on views around LGE model and process, partnership working and data sharing, and LGE impact on clients as perceived by staff and clients themselves. LGE model and process All respondents to the staff survey (n=15) said they supported the overarching LGE model, and had a good understanding of the processes involved. They also largely agreed that governance of the project was working effectively. Views on the LGE processes, including risk assessments, were almost entirely positive. However, Quality Assurance recommendations involving embedding an MPS intelligence officer to improve the quality of risk assessment, as well as introduction of case review for high risk cases, are yet to be implemented and will be explored within the final report Eligibility Criteria and Purpose Overall, staff feedback indicated positive views around many aspects of the programme such as matching interventions to needs, risk assessments, their own understanding of intervention strands and processes. However, views on the eligibility criteria were more mixed with only 8 out of 15 being positive. Several practitioners (n=4) did not believe that the LGE eligibility criteria allowed them to engage with the right clients (see figure 1). This was seen as a problem with clear communication of programme’s purpose, as one respondent comments: “Programme could be clearer on its offer and intrinsic purpose and how this is communicated to wider stakeholders”. Linked to the eligibility issue, it was noted in the initial progress report on LGE that the branding of the intervention as ‘London Gang Exit’ may be misleading as many clients have little involvement in gangs. A new brand strategy is being developed by Safer London to address this issue with involvement of the clients and key stakeholders. Figure 1. LGE process and staff understanding of them according to staff survey responses (n=15). 9 Partnership working and data sharing Practitioners held largely positive views of partnership working, with most agreeing that the agencies work collaboratively on LGE. Respondents had largely positive views on data sharing both internally and between key agencies24. Encouragingly, most also believed that data is used effectively when shared. However, the evaluation team have encountered difficulties in obtaining individual level data from LGE; this is likely to have been caused by data migration issues related to the implementation of a new case management system. Whilst this is a significant risk in terms of ensuring effective working as well as enabling robust evaluation, E&I continue to work with LGE to maximise the potential for greater access to individual level data to inform the final report next year. Staff training Although the majority of staff (n=11 out of 14) were satisfied with the quality of training received, a number of training needs were identified, including suicide prevention and conflict management. Some frontline staff also indicated they required more training in working with individuals with moderate to severe learning difficulties. This training need is emphasised by the fact that 33% (n=17) of clients with disabilities had learning difficulties, which represents 8% of all open and completed cases (n=225). Staff views on client impact Almost all respondents (n=14) agreed that LGE has a positive short and long-term impact on young people (see figure 2). However, it was also recognised that clients need to be motivated to change for the whole duration of the intervention, which can be challenging for clients in times of heightened emotional stress. In those times both staff and clients themselves suggested that open channels of communication and responsiveness to changing needs are key to keeping the client on track. However, it should be noted that programme staff typically have positive views on their own work. Figure 2. Staff perceptions of LGE’s impact on clients and wider gang issues (n=14). 24 29% responded “neither agree nor disagree”, which for some may mean they do not have experience of sharing data with other agencies. 10 The length of the intervention was also raised in terms of ensuring a sustainable impact. However, it should be noted that the duration of the programme has already been extended from six to nine months in the early implementation period as it was recognised that complex cases may require a longer intervention. “I do believe the programme should be longer as practitioners need to gain the trust of the clients, unpick years of learned behaviour and support the client to relearn positive and healthier behaviours” LGE staff were also asked to identify elements that work particularly well, that have improved and those than still need improving. The table (Table 3) below summarises key themes from these open-ended questions. Table 3. Key themes from answers to staff survey questions on key positives, improvements made and remaining challenges for LGE. Question Key themes Quote Elements that work well LGE staff identified a range of elements that work particularly well on LGE including: client advocacy, housing intervention strand and partnership working. “I believe the housing advice and advocacy element is particularly effective as all who present with housing need receive advice and that manages expectations which is very important in the current housing crisis we have in London today.” Elements that have improved Respondents believed that multiple processes have improved throughout the duration of the project. The most frequent answer was the centralisation of the referral centre. This is encouraging as the referral process was identified as one of key issues in the previous E&I evaluation report as well as the Quality Assurance report. “Since I started working with LGE I have seen a more streamline and effective method of referral since the referral team and delivery team were co-located.” Elements that still need improving The most frequently mentioned area for improvement was an introduction of more consistent training for all staff. Other comments included more clarity about the future of the programme, increasing the number of frontline staff, improved stakeholder engagement and extending the duration of the programme. “Much more solid training being provided to all the team (in strategic and operational areas) so that everyone is equipped with the same 'tools and the service becomes more consistent.” 11 Client Perspectives of LGE Five clients were interviewed about their experience of LGE, two of which had completed the programme25. It should be noted that clients interviewed were selected by case workers and may be more engaged than others on the programme26. Engagement with case worker Findings from the interviews suggest that developing a trusting relationship with the case worker is crucial to engaging clients. LGE clients spoke about the following elements, which helped them to stay engaged: • Positive first contact - clients spoke of a tendency not to trust other people, ask for help or accept it when offered. Positive first conversations with the case worker convinced them to join the programme. • Responsiveness to emerging issues - all clients mentioned that it was very easy to get in contact with their case worker in-between meetings, or move appointments. Scheduled meetings are important as they can bring structure to the often chaotic lives of clients, but responding to changing circumstances ensures the clients have support in crisis situations. “If I don’t know what I’m doing or feel like I’m going through a wrong stage I have got them as back up to just ask them and they can guide me through.” (Interviewee 1) • Support beyond the duration of the programme - allows clients to stay on track, but is offered additionally by certain case workers rather than as part of the programme’s offer. Self-perceived impact All five clients interviewed believed LGE had a positive impact on their life, although they sometimes struggled to fully articulate and reflect on the journey they had gone through. This issue was flagged by LGE’s front line staff as a potential barrier to capturing the full impact of the programme. Nonetheless, the interviewees spoke about some enablers for the positive change. Moving away from the locations associated with risk or learning how to avoid potentially harmful situations, frequently featured as an element which reportedly has a particularly strong impact on exiting the gang lifestyle, improving feelings of safety and personal wellbeing. “If I didn’t join this programme I would still be in the estate I was in, getting in trouble, I had problems in that area which [case worker] knows about and he helped me get away from it and if I hadn’t have met him I would probably be in there and [ultimately] end up back in jail.” (Interviewee 2) Clients also felt that support and advice they received as a part of LGE in finding a job or re-entering education were key outcomes from the programme and would allow them to achieve a better future and stay away from crime. “I know I definitely wouldn’t be working… she pushed me to find work.” (Interviewee 3) 25 Another interviewee dropped out of the programme after three months. He explained that he was too busy with both working and studying but he still believed LGE had a big impact on his life. 26 This is demonstrated by their willingness to use their own time to be interviewed (no incentives were used for interview recruitment). 12 4. Impact on offending: Preliminary findings Offending history The following analysis was conducted for completed (n=46) and open cases (n=67) only and did not include information about clients who left the programme before its completion27, which can bias the sample to show more positive results that it would be for the entire cohort. A total of 702 cases were identified on CRIS in the time period between January 2014 and June 2017 where a client was named as a suspect or accused (receiving a police charge) for any offence, which represents an average of 6 offences per client. Another 144 cases were found where a LGE client was a victim, which equated to over 1 victimisation per client. 72% (n=81) of the LGE cohort featured on CRIS as either suspect, accused or victim during the period of analysis. This means that 28% (n=32) of the analysed cohort did not feature on CRIS during the entire period of analysis, emphasising the wide range of individuals referred to LGE. The maximum number of offences linked to a single client in this time period was 32. All LGE clients with CRIS records were a suspect or accused at least once in the analysis period. 88% of all clients who were a suspect or accused (n=71) were involved in violence against the person28. 48% (n=54) of clients had been recorded as a victim, of which 80% (n=43) related to violent offences. Almost half the cohort (48%, n=54) had been both a victim and suspect/accused since January 2014 and 34% had been both a suspect/accused and victim in a violent incident, highlighting significant victim-offender overlap. LGE offending and victimisation (before and after referral comparison) 9 month before and 9 month after referral comparison From the 113 in-scope cases, indicative impact of LGE on offending and victimisation was explored for 70 cases where at least nine months has elapsed post-referral. This maximised the case sample inclusion and period after referral. Please note that analysis is indicative only. Caution should be exercised when interpreting findings and the following caveats should be kept in mind: • The small sample size available for analysis, relatively low prevalence of offending, and basic before and after methodology this necessitated, mean that robust causal statements as to LGE’s impact on offending behaviour are not possible at this stage of the evaluation. The final evaluation report seeks to utilise a matched control methodology, allowing for a more robust exploration of impact. • Due to restricted time periods for analysis. CRIS data was used; this only allows insight into arrests and charges, and does not represent proven offending (i.e. individuals found guilty at court). Offending analysis was conducted for nine months before and nine months after each referral29 to establish if there was any change in offending. It should be noted that 31 (44%) of LGE clients in this analysis did not have any suspect/accused CRIS records nine months before or after referral. Out of those, only two were recorded as victims in the same period of time. The maximum number of offences per client before referral was eight (most frequent score was two), which decreased to two after referral suggesting the decrease is driven by a small number of cases. However, a decrease in the count of incidents in all 27 This information was requested by MOPAC Evidence & Insight but LGE could not provide it. 28 Violent offences in police recorded data are referred to as ‘violence against the person’ and include homicide, violence with injury, violence without injury and weapons possession offences. The category does not include sexual offences. 29 E&I were not provided with the actual start date, thus referral date was used as the best alternative. 13 three major categories, including offending (n=-7), violent offending (n=-8), and victimisation was found (n=-13) (see Table 4 for details). The differences were not statistically significant30. Table 4. Change in CRIS offending and victimisation between 9 months before and after referral. 31 Offending Violent offending Victimisation Group Before After Change Before After Change Before After Change Combined (n=70) 95 88 -7 (-7%) 35 27 -8 (-23%) 31 18 -13 (-42%) Completed (n=40) 50 38 -12 (-24%) 19 12 -7 (-37%) 15 8 -7 (-47%) Open (n=30) 45 50 5(+11%) 16 15 -1 (-6%) 16 10 -6 (-38%) Overall, no decrease in the proportion of clients who offended was found (n=30 before and after referral). However, a small decrease was seen for clients who engaged in violent offending or were victims (see Table 5). Generally, those clients who were suspect/accused nine months before referral also offended nine months after, but with less frequency. The only notable exception was shown by three individuals who had no previous records, but appeared 6, 5 and 4 times as suspect/accused respectively nine months after referral. Table 5. Number of clients with CRIS records 9 months before and after referral. Offending Violent offending Victimisation Group Before After Change Before After Change Before After Change Combined (n=70) 30 30 0 (0%32) 20 17 -3 (-4%) 20 10 -10 (-14%) 9 month before referral and 3 to 12 months after referral comparison The same analysis was then conducted for nine months before the referral compared to nine months from three months point (three months point until twelve months post referral) of the programme representing one-third of the typical programme duration (see Appendix 1 for visualisation). This was to allow time for LGE to start having an impact on clients’ behaviour as interventions rarely have an immediate effect. The percentage decrease was greater (-20%, n=-11 compared to -7%, n=-7, see Table 6), but the number of cases available for analysis in this time period halved (n=37) giving us less confidence in the findings. For this sample a decrease in victimisation (-52%, n=-11) was statistically significant although still should be heavily caveated due to the small base size33. 30 Paired samples t-test for 70 cases. Offending p=0.716; Violent offending p=0.369; Victimisation p=0.085 31 Completed cases received full intervention; open cases are ongoing; marked as ‘combined’ are both completed and open cases. E&I were not provided with information on non-completed cases, so this group is not included in the analysis. 32 Percentage of the entire cohort (n=70) 33 Paired samples t-test 95% confidence level, p=0.014. 14 Table 6. Change in CRIS offending and victimisation between 9 months before and after referral, with 3 months post referral excluded to allow for behavioural change. Offending Violent offending Victimisation Group Before After Change Before After Change Before After Change 9 month (n=70) 95 88 -7 (-7%) 35 27 -8 (-23%) 31 18 -13 (-42%) 9 (3) month (n=37) 55 44 -11 (-20%) 23 18 -5 (-22%) 21 10 -11 (-52%) Table 7. Number of clients with CRIS records 9 months before and after referral. Offending Violent offending Victimisation Group Before After Change Before After Change Before After Change 9(3) month (n=37) 18 15 -3 (8%34) 12 11 -1 (-3%) 13 6 -7 (-19%) Again, less than a half of all analysed clients had records of previous offending (see Table 7). Those clients tended to offend in the nine month post period but with lesser frequency. The same trend was seen for violent offending. Propensity of victimisation seems to be consistently decreasing both in terms of number of clients who are victims and the cumulative count of incidents suggesting that LGE works well to decrease harm. However, caution should be applied when interpreting these findings, due to relatively low sample sizes. Subgroup analysis A total of 53% (n=37) of all clients used for nine month analysis appeared on the MPS Gangs Matrix. There were no large differences in offending and victimisation at the baseline between the group appearing on the Matrix and a non-Matrix group (see Figure 3 below and Appendix 2, Table 9 for details). However, there was a statistically significant difference in the change in general offending between those groups (33%, n=17 for Matrix and -56%, n=24 for non-Matrix) with those appearing on the Matrix showing an increase in offending post referral35. Additionally, it was found that for the clients appearing on the Matrix, every individual with a history of offending 9 months after referral also offended in the 9 month period after referral; further 6 clients who did not offended 9 months before, had been recorded as a suspect or charged after referral (see Appendix 2, Table 10). This may suggest more intensive treatment is required for those with stronger connections to gangs. 34 Percentage of the entire cohort (n=37) 35 Independent samples t-test, p=0.029 (offending); No significant difference for violent offending or victimisation p=0.102 and p=0.444 respectively. 15 Figure 3. Count of offences for the Matrix and non-Matrix groups before and after referral. As might be expected, there is a more positive impact for completed cases than for open cases, although the difference is not statistically significant36. Similarly, a difference between impact on being accused or suspect was noted, but was also not statistically significant37. Further cohort breakdowns and comparisons on the basis of demographics or offending history were not possible due to the small sample size available for this analysis. Contingent on sufficient throughput, this will be explored in the final evaluation report. Distance Travelled Assessment Tool The LGE delivery consortium uses the Distance Travelled Assessment Tool to measure the progress made by individuals over the course of their intervention against set outcomes38. Both the client and support worker complete the assessment using a 5-point grading scale for each element at set points including the initial appointment, six weeks later once the support worker has developed a relationship with the young person, and again at 3 months and 6 months. Although initial findings appear encouraging, the evaluation team were not able to access results at an individual level. LGE found 79% of completed cases (n=51) engaged in gang exit activity, an 80% reduction in reports of harm/feeling vulnerable and 75% engaging in activity to address offending behaviour. However, caution should be applied when interpreting these findings as they are entirely based on self-reporting, and the rigour in data collection could not be accurately assessed. 36 Independent samples t-test, p=.104 37 Independent samples t-test, p=0.505 38 Outcomes assessed include: gang activity, harm from gangs, offending behaviour, housing, health & well-being, relationships, family and education, employment & training. 52 43 69 19 0 10 20 30 40 50 60 70 80 Matrix Non-matrix Before referral After referral 16 5. Discussion This report presents findings from an interim evaluation of LGE. A range of research methods were applied to assess progress from the last progress report, including staff survey and client interviews. The report also explores the early indicative impact of LGE on offending and victimisation. Given the low throughput of clients at this stage, caution should be applied when interpreting these findings. A number of improvements have been noted in LGE processes and implementation since the preliminary progress report, indicating the benefits of ‘action learning’. Feedback suggests that the integration of the referral centre and LGE staff, as recommended in previous process learning, has been received positively. Furthermore, staff feedback indicated positive views around many aspects of the programme such as matching interventions to needs, risk assessments, their own understanding of intervention strands and processes. Several challenges remain. Some staff believed the programme could be clearer on its purpose and how it is communicated to wider stakeholders, which can be linked to very broad eligibility criteria. There is also a need to improve and centralise collection of client data to enable efficient data sharing with key partners, including MOPAC. It would be beneficial to the risk assessment procedure to address Quality Assurance report recommendation and gain better access to MPS intelligence. Practitioners also believed case workers need additional training in relation to working with clients with learning disabilities. This can prove to be an increasing issue as learning difficulties was the most common disability as identified by the case workers. Further, there is a need for a clear process for high risk cases and reviewing them. Early analysis was conducted to provide some indication of the offending and victimisation patterns of LGE clients. The analysis of CRIS records shows that the frequency of offending and victimisation seem to be driven by a small proportion of clients. Only less than a half of the analysed cohort had offended nine months before referral and mostly the same clients offended after referral, but with slightly smaller frequency. The same trend was seen for the clients who were victims, but the decrease in a number of clients who became victims after referral was greater than for offending. The analysis of cumulative count of incidents shows a decrease in the number of CRIS records relating to offending, violent offending and victimisation after referral to LGE. However, it too soon to draw firm conclusions around offending or victimisation given a relatively small sample size, short follow up and most importantly simple methodology with no comparison group. The final evaluation will seek to provide a robust impact evaluation, but this will only be possible with sufficient throughput and access to individual level data. 17 Appendices Appendix 1. Timelines for offending and victimisation analysis. Figure 4. Visualisation of timelines used for offending and victimisation analysis. 18 Appendix 2. Detailed offending change figures. Table 8. Detailed before and after figures for offending analysis (count of offences). Offending Violent offending Victimisation Time period Group Before After Change Before After Change Before After Change 9M (n=70) Combined 95 88 -7 35 27 -8 31 18 -13 Completed 50 38 -12 19 12 -7 15 8 -7 Open 45 50 5 16 15 -1 16 10 -6 9(3)M (n=37) Combined 55 44 -11 23 18 -5 21 10 -11 Completed 43 29 -14 18 13 -5 12 5 -7 Open 12 15 3 5 5 0 9 5 -4 Table 9. Detailed before and after figures for the matrix and non-matrix group (count of offences). Offending Violent offending Victimisation Group Before After Change Before After Change Before After Change Matrix (n=37) - count of offences 52 69 17 21 24 3 15 11 -4 Matrix (n=37) – count of clients with CRIS records 18 24 6 11 14 3 11 6 -5 Non Matrix (n=33) - count of offences 43 19 -24 16 7 -9 14 3 -11 Non Matrix (n=33) - count of clients with CRIS records 12 6 -6 9 3 -6 9 4 -5 19 Appendix 3. Evaluation timeline and milestones. Jun-17 Jul-17 Aug-17 Sep-17 Feb-17 Mar-17 Apr-17 May-17 Oct-17 Nov-17 Jun-18 Jul-18 Mar-18 Dec-17 Jan-18 Feb-18 Apr-18 May-18 Client Interviews Assist with Client Interviews schedulling Staff Survey 2 LGE Evaluation INTERIM REPORT (charges and arrest, interv iews and staff surv ey findings) FINAL REPORT Staff Survey 3 Client Interviews DEADLINE FOR MOPAC RECEIVING PARTICIPANT DATA DEADLINE FOR MOPAC RECEIVING PARTICIPANT DATA Assist with Staff and Client interview scheduling Headline findings discussion @ Quarterly Contract Quarterly meetings Confirmatio n of the cost of the programme for VFM analysis Staff interviews 20 Appendix 4. Key recommendations from previous reports. MOPAC Evidence & Insight LGE initial progress report (December 2016) Key recommendations 1. The branding of ‘gang exit’ implies a service designed to solely help gang members leave, yet the programme has a much broader remit to help those affected by gangs, or who are at risk of harm from gangs. This has the potential to exclude certain groups. Looking forward, a consistent level of engagement will ensure the young people most in need are getting access to the LGE. 2. A shortage of LGE staff caused initial problems, with specialist caseworkers taking on clients in the absence of support workers. However, this was identified as an initial mobilisation issue, and staff are now confident there are sufficient contingency plans in place should caseloads reach full capacity (e.g. the use of volunteers). 3. There are barriers associated with the separateness of the referral centre and LGE, which affects the ability to retain the integrity of the programme. Greater knowledge and experience sharing is recommended to ensure a more streamlined feedback loop. Training in specialist areas to upskill all members of the team would also help to improve this. Johnathon Toy, Pan London Gang Exit Programme – Quality Assurance Final Briefing (October 2016) Key recommendations 1. It is recommended that a review takes place to understand why the inconsistencies exist between boroughs with known gang problems and what the blockages are for those boroughs who are not referring. An improvement action plan should be implemented and overseen by the service commissioners. The review should consider delivering a more targeted training programme to key borough or services such as education providers, including pupil referral units, academy chains and the secure estate. 2. As recommended in previous reports, it is strongly recommended that the Motivational Assessment Tool is removed from the referral team process and integrated into the delivery consortium assessment interviews. This would allow the delivery consortium to measure the motivational change of the client during the delivery of the intervention. 3. Commissioners and senior managers need to work together to resolve the definition of the risk level of clients eligible for the scheme as a matter of urgency. 4. Improvements to the case management process between the referring agency, referral team and delivery consortium. a. It is recommended that it is the referral team that contact the referring agency, as soon as the case has been approved. b. The delivery consortium should have the authority to contact the client direct, at the point when the referral team know the case will be approved. c. It is recommended that the delivery consortium introduce a system of a lead worker and a co-worker to the client. The lead worker would carry out the majority of the support, the co-worker would be known to the client and have regular but sporadic engagement. d. Greater freedom of the location of initial engagement – The key workers and operational managers stated that homes visits with the clients are not undertaken as part of the early engagement except as “a last resort”. 21 e. Greater integration of the MASH – commissioners and senior managers should review the role of MASH in this process and the opportunities or closer alignment with borough based MASHs. f. Referral team managers and the operational managers of delivery consortiums should undertake a case review and report on cases where a case has been referred and the interventions requested forms part of a statutory provision. g. is now strong evidence that with the calibre of the staff within the referral team and the key workers, there is an opportunity for greater cases based integration. The case based integration would include a very early dialogue between the referral team officer and a key worker when a case comes in and an ongoing dialogue in the early stages of the case acceptance and progress. 22 References i MOPAC. (2017). A Safer City for All Londoners: Police and Crime Plan 2017-2021. Retrieved from: https://www.london.gov.uk/sites/default/files/mopac_police_and_crime_plan_2017-2021.pdf ii MOPAC (2017). The London Knife Crime Strategy. Retrieved from: https://www.london.gov.uk/sites/default/files/mopac_knife_crime_strategy_june_2017.pdf iii TSIP. (2016). London Gang Exit Programme: Theory of Change and Programme Outcomes February 2016. iv J., Toy (2016). Pan London Gang Exit Programme – Quality Assurance BRIEFING March 2016, Commissioned audit by CRC. MOPAC Evidence & Insight Stalking in London: a deep dive July 2024Embargoed until 00:01 on Tuesday 23rd July. Background and methodology to the Stalking Deep Dive • Stalking is fixated, obsessive, unwanted & repeated behaviour, which causes fear of violence or engenders alarm & distress in the victim. A national priority. • High prevalence in England & Wales - 1 in 5 women & 1 in 10 men have experienced stalking in their lifetime.1 2 • Other recent research highlights a strong correlation between stalking & homicide, suggesting the need to effectively respond to allegations. • National Stalking Consortium (including stalking experts, frontline services & victims) launched a super-complaint against police in England & Wales in 2022 highlighting systemic failings in the police response, resulting in poor victim outcomes & putting victims at risk. Super-complaint currently being investigated by HMICFRS, IOPC & College of Policing - preliminary findings expected mid/late 2024. • In London, better supporting stalking victims is a priority in the Mayor’s Police & Crime Plan 2022-25 (MOPAC, 2022a). In response, the London Victims Commissioner requested MOPAC’s Evidence & Insight (E&I) to conduct ‘deep dive’ research into stalking. • Data used in this report largely runs until 2023. Since this time there have been some changes in practice, but the main themes should be consistent. The Deep Dive approach An established & diverse methodology which has been previously adopted across many topic areas. Key research questions are: Analysis of secondary data sources: Aggregated MPS stalking data (2015-2023); Stalking Threat Assessment Centre (STAC) MPS caseload data (2023); aggregated PS (Probation Service) London data (2023) Scoping review of the literature Interviews with key stakeholders (conducted Oct-Nov 2023) Deep-coding stalking police crime reports (CRIS) July 2020-June 2021 Rapid Evidence Assessment Victim interviews (conducted July 2024) Deep coding exercise of 400 MPS crime reports to better understand: • Stalking victims & perpetrators. The nature of stalking. • Stalking investigations & procedures, including drivers of police • no further action (NFA) & victim withdrawal. Stalking perpetration in London ▪ Victim characteristics ▪ Perpetrator characteristics ▪ Offence characteristics Criminal justice response to stalking ▪ Case attrition analysis with statistical modelling ▪ Victims’ CJS experiences Key learning from the Deep Dive The criminal justice system Overall, the increase in stalking cases in London has not been matched by increase in charge/summons, caution or community resolution. Of the coded cases - victims withdrew in 45% (n=169) of cases. • 89% of these described as ‘not willing to prosecute or want a CJS outcome’. • 36% withdrew in the first week indicating a narrow window to engage. Police decided to close 41% (n=155) with ‘no further action’ (NFA). • Lack of evidence cited by police in 57% coded NFA’d cases. • 42% of those cases NFA’d were within the first 4 weeks. Investigation characteristics are the strongest predictors of victim withdrawal & NFA. • Victim withdrawal less likely: suspect arrested/interviewed or witness present. • Victim withdrawal more likely for s4a, when special schemes put on victims’ address, & when victims offered/provided support. • Cases with victim technology less likely to be NFA’d. • Repeat victims of stalking were less likely to be NFA’d. • Vulnerable victims and DA victims more likely to be NFA’d. Stalking not consistently identified • Stalking offence reclassification appears frequent (in MPS & CPS). 8 in 10 offences not initially identified as stalking. CPS changed charge in 25/33 coded cases. Gaps in investigations • S-DASH risk assessments completed in only 7% (n=27) of coded cases. • Victim impact in Victim Personal Statement recorded in less than 10 coded cases. • Evidence of engagement between OICs & STAC officers in coded cases was limited. • Potential technology evidence referenced in 40% of coded cases, suspect & victim technology retrieved in minority of cases. • Other MOPAC research found stalking had highest proportion of cases NFA’d due to ‘victim supports, evidential difficulties’ compared to other VAWG groups. SPO use • MPS target of issuing SPOs in 2% of stalking cases, SPOs obtained in 1.4% of stalking offences in London in 2022. In terms of volume, MPS issued more SPOs than other forces, and there have been recent positive efforts MPS to support officers. Victim voice • The victims we spoke to were impacted negatively by the CJS process, lacked confidence in police investigations, & experienced poor advice & communication from police & CPS. These interviews, interviews with victim advocates, & wider literature also indicated a lack of public awareness of stalking behaviour. • No current robust capture or feedback loop of victim voice. An increasing prevalence • In London, an 11-fold increase in cases since 2016/17, in part due to reclassifying domestic harassment cases in 2019. • Estimated that 1 in 5 women & 1 in 10 men have experienced stalking in their lifetime. • Per 1000 population police recorded stalking is lower in London (6.98) than E&W (11.71) or most similar forces (17.89). • 11,721 recorded stalking cases in London in 22/23 with considerable Borough variation. Victims • Since 2016/17, around 80% of Police recorded stalking victims in London each year have been female. • Younger people more likely to be victims of a more serious stalking offence (s4A). • In the coded cases, half of stalking victims (n=189) have previously been a victim of any crime. With a quarter (n=100) being victim of stalking/harassment previously. • Only a minority of coded victims were recorded by police to have mental health issues (10%) or be vulnerable (14%), & the proportion identified as being vulnerable has been declining in recent years. This likely indicates investigation gaps in data recording on victim vulnerabilities. Suspects • The majority of stalking perpetrators male. • Coded cases indicate 62% of suspects had previous allegations of a crime & 16% had been previously convicted of an offence. • Just over a third of suspects (n=131) were reported to have mental health issues. 29% of suspects were reported to have substance misuse issues. 53% of stalkers on probation in London have psychological problems. The offence • Stalking behaviour is varied. 66% of cases involved phone calls/text; 38% surveillance/ spying. 34% unwanted face to face contact. • Most stalkers engaged in at least 2 different types of stalking behaviour. S4a cases included more types & more serious behaviour. • Escalation was noted in half of cases (n=195). However, in a quarter of these the victim stated escalation whereas the police did not. • In a quarter of coded cases, stalking behaviour continued after the initial report. • Majority of stalking offences (82%) in London flagged as DA related. • Victim-suspect relationship was only recorded for 9% of MPS dataset. Support partnership working 1. Promote cross-BCU working of stalking SPOCs - to share & disseminate best practice & common issues from across London. 2. OICs to collaborate more with victim advocates & STAC - acknowledge advocates as a resource to support better investigations, facilitating communication with victims. 3. Build on existing collaboration between CPS & MPS on stalking outcomes. Recommendations from the Deep Dive Increase evidence-based working, training & resourcing 1. Provide and evaluate specialist stalking training for relevant CJS agencies - particularly CPS & frontline police officers. Training may include definition, legislation & offence classification, the nature of stalking, stalking risk assessments & victim impact & responses. This report notes recent delivery of stalking awareness training with 782 MPS & London probation officers, and CPS lawyers, and the MPS e-learning stalking awareness package in final stages of development. Monitor and measure impact of these initiatives.​ 2. Systematically capture victim voice & feedback. There are current surveys available that could be utilised to better capture this group’s victim voice. This should happen and then be used to optimise provision of support and advocacy for victims. 3. Given the overlap between DA and stalking, there should be consideration to building in stalking components within these wider DA programmes (i.e., Drive & probation accredited programmes) and STAC may have a role to play here. 4. Attend and develop the evidence base in respect to Stalking. This would include incorporating offender need into design and ensuring robust evaluation. At the intervention level, the deep dive highlights a potential gap in picking up lower risk perpetrators (i.e., those not meeting STAC NHS level) and the wider literature identified certain approaches with better results (i.e., Dialectical Behaviour Therapy). This report notes the new multiagency early stalking intervention programme (ESIP), evaluation of ESIP is opportunity to develop evidence base. 5. Create public awareness campaigns: raise awareness of stalking behaviour & legal sanctions, particularly ex-partner & online stalking behaviours. 6. Make resources for victims available to aid understanding stalking legislation, investigations, court processes & their rights. 7. Ensure front line officers receive and engage with key learning from this research. Establish and monitor minimum standards for the CJS stalking response 1. Clearly defined minimum standards for stalking investigations - incorporate best practice & produce guidance tools: appropriate personal safety advice for victims, risk assessment for all stalking cases, monitoring SPOs, routinely collecting victim impact statements, use of evidence in stalking investigations, creating crime reports for further offending & consider that criminal justice routes should not be abandoned in favour of civil or mental health routes. 2. Monitor delivery of minimum standards across MPS/BCUs to assure application as intended, performance reporting and a clear feedback loop into learning; improve investigations & victim response. MPS wide SPO enforcement procedures & resource. 3. Fold partners into application of minimum standards - include victim communication guidance (with why investigations, legal proceedings or prosecutions are not continued). Address downgrading of stalking charges to harassment, & lack of stalking charges for breach of protective order cases. Improve data 1. Improve MPS data recording to better support stalking responses, especially capturing vulnerability, ethnicity & victim-suspect relationship. 2. Improve linking data for quicker access to background information on victims/suspects, linking previous incidents, and to support identification of stalking (as a behaviour offence). Assess extent that Connect system addresses this. The need for a strategic approach to managing high-harm offenders In 2023 MOPAC E&I explored the management of High Harm offenders in the MPS. This work identified a lack of central strategy and definitions, and consideration needed of resourcing, multi-agency response and monitoring/enforcement of orders. The management of stalking should be viewed within this wider organisational perspective ensuring a strategic and joined up response to the management of high harm offenders. Background & methodology 1. Stalking is a London & national priority 2. E&I approach to the Deep Dive Stalking is a London & national priority • In London supporting stalking victims is a key priority in the Mayor’s Police & Crime Plan (PCP) 2022-25₁: ‘Improving the service & support that victims receive from the MPS, support services & the CJS… Victims receiving a better criminal justice response & outcome, reducing the number of repeat victims of domestic abuse, sexual violence & stalking.’ • PCP outlines: ‘MOPAC will continue to work alongside the MPS, Probation, the CPS, & victims & survivor services to ensure the identification & response to stalking in London is improved & that all agencies meet the needs of Londoners resulting from increases in stalking reports to police.’ • Response - London Victims Commissioner commissioned MOPAC’s Evidence & Insight Unit to conduct ‘deep dive’ research into stalking victims & perpetrators in London. 1. MOPAC, 2022a Suzy Lamplugh Trust, 2021 Storey et al., 2023 Monckton-Smith et al., 2017 Suzy Lamplugh Trust, 2022a 2. 3. 4. 5. A lot is already known: • High prevalence in England & Wales - 1 in 5 women & 1 in 10 men have experienced stalking in their lifetime.₂ • An estimated 2.4% of men & 4.4% of women over 16 years were victims of stalking in the UK in 2022/23. • Research highlights importance of effectively responding to stalking. Stalking behaviour is found to have significant detrimental impacts on victims’ psychological wellbeing.₃ Research highlights a strong correlation between some stalking behaviours & homicide.₄ • National Stalking Consortium (including stalking experts, frontline services & victims) launched super-complaint against police in England & Wales in 2022: Highlighted systemic failings in police response to stalking, resulting in poor victim outcomes & putting victims at risk.₅ • Super-complaint currently being investigated by HMICFRS, IOPC & College of Policing - preliminary findings expected late 2024. ‘Stalking is a highly impactive crime, which leaves victims fearful of what their stalker may do next.’ NPCC’s Lead for Stalking & Harassment, Deputy Chief Constable Paul Mills ‘Stalking is a complex form of abuse, & it can have a devastating impact on the lives of victims & their families. Sadly, it can end in the most tragic circumstances.’ Minister for Victims & Safeguarding, Laura Farris • Aggregated MPS data on all stalking offences, victims & perpetrators between July 2015 & March 2023. • MPS data from Stalking Threat Assessment Centre (STAC) caseload (2023). • Aggregated PS (Probation Service) London data of stalking offenders (‘snapshot’ of live cases extracted in November 2023). Scoping review of literature - review of academic literature, including 20 studies & 7 data sources. Semi-structured Interviews with 17 key stakeholders (conducted Oct- Nov 2023), including: central MPS, STAC leads (MPS, NHS, PS London, & Suzy Lamplugh Trust) & VAWG local authority leads. Deep-coding of a random selection of 400 stalking police crime reports (out of a total of 10,436 offences between July 2020-June 2021). See methodology here. Rapid evidence assessment of perpetrator interventions. Victim voice: E&I researchers observed one focus group & 3 interviews with stalking victims chaired by the London Victims Commissioner (conducted July 2024). Research Limitations Police data (and London PS) & crime reports are not research specific tools, have data error and they only document reported cases and we know stalking in an underreported crime. It is however an organisational record of demand and response and so remains highly valuable. E&I approach to the Deep Dive E&I followed an established & mixed-method approach to the research: Rapid evidence assessment Reviewed the literature Stakeholder interviews This problem profile is linked to a wider series of analytics exploring vulnerability across London (i.e., victim withdrawal, Child Sexual Abuse, VAWG perpetrators) due for completion 2024. Coded cases MPS or PS London data Statistics relating to London Comparison to England & Wales Data used in this report is largely from 2023 & earlier. As such, there may have been some changes in practice since. Victim interviews 1. Defining stalking • Background information on stalking definitions & legislation. 2. Stalking in London • Describes prevalence & characteristics of stalking in London: • MPS, PS London, & STAC data. • Learning from the E&I coding exercise. 3. Criminal justice response to stalking • Regression analysis looking at drivers of police ‘no further action’ (NFA) & victim withdrawal within stalking. • Reclassification of stalking cases • Use of Stalking Protection Orders • STAC • Stalking victims’ experience of the CJS 4. Appendices 1. Deep coding method 2. Criminal justice journey 3. University of West London (UWL) systematic literature review 4. Perpetrator intervention programmes across London 5. Victim support services across London 6. STAC CJS input 7. Definitive assessment on disproportionality not possible 8. 2023 strategic review of MPS High Harm Offender Management 9. References Defining stalking 1. The basics - what is stalking? 2. Stalking typology 3. Stalking legislation The basics - what is stalking? • Echoing definitions in wider academic literature₁, the Suzy Lamplugh Trust (national charity for victims of stalking) defines stalking as: “A pattern of fixated & obsessive behaviour which is repeated, persistent, intrusive & causes fear of violence or engenders alarm & distress in the victim”₂ • Key to this definition is ‘pattern’ or ‘repeated’: stalking is a ‘behaviour’ or ‘course of conduct’ offence – behaviour conducted over time. Stalking is difficult to define because:₁ 1. It incorporates a wide range of behaviours. It might include: • Following, spying, loitering Can also include: • Unwanted phone calls, messages, emails, gifts/flowers (love bombing) • Verbal abuse • Befriending victim’s friends & family • Blackmail • Criminal damage • Cyberstalking: signalling a recent shift in stalking perpetration through digital technology, especially since Covid-19₃. • Identity theft & account hacking • Physical & sexual violence 2. Victim’s perceptions are integral to stalking behaviour, not just intentions & behaviour of perpetrator. 1. Scott, 2020 2. Suzy Lamplugh Trust, 2022b 3. Suzy Lamplugh Trust, 2021 • Stalking perpetrators are not a homogenous group - behaviour is often influenced by several factors, e.g., perpetrators needs, psychological characteristics, motivations, & victim relationship. • Several stalker typologies/classifications are based on a combination of these factors. Mullen et al.’s₁ typology is the most widely cited typology & used by STAC – the Stalking Threat Assessment Centre - in London2. The typology was originally developed on a small-scale clinical sample of serious & frequent stalkers. • Pie chart - 2023 suspect data for STAC reviewed MPS s4a stalking cases. Majority of STAC assessed s4A stalking perpetrators were ‘Rejected’ stalkers (n=2,781, 61%). The dominant typology of stalkers identifies five key types Incompetent stalker Usually lacking in social skills, often recognise their affections are not reciprocated, but still pursue victim with aim of securing a date /sexual encounter. High rates of ‘intellectual disabilities’. Intimacy seekers Motivated by infatuated desire for a relationship with victim, often suffering from erotomaniac delusions & mistaken belief their desire is mutual. High rates of delusional disorders. Predatory stalker Driven by sexually sadistic fantasies, making meticulous plans to launch a sexual attack on their victim, who is often a stranger. Rejected stalker Ex-intimate partner of victim, striving for reconciliation or revenge following relationship breakdown. High prevalence of personality disorders. Resentful stalker Behaviour is driven by feelings of revenge & the wish to instil fear & distress in the victim. Mullen et al.’s (1999) stalking typology: 1. Mullen et al. (1999) 2. A multiagency service in London (MPS, PS London, NHS, Suzy Lamplugh Trust) that assesses risk/threat to identify the appropriate intervention based on the stalking behaviour. MPS or PS London data 2.40% 4.30% 0.66% 61.35% 4.35% 10.24% 16.70% Typology of s4a stalkers on STAC caseload Incompetent Intimacy seeker Predatory Rejected Resentful Not stalking Missing / Unknown Stalking legislation is unclear • There is no UK legal definition of stalking₁. Stalking became a criminal offence in England & Wales in 2012 - following an Independent Parliamentary Inquiry into Stalking Law Reform₂. • The Protection of Freedoms Act₃ amended the Protection from Harassment Act 1997 to introduce two types of stalking offence (2A & 4A); • In 2019 the Home Office changed the threshold between harassment & stalking. All domestic harassment is now classified as 2a. Harassment v stalking? • For non-domestic stalking, the difference between harassment & stalking can be thought of on a sliding scale₅. The National Police Chief Council guidelines advise officers to consider the FOUR mnemonic: Fixated Obsessive Unwanted Repeated • Academic literature separates harassment from stalking as being more serious, occurs for longer & causes mental or physical harm to the victim; involves 1) multiple intrusions & 2) fear-provoking behaviours with significant long-term victim consequences₆. Stalking is fixated, obsessive, unwanted & repeated. Harassment is unwanted & repeated. Section 4A - stalking that has additionally caused fear of violence (i) or serious alarm or distress (ii). • Serious distress - behaviour with ‘substantial adverse effect’ on the victim’s day to day activities. • Ten years max. prison sentence. • Note: key difference between 2A & 4A is victim impact. • Note: experts want a single offence due to police classification difficulties. Section 2A – behaviour that amounts to stalking • Six months max. prison sentence. • Non-exhaustive list of behaviours including; following, contacting, publishing victim-related material, monitoring, loitering, interfering with • property & watching/spying. Harassment Stalking 2a Stalking 4a Unwanted, repetitive behaviour Domestic harassment, & non- domestic harassment PLUS fixated, obsessive Stalking PLUS substantial impact on victim The operational difference between harassment, 2a & 4a stalking for the Met can be considered on a sliding scale, depending on victim impact: 1. Scott, 2020 2. Richards et al., 2012 3. The Protection of Freedoms Act (2012) 4. The Stalking Protection Act (2019) 5. College of Policing, 2020 6. Mullen et al., 2001; Scott, 2020 • The Stalking Protection Act₄ introduced Stalking Protection Orders (SPOs) - civil order applications made by police to the magistrate’s placing restrictions (e.g. no victim contact) &/or perpetrator requirements (e.g., attend a perpetrator programme) to protect victims from risk. SPOs can be requested where: • Threshold to charge has not yet / will not be met, to allow early police intervention, or • Where suspect has been charged, to complement stalking offence prosecution. Stalking in London 1. Overall trends 2. Victim-suspect relationship 3. Stalking victim demographics 4. Previous victimisation & victim vulnerability 5. Stalking perpetrator demographics 6. Repeat offending & offending history 7. Perpetrator vulnerability 8. Stalking behaviours 9. Duration of stalking COVID 19 Restrictions Introduction of SPOs Home Office reclassification of stalking offences Recorded stalking in London has soared in recent years, largely explained by guideline changes • From MPS data - the number of recorded stalking offences increased 11-fold between 2016/17 & 2022/23. • This can largely be explained by the change in Home Office classification of domestic harassment to s2a stalking in 2019. However, MPS recorded s4a offences also increased in this period. • Of 11,721 MPS recorded stalking offences in 2022/23: • 74% section 2A stalking (n=8633) • 26% section 4A (i) or (ii) – stalking with fear of violence or serious alarm distress (n=3026) • 0.5% breaches of SPOs (n=62) • Greenwich had the highest number of stalking offences per 1000 population (1.8 per 1000) in 2022/23, while Richmond upon Thames had the smallest (0.8 per 1000). How does London compare to the rest of England & Wales? • Latest data from the Office for National Statistics (data for 12 months ending December 2023) shows the MPS are ranked 35th for Stalking & harassment offences across E&W. - MPS: 6.98 offences per 1,000 population. - E&W (excluding MPS): 11.71 offences per 1,000 population. - Most Similar forces: 17.89 offences per 1,000 population. • It is not clear why numbers are lower for London. It could be linked to differences in police interpretation, police recording, lower prevalence or differences in victim reporting. Kensington & Chelsea 1.6 62 8633 984 2042 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Number of stalking offences in London, by type of offence Breach of Stalking Order/Interim Stalking Order Section 2A offence Section 4A(i) Fear of violence Section 4A(ii) Serious alarm/distress Hammersmith & Fulham 1.5 MPS or PS London data Majority of stalking offences in London flagged as domestic abuse related by MPS. E&I deep coding finds victim & suspect were ex-partners 7 in 10 coded cases Deep coded cases provide more on victim-suspect relationship. • In 7 in 10 coded cases - the victim & suspect were ex-partners. • In 12% of cases (n=47) the victim & suspect were known acquaintances, e.g., friend. • In 8% of cases (n=29) the suspect was a stranger. • Male suspects more likely to be ex-partners than female suspects (71% to 61%). • Male suspects more likely to be a stranger (9% to 3%) & less likely to be a known acquaintance/colleague (13% to 20%) than female suspects. • Three quarters of cases (n=285) were coded as DA-related, which nearly matches MPS flagging of DA cases in the dataset. • Majority of stalking offences (82%, n=9593) in London flagged by MPS as domestic abuse (DA) related in 2022/23: • 86% (n=7409) of s2A offences • 71% (n=2150) of s4A offences Coded cases MPS or PS London data Victim-suspect relationship was only recorded for 9% of MPS dataset. 70% 2% 1% 1% 2% 3% 12% 0% 8% 1% Victim - suspect relationship in coded cases Ex-partner Current partner Suspect is victims parent Suspect is victims child Other family member Colleague or peer Known acquaintance Known online acquaintance Strangers Other 44% 51% 58% 60% 84% 86% 86% 46% 58% 58% 62% 73% 71% 71% 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 MPS recorded stalking cases flagged as domestic abuse related 2016/17 to 2022/23 2A DA Flagged 4A DA Flagged London’s victims of stalking are predominantly female. Younger people are more likely to be victims of a more serious stalking offence • Since 2016/17, around 80% of stalking victims recorded by MPS in London each year have been female. This is different to the CSEW expected levels & likely illustrates an issue with underreporting. • London police data indicates a higher proportion of female victims compared to the rest of England & Wales - where 66% of stalking victims are estimated to be female₁. • Majority of stalking victims are aged between 25 - 44 (62%) in 2022/23. A third are aged 25 to 34 years old (35%, n=2,885). Age breakdown has remained consistent over time. • Younger people more likely to be victims of a more serious stalking offence (s4A) than older age groups. • In 2022/23, the victim age group with the largest proportion of S4A offences was under 18 (39% n=88), followed by 18–24-year-olds (34% n=451). • In England & Wales, the highest estimated victimisation rates were for those aged 16-19 years at 9.7%, followed by 20-24 years at 8%1. • The CSEW finds that nationally 14% of stalking victims were stalked by a partner or ex-partner, 8% by a family member, 20% domestic stalking, 44% cyberstalking1. • Self-defined ethnicity in MPS data was not recorded for 35% (n=2908) of victims due to victim declining, not understanding, or no ethnicity recorded. • A third of stalking victims self-defined as white (34%, n=2782). • Highest estimated victimisation rates in England & Wales were amongst those of mixed ethnicity, at 12.1%, followed by those who were white at 3.4%.1 • Victim ethnicity is not that different to the population of London, however any findings on ethnicity are skewed by the high proportion of missing ethnicities. 1. ONS, 2023 Victim ethnicity (self-defined) 2022/23 MPS recorded stalking offences data 2022/23 MPS or PS London data 3% 16% 35% 27% 12% 5% 2% 0% Victim age groups 2022/23 Under 18 18 to 24 25 to 34 35 to 44 45 to 54 55 to 64 65+ Unknown 9% 9% 35% 3% 10% 34% Asian or Asian British Black, Black British, Caribbean or African Declines/Does not understand/not recorded Mixed or Multiple ethnic groups Other ethnic group White Coded cases highlight victims' previous victimisation & apparent gaps in recording vulnerability • Half of victims (n=189) within the coding exercise have previously been a victim of any crime. • This is higher than overall repeat victim proportion in MPS data, which was 37% (n=3,074) in 2022/23. • A quarter (26%, n=100) have previously been a victim of stalking or harassment-offences. Of these, nearly all (n=92) have been victims of stalking/harassment from same suspect. • 32% of victims have previously been a victim of domestic abuse. • In 29% of coded cases (n=109), the victim & suspect have children together. • 2 in 10 victims (n=86) have previously had allegations against them for any crime. - 5% have previous convictions (n=19). - 4% (n=15) have previously stalked or harassed the suspect, for instance the suspect has a SPO, RO or NMO against victim. Coded cases • Just 10% of victims (n=38) were reported by police to have mental health issues – this is markedly lower than the 41% in the 2019 Rape Review research, & similar to the 12% in the DA Deep Dive research in 2022. • 2019 survey of 96 stalking victims - 24% reported having mental health problems prior to being stalked, & 91% reported suffering mental health issues after being stalked₁. • In 8 cases - the victims’ severe mental health was considered to undermine allegations of the case. • E&I coded cases - only 14% of victims (n=54) recorded by the police as vulnerable. Considered underreported as victim emotional distress or fear was recorded in 54% of cases, & previous victimisation in 50% of cases. Deep coded cases also found 19 victims were reported to have mental health issues but not recorded as vulnerable by the police, 8 with physical health issues (including disability), & 8 known to police for safeguarding reasons but not recorded as vulnerable. • Overall, this likely illustrates the difficulty & inconsistency in the police capturing this sensitive information & is consistent with the declining trend in MPS recording victim vulnerability. 1. Taylor-Dunn et al., 2018 MPS data shows downward trend in recording of stalking victims as vulnerable MPS or PS London data Number of victims % of victims recorded vulnerable 134 201 271 365 365 846 891 747 0% 5% 10% 15% 20% 25% 30% 35% 0 100 200 300 400 500 600 700 800 900 1000 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 MPS data: number & % of stalking victims recorded as ‘vulnerable’ In London, 9 in 10 charged or cautioned stalking perpetrators are male, two thirds are 25-44 & half are white MPS stalking offences data 2022/23: • In London - 9 in 10 stalking perpetrators charged/cautioned are male. This has remained consistent over time & consistent literature which finds stalking perpetrators are predominantly male₁ ₂ ₃ ₄. • Half of those proceeded against were white (49%, n=339). This has remained consistent since 2016/17. • Ethnicity is not recorded for 14% (n=99) - recording of self-defined ethnicity recording has slightly improved over time. Data there is suggests suspects are broadly proportionate of London. • Just over half (54%,n=381) were charged/cautioned for s2a stalking. The other 46% were charged with s4a stalking. • Perpetrators are most likely to be aged 25-44 (65%, n=1,818) - with 38% aged 25 to 34 years old. These patterns have remained the same since 2016/17 & are similar to victim profiles. Demographics of people proceeded against by MPS for stalking offences in 2022/23 1. Meloy, et al., 2000 2. Mullen et al., 1999 3. Purcell et al., 2001 4. Sheridan et al., 2001 The E&I deep coded dataset includes all suspects of stalking – not just those that are proceeded against. • Of these, nearly a quarter (n=90) of suspects are female. This suggests that male perpetrators are more likely to be proceeded against. • Suspect ethnicity is under-reported: Ethnicity not recorded for 43% of suspects in coded dataset. Coded cases MPS or PS London data 14% 15% 14% 4% 4% 49% Ethnicity (self-defined) Asian or Asian British Black, Black British, Caribbean or African Declines/ does not understand Mixed or Multiple ethnic groups Other ethnic group White 0.1% 11.1% 38.1% 29.0% 14.0% 4.9% 2.7% Age Under 18 18 to 24 25 to 34 35 to 44 45 to 54 55 to 64 65+ 92% 8% Gender Male Female The suspects are generally known to the police through previous allegations or convictions History of aggressive/controlling behaviour – PS London data (n=353) Coded cases PS London data shows DA offences prevalent among stalking perpetrators on probation: • Nearly two thirds (61%, n=217) were recorded as perpetrators of DA. • Similarly, 7 out of 10 stalking. perpetrators on probation (n=246) had some recorded problem or a serious problem with aggressive & controlling behaviour. • 3 out of 10 had a serious problem (n=106). Deep coded cases show repeat offending among stalking suspects: • 62% have previous allegations of another crime (n=234) • 16% (n=60) have been previously convicted of another crime • Prevalence of previous DA & violence (including sexual violence) allegations was substantial (36% & 30% of suspects respectively). • 3 in 10 suspects (29% n=110) had previous stalking or harassment-related allegations (including malicious communications). Of these: - 81% (n=89) had previous same victim stalking or harassment allegations. - 19% (n=21) had previous stalking or harassment allegations from a different victim. • 28% of deep coded cases had a suspect with a PNC ID. 104 suspects in the deep coded dataset had a PNC ID – analysis of these suspects’ criminal histories (charges) show: • Average of 9 previous charges in their background (range from 1 - 61). • Average of 1.6 charges for a stalking or harassment related offence, ranging from 0 to 28 charges. 61% had at least one charge for a stalking or harassment-related offence & 33% had at least previous 2 charges, & 7% had 5 or more charges for stalking or harassment. MPS or PS London data 62% 16% 29% 36% 30% Previous allegations Previous convictions Previous stalking or harassment allegations Previous domestic abuse allegations Previous violence allegations Repeat offending among suspects in coded cases 24% 40% 30% 6% No problem Some problem Serious problem Missing Risk of serious harm (RoSH) for offenders on probation Stalking perpetrators present considerable vulnerability Stalking perpetrators on probation display a wide range of vulnerability: • 28% of stalking perpetrators on probation (n=100) were on medication for mental health problems. • 16% (n=56) have received, are currently receiving, or will receive psychiatric treatment. • Over two thirds (68%, n=239) had either some problem, or a serious problem, relating to difficulties coping. • 53% had some or a serious psychological problem or depression. Nearly one third (31%, n=109) had a serious problem relating to having either self-harmed, attempted suicide, or had suicidal thoughts. • Three quarters (78%,n=277) had an identified need around relationships. PS London identified mental health needs of stalking perpetrators PS London data reveals more detail as to suspect needs. Note: PS London data refers to 353 active PS London cases in November 2023. This data therefore refers only to convicted stalking perpetrators. • In the deep coding - just over a third of suspects (n=131) were reported to have mental health issues. 29% of suspects were reported to have substance misuse issues. This likely underreports the issue given police recording. In 8 cases, suspect was sectioned/being treated at a mental health hospital on case closure. 3 in 10 stalkers on probation have a high or very high risk of serious harm: • Majority of stalking offenders (70%, n=246) on probation classified as at ‘medium’ risk of serious harm (calculated using the Risk of Serious Harm Assessment (RoSH)). • Majority of stalking offenders were assessed as being at ‘low’ risk of re-offending generally (71%, n=251) (according to the OGRS) & committing non- sexual violence offences specifically, (64%, n=225) two years post-probation (according to the OVP). Coded cases MPS or PS London data 6% 7% 6% 6% 6% 6% 21% 12% 10% 9% 31% 7% 47% 41% 32% 35% 17% 26% 40% 52% 50% 63% 70% Difficulties coping Psychological problems / depression Social isolation Offender's attitude to themselves Self harm / attempted suicide / suicidal thoughts Psychiatric problems No problem Some problem Serious problem Missing < 0.5% Deep coding shows that behaviours included within stalking vary & depend on relationships Coded cases Breaking down types of behaviour by victim-suspect relationship: • Current or ex-partners were most likely to perpetrate third party harassment (27%, n=74) or threaten self-harm (13%, n=35) than family members, other acquaintances or strangers. • There was no statistically significant difference between the number of stalking behaviour types for DA & non-DA related cases. • This contrasts slightly to previous research that found ex-intimate stalkers presented more types of behaviours, & with more severity than acquaintance or stranger stalkers.₃ • Strangers most likely to do surveillance (67%, n=20) & love-bombing (23%, n=7). Physical locations involved in offence: • Just under a third (n=114) of coded cases involved no physical location – involving only phone calls, texts or online contact. • Nearly half of cases (n=182) included offending at victim’s home, 12% at victims’ school or workplace, & 21% in other public locations, e.g. street, public transport. 1. ONS, 2023 2. Stevens et al., 2021 3. White et al., 2020 Number of stalking behaviour types: • Most cases (70%, n=262) involved at least two stalking behaviour types. • Average of 2.4 stalking behaviour types per case, ranging from 0 to 7 behaviours. • S4a stalking cases included significantly more types of stalking behaviour on average: 2.9 types compared to 2.2 types for s2a cases. • Suspects with previous allegations engaged in significantly more stalking behaviours on average (2.5 to 2.2 types). Coded cases involved a range of stalking behaviour types: • 66% of cases (n=249) involved phone calls or text. • 38% (n=143) surveillance, spying, following or loitering, & 34% (n=129) unwanted face to face contact. • Online contact observed less than expected – one quarter of cases (n=88). CSEW statistics suggest nearly half of stalking victims (44%) have experienced cyberstalking.₁ • S4a cases were more likely to include third party harassment, surveillance, unwanted face to face contact & any threats. 0% 2% 2% 3% 3% 7% 9% 10% 10% 16% 23% 24% 34% 38% 66% Sexual contact Physical assault Weapons involved Threats: sexual Threats: damage to property Forced entry Damage of property Love bombing Threats: self harm or suicide Threats: physical harm Online contact, cyberstalking etc Third party harassment Unwanted face to face contact Surveillance Phone calls or texts Types of stalking behaviour in coded cases 101 112 74 43 21 9 3 0% 5% 10% 15% 20% 25% 30% 35% 0 20 40 60 80 100 120 1 behaviour 2 behaviours 3 behaviours 4 behaviours 5 behaviours 6 behaviours 7 behaviours Number of different stalking behaviours involved in offence s2a stalking s4a stalking Deep coded cases provide insight into length of time stalking behaviours occur: • In two thirds of cases reported stalking occurred for 3 or less months (n=241). • Stalking behaviour occurred for more than a year in 14% of cases (n=51). • S4a cases were more likely to involve longer term offending. • As were cases in which any threats were made, unwanted face to face contact, & third-party harassment: 1. James & Farnham, 2003 Coded cases Reported escalation of behaviour: • Escalation of behaviour – either severity or frequency – was reported in half of cases (n=195). This includes cases in which the victim states that behaviour has escalated and/or the officer reports behaviour has escalated in the report details. • In a quarter of these (n=53), the victim stated escalation, whereas the police did not report any escalation. • Previous research suggests stalking behaviour occurring over shorter periods of time with quick escalation is an indicator of higher levels of violence₁. DA v non-DA S2a v s4a Any threats Unwanted contact 3rd party harassment Length of time stalking behaviour occurred, by types of stalking & behaviour •In two thirds of cases reported stalking occurred for 3 or less months (n=241). •Stalking behaviour occurred for more than a year in 14% of cases (n=51).•S4a cases were more likely to involve longer term offending•As were cases in which any threats were made, unwanted face to face contact, & third- 8% 20% 20% 16% 9% 7% 7% 4% 2% Not multiple incidents over time Less than 2 weeks 2 weeks to 1 month 1-3 months 3-6 months 6-12 months 1-2 years 3-5 years More than 5 years Length of time stalking behaviour occurred 51% 55% 54% 46% 38% 57% 43% 57% 41% 55% 37% 24% 32% 37% 43% 30% 40% 30% 43% 31% 13% 21% 14% 17% 19% 13% 17% 13% 16% 14% DA Non-DA s2a s4a Any threats of harm No threats of harm Unwanted face to face contact No unwanted face to face Third party harassment No third party harassment Long: more than a year Medium: 1 month to a year Short: less than a month Duration of stalking from time of report varies - longer duration linked to certain behaviours Continuation of stalking behaviour: • In a quarter of cases (n=93) stalking behaviour was reported as ongoing, with further offending reported during the investigation after initial reporting. • Over a third (n=141) were linked to other CRIS reports. Criminal justice response to stalking 1. Reclassification of cases 2. Top level view of police recorded case outcomes 3. Case attrition detail 4. Drivers of withdrawal 5. Drivers of police NFA 6. Stalking Protection Orders 7. Role of Stalking Threat Assessment Centre (STAC) 8. Stalking victims’ experience of the CJS The number of stalking offences in London with a charge /summons, caution or community resolution has decreased over time, & there is high victim withdrawal & police NFA in investigations • In 2022/23 – most stalking cases were either NFA’d by police (47%), or NFA’d after victim withdrawal (44%). • Majority of cases are NFA’d due to evidential difficulties. • Consistent with England & Wales1. • Proportion of stalking cases NFA’d due to victim not supporting further action increased between 2019/20 & 2020/21 & has remained consistent in last three years. The E&I Victim Withdrawal research compared outcomes across crime types using a bespoke data set from police recorded data (April 2021 – March 2022 reporting outcomes varied by offence type. In that research, the stalking offence type had the highest proportion of cases NFA’d by police during the investigation due to ‘victim supports, evidential difficulties’ - 35% of cases - compared to other VAWG offences. • Proportion of reported stalking offences in London resulting in a charge /summons, caution or community resolution has decreased. Proportion of stalking offences resulting in a charge/summons, caution or community resolution over time Home Office reclassification of stalking offences 1. Home Office, 2023a MPS or PS London data • MPS data from 2022/23 - only 9% of stalking cases resulted in a charge /summons, caution or community resolution. • In England & Wales - rate of stalking cases resulting in a charge /summons, caution or community resolution is lower than London. • According to Home Office data - in England & Wales 6.6% of stalking reported to police 2021/22 resulted in a CPS charge1. • A smaller proportion results in conviction - 1.4% in England & Wales 2021/22 resulted in a conviction1. 6% 4% 3% 3% 3% 3% 21% 22% 29% 41% 42% 41% 38% 43% 38% 36% 35% 37% 12% 13% 10% 8% 9% 9% 1% 1% 1% 22% 18% 19% 11% 10% 10% 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 Judicial outcome Other outcome Outcome 18 - Investigation complete: no suspect ID Outcome 15 - Evidential difficulties: suspect ID; victim supports Outcome 16 - Evidential difficulties: suspect ID; victim does not support Outcome 14 - Evidential difficulties: no suspectID; victim does not support 22% 18% 19% 10% 9% 9% 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 Stalking case police recorded outcomes The deep coding allows us to explore case attrition in more detail Coded cases Victim withdraws during investigation = 169 (45%) Police NFA (when the victim did not withdraw) = 155 (41%) Charged by CPS = 33 (9%) Submitted to the CPS = 42 (12%) Victim withdraws during legal proceedings = 3 Total coded stalking cases in dataset = 376 Legal proceedings • Victims withdrew during police investigation in 45% of cases. A further 10% of victims did not explicitly withdraw but stopped cooperating with/responding to police during investigation. • Only 8% (n=15) of coded cases which had victim withdrawal statements taken. • Police decided to close 41% (n=155) of investigations with ‘no further action’ (NFA). Of these - the suspect was not identified in 19% (n=29) of cases. • Just 12% of cases were submitted to the CPS. • Only 10 CPS charges were for stalking – 4 s2a & 6 s4a, despite police classification being stalking for all 33 cases. • Approximately half of CPS charges were for harassment (17/33). • Suspect sectioned or placed in a hospital for mental health reasons in 8 coded cases without charge or caution. Police investigation • Suspect arrested or interviewed in 28% (n=106): arrested in 26% (n=98) & interviewed in 26% (n=97) of cases. • Potential technology evidence was referenced in 40% of cases, but: - Suspect technology seized in 13% of cases (n=50) - Victim technology provided in 17% of cases (n=62) • Of the suspects arrested/interviewed half were bailed (52%, n55). • S-DASH risk assessments completed in only 7% (n=27) of coded cases, against best practice. • Recording of victim impact in the Victim Personal Statement was referred to in less than 10 cases. • 38% (n=141) of coded cases linked to other reports. • Of the 97 interviewed: - 25% (n=24) gave no comment. - 74% (n=72) denied or - minimised offence. - 26% (n=25) made counter allegations. See Appendix for more detail Predictors of victim withdrawal • Victim withdrawal less likely (9x less) when suspect was a stranger. • Withdrawal less likely when the offence only included one incident • (x8 times). Overall aspects of the investigation were the strongest predictors of withdrawal. • S4a cases were more likely to withdraw (2x more likely). • When the suspect was arrested/interviewed – victim less likely to withdraw (8x). • Cases with witnesses were less likely to withdraw (7x less likely). • Where police report victim was frightened - lower likelihood (5x less likely). • Police special schemes (x3 time more likely) & police offer any other victim support (x2 times more likely) to withdraw. • Older victims (35-44 or 45-54) were less likely to withdraw (5x & 6x less likely) • Repeat victims of stalking (&/or harassment) were 5x less likely to withdraw. • Victims with previous allegations were less likely to withdraw. • VW 6x less likely where suspects aged 35-44 (compared to aged 25-34). • VW 5x less likely when suspects have previous convictions - potentially linked to suspects risk level. Coded cases Statistical modelling allows the Deep dive to explore what the significant predictors are of victim withdrawal… Block Explanatory variable Influence on likelihood for victim to withdraw from case during investigation (odds ratios, holding all other model variables constant) Victim characteristics Victim is aged 35-44 compared to 25-34 5 times less likely Victim is aged 45-54 compared to 25-34 6 times less likely Victim is a previous victim of stalking 5 times less likely Victim has previous allegations 7 times less likely Model fit for block specific model 0.1001068 Other, non-significant variables included in overall model Victim sex, victim ethnicity, victim reported to have mental health issues Suspect characteristics Suspect is aged 35 to 44 compared to 25-34 6 times less likely Suspect has previous convictions 6 times less likely Model fit for block specific model 0.1102639 Other, non-significant variables included in overall model Suspect sex, suspect has mental health issues, suspect was previously a victim Offence characteristics Offence included only one incident 8 times less likely Victim & suspect are strangers (compared to current or ex partners) 9 times less likely Model fit for block specific model 0.0667473 Other, non-significant variables included in overall model Offence involved a physical location, offence involved threats of harm, offence involved third party harassment Investigation characteristics Case classified as s4a 2 times more likely Suspect was arrested or interviewed 8 times less likely Case had witnesses 7 times less likely Police report victim feared for safety or was frightened 5 times less likely Police placed special schemes on victim's address 3 times more likely Police offer or provide any victim support to victim 2 times more likely Model fit for block specific model 0.304959 Other, non-significant variables included in overall model Police make reference to potential technology evidence, suspect was identified, SASH completed, case had linked CRIS reports Overall model fit 0.4147998 (n=334) Other captured details allows the deep dive to go further - not wanting a CJS outcome or unwilling to prosecute a common reason for victim withdrawal E&I coded data for amount of victim withdrawal matched trends in MPS data: • 18% withdrew within one day of reporting (n=30) – explicitly not wishing to prosecute at the time of reporting. • Withdrawal statements only obtained for 8% (n=14) victims who withdrew (n=169). • In 89% of cases where victim withdrew, police recorded it was due to the victim not wanting a CJS outcome or was unwilling to prosecute. • Victim advocates report that the victims often just want the behaviour to stop & to feel safe, rather than wanting to secure a prosecution. • Victim advocates highlighted a lack of trust in criminal justice agencies & processes; involvement in police investigations & legal proceedings can also be traumatising. Coded cases Victim withdraws during investigation = 169 (45%) 2% 89% 19% 6% Due to safety risk or fear of reoffending Not willing to prosecute or does not want a CJS outcome The behaviour has stopped Other Reasons for victim withdrawal 18% 18% 13% 20% 18% 11% Same day or next day as report made (n=30) First week of report (n=31) Second week (n=22) 2-4 weeks after report (n=33) 1-3 months (n=30) More than 3 months (n=19) % of victim withdrawals at different time periods Coded cases Predictors of police No Further Action Investigation factors again the strongest factors to predict NFA. • When victim technology was provided & cases have witnesses - police less likely to be NFA (both 10x less). • Suspect denying or minimising the offence was less likely to NFA (8x less). • Cases where special schemes set up on victims’ addresses were 7x less likely to be NFA. • DA related cases more likely to be NFA’d (x10 times) compared to non-DA related cases. Notably, no offence behaviours came up as statistically significant drivers of police NFA in the model. • Repeat victims of stalking (&/or harassment-related offences) less likely to be NFA’d (x7 times). • Consistent with other projects - victim recorded by police as vulnerable or perceived to have mental health issues is linked to higher (x7 times) likelihood of police NFA. Note: poor data collection is likely here - only 10% of victims reported to have mental health issues, & 14% recorded as vulnerable. • Male suspects 9x less likely to be NFA’d by police. • Suspect ethnicity unknown/not recorded is 11x more likely to be NFA’d by police. Statistical modelling allows the Deep dive to explore what the significant predictors are of NFA… Block Explanatory variable Influence on likelihood for NFA during investigation (odds ratios, holding all other model variables constant) Victim characteristics Victim aged 45 & over, compared to 25 to 34 10 times less likely Victim ethnicity is Asian or Arab, compared to white 9 times less likely Victim reported to be vulnerable or percieved to have mental health issues 7 times more likely Previous stalking victim 7 times less likely Model fit for block specific model 0.1471446 Other, non-significant variables included in overall model Victim sex, victim was previously a stalking victim Suspect characteristics Suspect aged 45 & over, compared to 25 to 34 17 times more likely Suspect is male, compared to female 9 times less likely Suspect ethnicity is unknown, compared to white 11 times more likely Model fit for block specific model 0.3439571 Other, non-significant variables included in overall model Suspect has mental health issues, suspect has previous domestic abuse allegations, suspect has previous convictions Offence characteristics DA related 10 times more likely Model fit for block specific model 0.1300463 Other, non-significant variables included in overall model Offence involved unwanted face to face contact, offence involved third party harassment, offence involved surveillance, spying, following or loitering Investigation characteristics Suspect denies or minimises offence 8 times less likely Victim technology provided to police 10 times less likely Case had witnesses 10 times less likely Special schemes set up on victims address 7 times less likely Model fit for block specific model 0.4061332 Other, non-significant variables included in overall model Offence classified as 2a or 4a, any police delays recorded Overall model fit 0.5997239 (n=196) Availability of evidence affects NFA • Police cited lack of evidence as a reason for NFA in 57% (n=89) of coded reports. • In 24% of these - non-cooperation from victims cited as NFA reason. In practice, key stakeholder interviews indicate victimless prosecutions are not possible in stalking cases as evidencing victim impact is key to offence. • In 18% of NFA cases - police noted a non-criminal justice outcome would be more suitable for the case, in 5 of these cases, MPS gave suspect a warning or ‘strong words’. In 4 cases suspect was sectioned. • 42% of investigations where the victim didn’t withdraw were NFA’d in first four weeks of reporting, with 19% in first week of reporting. Coded cases Police closed a further 41% (n=155) of investigations with ‘no further action’ (NFA). 4% 15% 10% 13% 23% 13% Same day or next day as report made (n=9) First week of report (n=31) Second week (n=20) 2-4 weeks after report (n=26) 1-3 months (n=47) More than 3 months (n=27) % of police NFAs (excluding victim withdrawal) at over time (n=155) 57% 24% 18% 12% Lack of evidence or evidential issues (n=89) Non-cooperation from the victim (n=37) Police feel non-CJS outcome would be most suitable (n=28) Suspect not identified (n=19) Reasons for police NFA (excluding victim withdrawal) (n=155) Reclassification of cases after reporting & at charge is common Initial classification Final police classification CPS or police charge Classification changes for 33 charged coded cases CPS changed charge to non-stalking offence in 25 of 33 charged deep coded cases. • Key stakeholder interviews suggested CPS favour prosecuting for harassment due to misunderstanding some stalking behaviours & perceived difficulties in prosecuting 2a stalking at court. • 2023 survey of victim advocates - 77% did not believe CPS had adequate understanding of stalking, linked to victim withdrawal1. • Key stakeholders noted CPS charge for breaches of orders without charging for stalking. Occurred in 7 of our coded cases. • MPS STAC officers meet with CPS North & South monthly to joint review cases & determine whether charges correct, this helps highlight issues. Inspection of police & Crown Prosecution Service found stalking was misunderstood by police & CPS, & often improperly recorded (HMIC & HMCPSI, 2017). “Unless it was… traditional stalking behaviours in a 2A … such as following & spying, CPS would probably rather go for basic harassment, because of trying to get through to a jury that this is stalking” Key stakeholder 8 in 10 coded stalking offences were not initially identified as stalking by responding officers. 42% were initially classified as harassment. Our key stakeholders highlighted several reasons: • National stalking legislation is very similar to harassment & awareness of 2019 Home Office changes is insufficient. • Police case recording systems make identifying stalking offences by linking cases harder. • Understanding of stalking may vary across police teams. Potential uncertainty around stalking definitions in deep coded cases. • One or more OICs stated either no crime or no stalking had occurred in at least 20% of coded cases (n=73). A recent Suzy Lamplugh Trust survey of stalking advocates found lack of police understanding of stalking behaviours was either common (55%) or extremely common (45%)1. 2a or 4a offence is based on officer’s judgement of victim impact, but officers often unable to distinguish. • Experts feel 4a classifications are ‘underused’ with consequences for risk management. • Victim impact often not referred to in deep coded cases: police reported victims experiencing emotional distress/mental toll, frightened or fearing for safety, due to offence in 54% of coded cases. Rates higher for DA-related offences (56%), compared to non- DA related (46%). • Interviewed stalking advocate & experts & the National Stalking Consortium super complaint in 2022 argued for a single stalking offence as all stalking engenders victim fear, alarm, & distress & the current two-tier framework fails to protect victims2. 1. Suzy Lamplugh Trust, 2023 2. Suzy Lamplugh Trust, 2022a. Coded cases Gaps in monitoring & enforcement of Stalking Protection Orders (SPOs) observed despite some recent improvements in SPO process Recent improvements to SPO process: • Key stakeholder interviews indicate the MPS has: - escalated issues with granting SPOs to HMCTS/judiciary (courts rarely grant both SPOs & restraining orders, Breach of bail conditions does not have same repercussions as SPOs or ROs). - improved ability to obtain SPOs, with new support for applications through Risk Management Units & dedicated SPO SPOCs in BCUs. • New Home Office guidelines for SPOs2 have lowered criminal standard required to obtain SPOs. • In practice, victim cooperation needed to obtain SPO: technically SPO applications can be victimless, but not in practice, which can be a barrier. Recent guideline changes may help this. • Early stalking intervention programme: multiagency intervention delivered by Police, NHS and Suzy Lamplugh Trust to those on SPOs to address stalking behaviours early:​ • Police applications and monitoring of SPOs.​ • Specialist Independent Stalking advocacy to ensure that victims’ needs are understood throughout the criminal justice journey.​ • Low-intensity mental health interventions for perpetrators. Enforcing orders Often not clear who is monitoring SPOs. Key stakeholder interviews highlighted: • Different procedures across BCUs. • Examples of proactive enforcement of orders exist, e.g., East BCU POU calling victims to monitor SPOs, but perceived under- resourcing limits active monitoring. • Every SPO should have an allocated officer – this is under review. • Perceptions of a lack of proactive enforcement by the MPS means enforcing breaches often falls to the victim. • However, victims might not be able to identify some aspects, e.g., mobile phone usage. • Police often only reactively enforce SPOs, when victims report breaches – highlighted as national problem. • Victim court attendance required for SPO breach convictions. • SPOs are victim specific so victims will be involved in policing it to a degree, but it should not be victim’s priority. • The MPS have a key performance target of issuing stalking protection orders in 2% of stalking cases. • SPOs were obtained in 1.3% of MPS recorded stalking offences in 20211 & 1.4% of stalking offences (s2a & s4a) in 2022 (n=164) (according to MPS data). Data was not available for 2023. • Though SPOs in London are low, they are higher than elsewhere. In 2021, 87 SPOs were issued in the Greater London area, more than twice the next highest area, Kent (32 SPOs issued).3 • In our coded cases from 2020-2021, SPOs were obtained in 2% of cases, while non-molestation orders were obtained in 7%. Obtaining orders STAC have a role in reviewing breach cases, & can offer OIC support • STAC reviewed 906 breach of orders reports in 2023. 15% (n=137) of these were breaches of SPOs, while the majority (70%, n= 634) of these were breaches of ROs. • However, in only 6% (n=52) of coded cases did STAC offer advice to BCU SPOC. Victims reported: difficulties obtaining SPOs, mis-information about SPOs from police, concern perpetrators will breach them.​ 27 9 6 1 NMO obtained SPO obtained RO obtained DVPN/O obtained Orders obtained during investigation – deep coded cases ‘Every officer in a stalking case should be saying why they are not applying for one – this isn’t happening – some officers don’t know what they are.’ Key stakeholder 1. MOPAC, 2022b 2. Home Office, 2024 3. Home Office, 2023b STAC is a leading stalking resource, but knowledge may not be reaching OICs & the CPS STAC case study: benefits of multi-agency working Individual had a delusional fixation & paranoid beliefs about a neighbour who he thought he was in a relationship with, presenting as both a resentful stalker & intimacy seeker (according to the stalking typologies). STAC discussed ongoing case for 6 months at multi-agency meetings. STAC professionals attended multi-agency public protection arrangement (MAPPA) & other meetings and supported him into a secure hospital through liaison diversion. STAC supported an SPO for victim who was referred to a Suzy Lamplugh Trust advocate. STAC PS London supported perpetrator’s probation officer with PSR. Made sure case wasn’t closed to ensure criminal justice route & mental health route taken at same time – best practice response. STAC experts highlighted role of STAC in creating system changes within various agencies, upskilling & building capacity through advice, consultations, & training. Evidence of engagement between OICs & STAC officers in deep coded cases was limited, though this engagement may not have been routinely recorded on CRIS. • In coded dataset, 91% of police reports (n=343) had evidence of input from STAC MPS officers. Yet, input mostly involved copying & pasting information about stalking, advice on handling stalking cases e.g. use of SASH risk assessment, & how to access support from STAC. Evidence of OICs following this advice in CRIS reports was limited, e.g., S-DASH’s were completed in 7% of cases. • Little evidence of OICs actively engaging with STAC MPS officers: only in 6% of cases (n=24). Evidence of OICs engaging with other STAC partners e.g. the NHS or PS London in only 1% (n=5) of cases. Note: coded case sample comes from June 2020-2021, & there may have been changes since this time. Key stakeholder interviews highlighted front-line MPS & PS London officers are under-resourced with high workloads. • PS London reliant on services approaching them for advice & consultations – meaning there may be MPS officers who would benefit from this but not asking. STAC stakeholders felt that their multi-agency working helps all agencies better respond to stalking, through quick information sharing, & drawing on each-other’s resources. • Daily management meetings & co-location of all agencies allow for efficient multi-agency working, sharing of expert advice, including victim perspective through Suzy Lamplugh Trust. Challenges at STAC • Perceived mismatch in resources between agencies at STAC. STAC police officers are looking into increasing (doubling) their resource, but the perception is that other agencies are already at top capacity. • Sense of power mismatch by victim advocates vs. other (perpetrator-focused) partners. Suzy Lamplugh Trust excluded from some (Operation Griffin) meetings due to being a 3rd sector agency. • STAC MPS officers’ daily review of cases difficult due to poor information systems & data collection. Requires many manual searches to highlight & find risk that is possibly missed elsewhere. Currently considering alternative models for case reviews due to increased demand from change in HO guidelines & increased reporting of stalking, e.g., localised BCU ‘clinics’, like Hampshire/Cambridgeshire. See the appendix for a description of how STAC inputs into each stage of the criminal justice journey. Stalking victims’ experience of the CJS was poor Examples of poor advice - that goes against College of Policing best practice & may undermine investigations or increase risk - also found in deep coded cases. E.g., change phone number, block perpetrator on social media or ignore behaviour. Victim advocates told us victim blaming in stalking is common. From the literature, many stalking victims feel blamed, not taken seriously by police, & reported inappropriate police response2. OICs recorded doubting victim reliability in 7% of deep coded cases (n=26) which is lower than previous MOPAC deep dive products. Ex-partner stalkers often viewed as less dangerous than stranger/acquaintance stalkers3. Victims doubt the ability of the CJS to tackle stalking: • Some participants felt that legal sanctions – including bail conditions, orders & prison sentences – are not enough to deter or stop stalking behaviour. One participant stated: ‘there will never be justice’ for them. Victims reported instances of poor advice from police: • For some participants, responding officers did not identify stalking behaviour, or were told that an incident did not amount to a crime, with patterns in behaviour not identified. It is ‘lucky’ to get an officer who understands & responds. • Victims belittled or not taken seriously by police, told to simply ‘let it go’ or ignore the perpetrator. Some feel the threshold for police to act is too high. • Sometimes given advice that they felt put them in more danger. • Mis-information about SPOs. Poor communication from investigating officers & courts a significant issue for victims – highlighted as linked to lack of trust, frustration & upset: • Lack of continuity of police officers on cases. • Some participants did not know if their suspect was charged or not. Court adjournments not communicated in a timely manner is frustrating. • Victims do not get to speak to the prosecuting barrister, and so worry that important information may be missed as a result. Victims lack confidence in police investigations, due to poor experiences: • Feeling that police do not prioritise stalking cases. Victims feel like they must push for action in cases: ‘victim as their own policeman’. • Concerns that when police have not initially identified stalking correctly, case is not investigated properly, e.g., correct statements not taken. • Concerns that police have missed things or haven’t logged or linked offences properly, & this will undermine a court case. Victims who took part in victim commissioner interviews highlighted negative experiences of the CJS: STAC stakeholders told us that unless underlying causes of stalking are addressed – e.g., fixation & obsession – stalking behaviour may continue even with a judicial outcome, meaning victims do not feel safe at case closure. Stakeholders maintained that the criminal justice process is key aspect of addressing stalking, in part, as services can input to support victims & address underlying causes of stalking, in particular mental health or behaviour interventions. The overall impact of the CJS process on victims: • Experience with both the MPS & CPS mentally draining: the CJS process can be as traumatic as the stalking itself. • Lack of communication about cases from CJS agencies & delays due to court adjournments highlighted as key factors contributing to a poor experience. 1. Weller et al., 2013 2. Taylor-Dunn et al., 2018 3. Chung & Sheridan, 2021 Lack of understanding about stalking as an offence & CJS processes among victims: • Some participants did not realise they were experiencing stalking at first. • Gaps in understanding of police investigations, CPS processes & rights as a victim. • Lack of understanding exacerbated by lack of updates & responses from agencies. • Some participants felt victim advocates could do more to explain CJS processes. Victim advocates highlighted a lack of awareness of stalking among victims, perpetrators & wider public, particularly for non-traditional stalking behaviours. Previous research on public perceptions of stalking also found lack of understanding & stereotypes around stalking behaviours1. Appendices 1. Deep coding method 2. Criminal justice journey 3. University of West London (UWL) systematic literature review 4. Perpetrator intervention programmes across London 5. Victim support services across London 6. STAC CJS input 7. Definitive assessment on disproportionality not possible 8. 2023 strategic review of MPS High Harm Offender Management 9. References Appendix 1: E&I deep coding of MPS crime reports • Due to poor or limited MPS data recording - E&I conducted a deep coding exercise of 400 MPS crime reports, to better understand: • Stalking victims • Perpetrators • The nature of stalking offences • Stalking investigations & procedures • Drivers of police no further action (NFA) & victim withdrawal. This replicates methodology from MOPAC research in the London Rape Review (2019) & DA deep dive (2022) • Extracted from the MPS Crime Recording & Information System (CRIS): 10,436 stalking-related cases from July 2020 - June 2021. • These extraction dates ensured enough time had elapsed for case conclusion. • 6,082 CRIS reports were isolated: • Section 2a stalking offences (72%) • Section 4a stalking offences (27%) • Breach of stalking protection order (SPO) (0.4%). • Random sample of 400 cases were selected, representative of the full dataset in terms of proportion of different stalking offences. • E&I researchers coded 400 reports for key victim, suspect, offence, investigation, legal proceeding & outcome characteristics. • Final caseload was 376 cases, after removing unsuitable cases (see diagram to the right). • Coding exercise produced a rich coded dataset, allowing more detailed & robust analyses of stalking cases compared to using MPS data alone. • It is not something the MPS can standardly run for ongoing oversight - this is crucial to address, as without a process for monitoring victim engagement, any improvements will be hard to measure. • The following slides mainly draw on findings from the coded dataset to give a richer understanding of stalking in London. • The next section - ‘Understanding criminal justice journeys’ - presents results from analysis of the stalking dataset to understand the main drivers of 1) stalking victim withdrawal, & 2) police NFA. Coded cases Entire data pull: 10436 cases • All stalking-related CRIS reports from 1st July 2020 to 30th June 2021 Stalking offences only: 6082 cases • Inclusion of CRIS reports with current classification of 2a stalking, 4a stalking or breach of SPOs • 72%% 2a, 27% 4a, 0.4% breach of SPO Random sample: 400 cases • Random sample of 400 selected • 72% 2a, 28% 4a, 0 breach of SPO Final caseload: 376 • 4 deleted: (3 x no allegations made, 1x restricted report) • 22 duplicated reports removed, 4 added Appendix 2: Criminal justice journey (Part 1) Police investigation 8 in 10 stalking offences (n=307) not initially identified as stalking by responding officers in coded dataset. Same trend found in MPS data. Majority of reclassified cases initially classified as harassment (41%) or domestic incident (24%). Of the 18% (n= 69) cases initially identified as stalking, 7 cases reclassified in seriousness of offence: 4 from 2a to 4a, & 3 from 4a to 2a. Victims made a report in 88% of coded cases. 2022/23 MPS data - majority of reports made by phone call (69%, n=7950), 13% in person & 7% online. Police investigations & legal sanctions Initial police response Suspect arrested or interviewed in 28% (n=106) of cases: arrested in 26% (n=98) of cases & interviewed in 26% (n=97) of cases Suspect not identified in 13% of cases (n=49) Of the 106 suspects arrested or interviewed: 52% (n=55) bailed & 15% (n=16) remanded. Of the 97 interviewed: 25% (n=24) gave no comment interview 74% (n=72) denied or minimised offence 26% (n=25) made counter allegations of stalking 74% 5% 1% 1% 1% Suspect identified same day or within next day of report 2-7 days 1-2 weeks 2-4 weeks More than 1 month Time to suspect identification 38% (n=141) of coded cases linked to other reports; 25% reported further offending - third of which had official crime report. Coded cases Appendix 2: Criminal justice journey (Part 2) Police investigation cont... Suspect technology seized in 13% (n=50) of cases –nearly always at arrest. On average suspect technology seized 37 days after crime report initiated (n=50). Third party evidence (including CCTV) was seized in 9% (n=33) of cases. 24% (n=91) of cases had a witness recorded. Victim technology provided in 17% of cases (n=62) – most often screenshots of text messages or emails. On average victim technology provided 23 days after crime report initiated (n=62). In 16 cases - MPS requested evidence from victim & victim did not provide. In 6 of these - victim refused to provide requested technology evidence, e.g., because of the time they would be without a mobile phone. Police investigations & legal sanctions Coded cases In 40% of cases potential technology evidence was referenced, but…. 6% 14% 4% Delays in obtaining evidence Delays in police work Delays due to COVID Recorded police delays Evidence of BCU SPOC for stalking input in 30% of reports (n=114). SPOC input >50% of cases only evident in two BCUs: West BCU (68%, n= 27) & North BCU (57%, n=20). SPOCs instated a few years ago - aim to disseminate specialised stalking knowledge to BCU officers on stalking cases. Many SPOCs sit within RMUs & are SPO coordinators - supporting OICs with SPO applications. 91% of police reports (n=343) had evidence of input from MPS STAC officers. Little evidence of OICs actively engaging with STAC MPS officers: only in 6% of cases (n=24). Evidence of OICs engaging with STAC partners e.g. NHS / PS London in 1% (n=5) of cases. Police investigation cont... Appendix 2: Criminal justice journey (Part 3) Police investigations & legal sanctions Coded cases No victim distress or fear reported in 46% (173) of cases. Victim emotional distress or mental toll recorded in 31% (n=118). Victim frightened or fearing safety recorded in 42% (n=161). Recording of victim impact in the Victim Personal Statement was referred to in less than 10 cases. S-DASH risk assessments completed in a minority of coded cases (7%, n=27). Police offered or provided some kind of victim support to 56% of victims. Victims declined victim support in 35% of cases (this includes victims already receiving support from another service). In 9% of cases, victim or suspect was rehoused – in 7% of cases this was the victim. Note: in many more cases, victims rehoused themselves Special schemes, e.g., panic alarms, set up on victim addresses in 19% (n=) of cases. Protective orders obtained in 11% of coded cases. • SPOs obtained in 2% (n=9) of cases. • Non-molestation orders more frequently obtained (7%, n=27) – usually with external support e.g. National Centre for Domestic Violence (NCDV). • The User Satisfaction Survey does not capture detail from stalking victims. However, there is a wider survey designed to capture this feedback. However, dissemination & completion is low. • Victim dissatisfaction with police was recorded in just 6% (n=23) of coded police reports, potentially underreported. • Victim advocate interviews highlighted several reasons for stalking victim CJS dissatisfaction: • Victims report poor responses from agencies e.g. police not taking seriously. Leads to low CJS confidence. • Lack of updates on investigation - partially due to resourcing: OICs may have up to 60 cases. • Communication during court process - reported as poor & confusing for victims1. • Traumatising experiences - facing perpetrators in court & being cross-examined. • Victims feeling unprotected & demoralised when stalkers receive just a fine or suspended sentence. • Length of time until trial means remanded perpetrators may have serviced enough time & are released. • Victims do not feel supported by criminal justice agencies. Support services don’t seek victim input. 1. Suzy Lamplugh Trust, 2023 7% 2% 19% 56% 14% 11% Victim rehoused by official agency Suspect rehoused by official agency Special schemes set up on victims address Any victim support offered or provided MARAC referral Order obtained Police response to victims 7% 2% 2% 0.30% DVPN/O obtained RO obtained SPO obtained NMO obtained Appendix 2: Criminal justice journey (Part 4) Legal proceedings The CPS decided to charge in 33 cases. Length of time to CPS submission – cases submitted once Just 42 cases (12%) submitted to CPS: 30 submitted once, & 12 submitted at least twice (missing for 2). On average it took 71 days from initial reporting to final CPS submission (missing for 12). In the 6 cases where CPS decided not to charge, 2 were classified by police as s2a & 4 as s4a. In addition to CPS charges, 4 cases received a police charge, & 3 a police caution. One charge was for breach of non-molestation order, & one for s2a stalking. Police investigations & legal sanctions 2 5 0 5 9 9 0 2 4 6 8 10 Same day or next day as report made First week of report Second week 2-4 weeks after report 1-3 months More than 3 months Coded cases Appendix 3: Evidence base for early intervention limited. Evidence of effectiveness found for some post hoc interventions The Evidence and Insight unit commissioned an academic team led by the University of West London (UWL) to conduct a systematic literature review of best evidence on interventions, services, tactics, and partnerships aimed at reducing or stopping offending across a range of VAWG offence types, including stalking. Sources included: • Peer-reviewed systematic literature reviews, meta-analyses, randomised control trials, rapid evidence assessments and VAWG related violence risk assessment manuals. • Publication written in English. • Articles published between 2012 and 2023. The stalking category included: • 4 reviews exploring perpetrator risk factors (3 systematic reviews and 1 meta-analysis), all of which rated high-quality by the authors of the report. • 4 articles investigating stalking interventions (1 systematic review, 2 rapid evidence assessments, 1 randomised controlled trial). The UWL literature review finds • Prior stalking is one of a range of offending behaviours that make the odds of relationship aggression more likely1. • The most evidenced risk factors that make individuals more likely to engage in stalking are: - mental health problems and addiction, - Relationship problems, - trauma or childhood victimisation. • Those who stalk a stranger or an acquaintance are more likely to present with mental health problems than those who stalk ex-partners2. Stalking interventions evidence base The UWL literature review found a gap in the evidence base for protective factors against stalking, which suggests a holistic understanding of the underlying factors in stalking and therefore informed opportunities for early intervention is limited. The review found some evidence for the impact of: • Protection orders (PO)3,4 - Reduce the severity of re-victimisation (not prevalence). - Increase likelihood of arrest and charge of intimate partner stalkers when PO already in place. - More effective when victims do not have ongoing contact with perpetrator. • Dialectical Behaviour Therapy (DBT) and anger management-based CBT - Found as the most robustly evaluated intervention for stalking perpetrators, though evidence base is still limited. - Evidence of effectiveness compared to standard intermate partner violence interventions.2,4 • Acceptance and Commitment Therapy - Evidence of effectiveness compared to the ‘Duluth approach’ where stalking behaviours occur within romantic relationships.5 1. Clemmow et al., 2023 2. Leigh and Davies, 2022 3. Dowling et al., 2018 4. Johnson et al., 2020 5. Zarling & Russell, 2022 Rapid evidence assessment Appendix 4: Monitoring high-risk perpetrators: STAC, Operation Griffin & the PS London • Higher risk Tier 3 nominals placed on STAC database Operation Griffin (list of highest harm/risk/repeat stalkers). Can include non-convicted or potentially dangerous perpetrators, perpetrators on probation or managed by MAPPA, perpetrators nearing release. Reoffending is tracked through Operation Griffin. • Operation Griffin nominals have PNC marker, alerting STAC to any activity. Monitored by police, PS London & NHS at STAC in Operation Griffin meetings. Approx. 300 individuals on Operation Griffin database at any time (according to STAC information provided in 2022). • Operation Griffin meetings: • Attended by MPS, NHS & PS London only, involve intensive multi-agency management. • Like ‘mini-MAPPAs’. • Allow early intervention - prevent reoffending/re-victimisation. • Exclude victim advocates – for confidentiality reasons. Suzy Lamplugh Trust note this decision excludes victim voice, which can benefit & improve cases. • Impact: positive reduction of 50-60% of harm in 18-month window. Three-tier system for STAC’s response to stalking cases. Depends on level of STAC input, perpetrator risk & need: 1. Tier 1: case reviewed & advice given to MPS officer in command or probation officer. 2. Tier 2: multi-agency work undertaken by one STAC partner, e.g., specialist consultation with OIC or risk assessment etc. 3. Tier 3: (previously divided into 3 & 4) more intensive multi-agency working by multiple STAC partners, e.g. attending MAPPA meetings, perpetrator interventions etc. STAC data used in this report is a snapshot of the caseload in 2023, a later period to the sample of coded cases • According to STAC data – In 2023 MPS officers reviewed 4,533 s4A stalking reports. • Cases discussed in multi-agency Daily Management Meetings, with different expertise brought by different partners. • See the appendix for more information on STAC partners’ roles. Monitoring of convicted stalking perpetrators on probation • From PS London ‘snapshot’ data, in November 2023 there were 353 offenders on probation for stalking offences. • Almost two thirds (63%, n=221) convicted of section 4A stalking offence. • A third (32%, n=114) convicted of section 2A offence. • Only 5% (n=18) convicted of SPO breach. • Key stakeholder interviews highlighted Probation resourcing, Probation Practitioner knowledge, & availability of stalking specific structured or accredited interventions as points for consideration. • PS London at STAC support frontline probation officers. 3811 634 15 73 STAC tier level - reviewed 4a cases in 2023 Tier 1 Tier 2 Tier 3 and 4 Missing Appendix 4: Perpetrator interventions: psychological/ behaviour change programmes Research suggests psychological interventions that address underlying stalking causes are critical for preventing reoffending1. • Legal sanctions alone do not address underlying causes & may exacerbate problem. • Perpetrator interventions need to be multi-disciplinary & tailored, given differences in stalkers’ motivations, behaviours, psychological characteristics & need2. Challenges to provision of stalking perpetrator interventions in London: • Lack of stalking interventions outside of STAC. PS London does not have specific stalking behaviour programme. DA-related stalkers referred to Building Better Relationships DA accredited programme. • Lack of collaboration between STAC & DA perpetrator interventions, including pan-London DRIVE - missed opportunity. • Lack of earlier intervention programmes: STAC attempt to secure Home Office funding for psychoeducational programme (for lower-risk perpetrators receiving SPOs) was unsuccessful. STAC provide stalking perpetrator interventions on a pan-London basis, delivered by Barnet, Enfield & Haringey Mental Health NHS Trust • In 2023, 172 / 4,533 STAC suspects of s4a stalking were referred to NHS support. Perpetrator interventions at STAC involve: • In-depth perpetrator assessments, including clinically-focused Stalking Risk Profile which assesses risk, considers violence, recurrence & persistence. • Assessment driven tailored perpetrator interventions. • Intervention models such as CBT, acceptance therapy, distress tolerance, emotional regulation, interpersonal relationships, scheme therapy, ENDR. • Clinical supervision to discuss/review cases, intervention plans, psychological assessments. • Formulation of a relapse plan – emphasis on perpetrator as expert to identify warning signs of return to problematic thinking/behaviour. • NHS monitor STAC perpetrator outcomes on wellbeing, functioning, mood, reductions in rumination, & reoffending, working with PS London & MPS e.g. breaches of orders. Stalking perpetrator interventions are under-researched & lack robust studies assessing their effectiveness2. 2020 evaluation of MASIP pilot interventions (including STAC) found3: • Proportion of stalking cases ending in charge/summons higher for each police force with MASIP sites. • Not possible to determine if interventions provided by London STAC led to reduced re-offending due to how crime recorded & low number of perpetrator interventions provided. • Stakeholder interviews - multi-agency approach key to effective perpetrator interventions & perpetrators felt confident they had tools required to manage obsessive & fixated behaviours. • Economic analysis found a cost-beneficial effect for STAC when institutional costs (such as prison and secure hospital) are excluded. 1. Ostermeyer et al., 2016 2. Leigh & Davies, 2022 3. Tompson et al., 2020 Appendix 5: Victim support services STAC victim advocacy service: • Referrals mainly from STAC partners, also from national helplines & London Stalking Support Service • 2023 STAC data – 766 (17%) of 4,533 s4a stalking cases reviewed by STAC MPS officers referred to Suzy Lamplugh Trust for victim advocacy support. • Triaged using SASH assessment / professional judgement of risk & need. • For highest risk cases. • Risk indicators include; level of violence, involvement of children, current court case, stalking related break ins. • High proportion of vulnerable service users. • Two levels of support: • ‘Brief’ intervention – one off support, e.g. signposting services. • ‘Advocacy’ – long term support (max. caseload of 20). • Safety plans • Stalking legislation education. • Specialist services referrals, housing support, attending MARAC/MAPPA meetings. • Police advocacy & court support. • Advocates provide victim’s voice in CJS – raise concerns, obtain responses & information – helps victims cope with, & have better experience of, CJS & its process. • Advocates training in emotional & wellbeing support is important – due to impact of stalking. • Issues faced by the STAC victim advocacy service: • Not taken seriously by perpetrator focused STAC partners. • OICs not responding. • Court cancellations. DA support services for DA-related stalking. • Multi-agency support from MARACs – VAWG leads noted stalking related DA cases often discussed at MARACs • Stalking behaviour viewed as risk factor in DASH assessments. • Coded dataset - MARAC referrals made in 14% of cases. • Independent Domestic Violence Advisors (IDVAs) • National Centre for Domestic Violence (NCDV) – support applications for non-molestation orders. • Lambeth’s Gaia centre (provided by Refuge) – supports gender-based violence victims, including stalking. London stalking victim support services include: • STAC victim advocacy service – provided by Suzy Lamplugh Trust – highest risk cases • London Stalking Support Service (LSSS) – provided by Suzy Lamplugh Trust – lower risk cases • National Stalking Advocacy Service – provided by Paladin • National Stalking Helpline – provided by Suzy Lamplugh Trust • DA support services for DA-related stalking. Victim support services are under-researched. Limited evidence suggests that victim advocates support victims to continue with cases & self-manage risks: • Victims supported by stalking advocates are more likely to see stalkers charged & convicted1. • A small-scale study found victims believed advocacy services aided their ability to cope with the realities of stalking, made the victims’ journey through the justice system easier to navigate, provided them with the necessary emotional support & practical advice needed for their personal safety & to feel in control of their risk management2. 1. Suzy Lamplugh Trust, 2022c 2. Jerath et al., 2022 Appendix 6: STAC can input into all parts of the criminal justice journey STAC MPS PS London NHS Suzy Lamplugh Trust MPS officers at STAC review overnight stalking reports on CRIS, starting with 4A cases. STAC officers can contact relevant officer to advise & give specialist stalking advice. • According to STAC data of 4533 4a stalking suspects reviewed by STAC in 2023, STAC advised the BCU SPOC on 16% of cases (n=744). BCU officers & SPOCs can also contact STAC MPS for advice. In daily management meetings, PS London share relevant information on perpetrators if known to probation, which can support risk assessments. Liaise with police custody suites: could involve diversion, provision of an appropriate adult, advocating for the Mental Health Act assessment if suspect is in crisis. Support police in how to adapt interviews for suspects with mental health needs. Provide psychological assessments to support risk assessments. For suspects on remand, support prison in reach teams with mental health provisions in custody – might involve a hospital transfer, or care in the prison itself. If individual is in hospital, will liaise with hospital itself, e.g. provide info on risk elements like SPO & Restraining orders, concerns about ongoing contact etc, & nature of behaviour that has led them to where they are. Bridge the gap between client & police, liaising with OICs to find out about police actions & case progression. Help victims write their victim personal statement – emphasis to take back control by stating what impact the perpetrator has had on their lives. Can advise police on restraining order or SPO conditions. Make sure special measures are in place & make referrals to victim support services. Conduct consultations with probation officers to help draft pre-sentence reports (PSR), when judges or magistrates feel they need advice on what sentence to pass. They would review CPS papers & accused’s history & needs with the probation officer to make sure recommendations are suitable & take into account stalking behaviours. If not at custody threshold, STAC probation officer would advise a referral to a probation accredited intervention. NHS can consult PSR court reporters. Provide psychiatric reports & give opinions at court. Liaison with team and support of officer when able to be interviewed. Keep the victim updated throughout the process, & support expectations, especially as court cases can be heavily delayed. Talk victims through the court process to ensure they know how proceedings work. Pre-trial visits to familiarise with the court, including entrances & exits. Connect victims with court victim liaison officers. Can also attend court with victim. Police can assist with monitoring licence conditions – these might include monitoring use of phones & computers. Support community probation officers with stalking clients. They offer consultations to POs on typologies, concerns & risks, allowing a deeper reflection time on the specificities of supporting that individual & feed into risk management plans. If suitable, consultations can be multi-agency with other partners from STAC, especially NHS when mental health issues. They can also refer to behaviour change interventions at STAC, delivered by the NHS. May attend MAPPA meetings. Provide behaviour change interventions for perpetrators. Consultation with Probation staff, MAPPA attendance. Risk management advice, referral support, case assist and co-working with those in MH settings (e.g hospital order) with stalking behaviour or conviction. Police investigation Legal proceedings (Post) conviction Initial response Appendix 7: Definitive assessment on disproportionality in the CJS not possible from data available but victim & suspect ethnicity broadly proportionate with census data Suspects/accused Although not directly comparable as recorded differently and from different time periods, there is some indication of differences between perpetrator groups: • Female suspects appear to be less likely to be accused or convicted than male. • Asian & mixed or multiple ethnic heritage perpetrators appear to be more likely to be convicted. • Suspects under 24 appear to be less likely to be proceeded against. According to MPS data on 2699 stalking suspects who were proceeded against (received a caution or charge): • Females more likely to be proceeded against for s2A offence (57%, n=125) & males more likely to be proceeded against for s4A offence (53%, n=1,313). • Perpetrators of Asian or Asian British ethnicity more likely to be proceeded against for a s4A offence (47%, n=214). • In 2022/23 female perpetrators more likely to receive a caution (30%) compared to male perpetrators (11%) & less likely to receive a charge or summons (70% female vs. 89% male). • Between 2016/17 & 2022/23, 18-to-24-year-olds were age group most likely to receive a caution (20%, n=68). • White people proceeded against were most likely to receive a caution (16%, n=206) compared other ethnic groups. Suspects* Accused: charged or cautioned Convicted: on probation Coded dataset sample of n=380 July 2020- June 2021 MPS data (n=700) 2022/23 PS London data (n=353) Live cases extracted in November 2023 24% Female (n=90) 76% Male (n=289) 8% Female (n=56) 91% Male (n=639) 5% female (n=19) 95% male (n=334) 12% Asian (n=46) 15% black (n=58) 26% white (n=100) 3% other (n=10) Missing for 44% (N=166) 14% Asian (n=97) 15% black (n=107) 4% mixed/multiple ethnic groups (n=27) 48% white (n=339) 4% other (n=28) Not stated/refused: 15% (n=102) of cases 18% Asian (n=63) 16% black (n=55) 10% mixed/multiple ethnic groups: (n=35) 47% white (n=165) 5% other (n=17) Not stated/refused: 5% (n=18) of cases Under 18 years: 3.2% (N=12) 18-24 years: 19.1% (N=72) 25-34 years: 37.2% (N=140) 35-44 years: 24.5% (N=92) 45-54 years: 11.2% (N=42) 55-64 years: 3.5% (N=13) 65+ years: 0.8% (N=3) Over 35 years: 39.9% (N=150) Under 18 years: 0.1% (N=1) 18-24 years: 11.1% (N=78) 25-34 years: 38.1% (N=267) 35-44 years: 29.0% (N=203) 45-54 years: 14.0% (N=98) 55-64 years: 4.9% (N=34) 65+ years: 2.7% (N=19) Over 35 years: 50.6% (N=354) 18-25 years: 10% (N=35) 26-35 years: 40% (N=141) Over 35 years: 50% (N=176) Rape review 2019 DA Deep Dive 2022 Stalking Deep Dive 2024 N=501 N=277 N=378 April 2016 2019 July 2020-June 2021 White 66% 56% 60% Black 24% 22% 22% Asian 7% 19% 14% * Note there is no mixed or multiple ethnic group in the coded dataset. • Victim ethnicity in MPS data has substantial missing data. • Victim ethnicity in coded data is broadly in line with ethnicity of victims found in the 2019 MOPAC Rape Review & 2022 DA Deep dive, & with 2021 census data on ethnicity in London. 9% 9% 35% 3% 10% 34% Victim ethnicity (self-defined) 2022/23 (MPS data) Asian or Asian British Black, Black British, Caribbean or African Declines/Does not understand/not recorded Mixed or Multiple ethnic groups Other ethnic group 21% 14% 6% 6% 54% Ethnicity of London Residents (Census, 2021) Asian or Asian British Black, Black British, Caribbean or African Mixed or Multiple ethnic groups Other ethnic group White • Partnership working should be a priority for high harm offender management for effective identification, prioritisation, & management. • Address patchwork coverage of multi-agency arrangements to ensure that high harm offender management is consistent across London. • Continue exploration of multi-agency arrangements which go beyond crime- type siloes to address gaps. • Establish strategy for information sharing with external partners & police forces. Appendix 8: 2023 strategic review of MPS High Harm Offender Management MOPAC conducted a strategic review into MPS High Harm Offender (HHO) Management in 2023 & most findings are highly relevant to stalking, the table below sets out these overlaps. HHO strategic review finding theme Summary of 2023 HHO strategic review recommendations Where relevant to Stalking Deep Dive Governance & structure • Establish a central strategy & oversight framework to enable central overview of all groups of offenders who pose the highest harm. • Establish mechanisms for cross-strand/team working centrally & locally. • Overcoming siloed working crucial for monitoring delivery of minimum standards across MPS/BCUs for stalking investigations to assure application; improving investigations & victim response. Defining harm • (Re)define ‘high harm’ for the MPS. • Establish a central strategy for the prioritisation of high harm offenders. • Holistic view of offenders should be a priority. • Review registers, nominals lists & matrices of HHO used by MPS. For stalking offence types, success partly rests on: • Clearly defined minimum standards for recording victim impact (S-DASH & Victim Personal Statements. • Improved linking of data for quicker access to background information on victims/suspects, linking to previous incidents, to support identification of stalking (as a behaviour offence). • Clear understanding of any overlap between STAC nominal cohort & other HHO cohorts needed. Demand & resourcing • Establish an organisation-wide approach to talent management. • This research also identifies need to identify knowledge gaps & provide specialist stalking training for relevant CJS agencies. Systems & databases • Conduct a strategic review of all systems & databases relating to high harm offender management. Consider any blockages. • Update systems to ensure systems talk to one another & data can be easily extracted for performance monitoring purposes. Intelligence • Invest in intelligence functions (Esp. local level) to support identification. • Learn from best practice & innovations in intelligence tools across MPS. • Improve use of intelligence from partners & other police forces. • Important for recommendation on improved linking of data. Partnership working Partnership working is a key area of focus from this deep dive: • Continue cross-BCU working of stalking SPOCs to share & disseminate best practice & common issues. • Patchwork coverage of multi-agency arrangements not necessarily an issue in this stalking research as STAC generally available but improving use of this resource is seen as a priority by this research. • Continue efforts of collaboration between CPS & MPS on stalking outcomes. • Folding partners into application of minimum standards around victim communication. Judicial orders • Establish an organisation-wide strategy around the monitoring & enforcement of judicial orders. Roles & responsibilities around orders must be defined. • Review & evaluate the current MPS approach for use of orders. Ensure that appropriate scrutiny & oversight is in place. • This stalking research identifies gaps in the MPS obtaining, monitoring & enforcing SPOs & recommends minimum standards for this. • Improved linking of data important for stalking offences for quicker access to victim/suspect information, linking previous incidents, identifying stalking (as a behaviour offence). • Seek to understanding if new systems (e.g. Connect+) are addressing this. • Success of HHO recommendations for stalking rests on this research for improved MPS data recording, especially vulnerability, ethnicity & victim-suspect relationship. References Chung, K. L., & Sheridan, L. (2022). Perceptions of Stalking: Examining Perceivers’ Country of Origin, Perpetrator-Target Prior Relationship, & the Mediating Effect of Victim Responsibility. Journal of Interpersonal Violence, 37(21-22). https://journals.sagepub.com/doi/full/10.1177/08862605211042601 College of Policing (2023). Stalking or harassment. https://library.college.police.uk/docs/college-of-policing/Stalking-and-Harassment-2020.pdf Crown Prosecution Service. (2023). Stalking or harassment: Legal Guidance, Domestic abuse , Cyber / online crime. https://www.cps.gov.uk/legal-guidance/stalking-or-harassment Clemmow, C., Rottweiler, B., Marchment, Z., Wolfowicz, M., Kyprianides, A., Doherty, P., ... & Gill, P. (2023). Risk & protective factors for domestic abuse perpetration: Establishing the evidence base to support the prevention, management, and risk assessment of domestic abuse perpetrators in the UK. https://discovery.ucl.ac.uk/id/eprint/10165656/ Dowling, C., Morgan, A., Hulme, S., Manning, M., & Wong, G. (2018). Protection orders for domestic violence: A systematic review. Trends and issues in crime and criminal justice, 551, 1-19. https://www.aic.gov.au/publications/tandi/tandi551 Her Majesty’s Inspectorate of Constabulary & Her Majesty’s Crown Prosecution Service Inspectorate. (2017). Living in fear – the police & CPS response to harassment & stalking. https://hmicfrs.justiceinspectorates.gov.uk/publications/living-in-fear-the-police-and-cps-response-to-harassment-and-stalking/ Home Office. (2023a). Police recorded crime & outcomes open data tables. https://www.gov.uk/government/statistics/police-recorded-crime-open-data-tables Home Office. (2023b). Publication of Management Information on Stalking Protection Orders. https://www.gov.uk/government/publications/management-information-stalking-protection- orders Home Office. (2024). Stalking Protection Orders Statutory guidance for the police. https://assets.publishing.service.gov.uk/media/66222d8dcab76a6fd1757d73/Stalking_Protection_Orders_Statutory_Guidance_2024.pdf James, D. V., & James, Farnham, F. R. (2003). Stalking & serious violence. Journal of the American Academy of Psychiatry & the Law, 31(4) 432-439 https://jaapl.org/content/jaapl/31/4/432.full.pdf Jerath, K., Tompson, L., & Belur, J. (2022). Risk Management in Stalking Victims: A Multi-Agency Approach to Victim Advocacy. Journal of Interpersonal Violence, 37(11-12). https://journals.sagepub.com/doi/full/10.1177/0886260520980402 Johnson, H., Derefaka, G., & Wheatley, R. (2021). The effectiveness of interventions for male perpetrators of stalking: A Rapid Evidence Assessment. Forensic Update, 138. DOI: 10.53841/bpsfu.2021.1.138.64 Leigh, S. & Davies, J. (2022), A rapid evidence assessment of psychological treatment approaches for stalking behaviour, The Journal of Forensic Practice, (24)1, 48-62. https://www.emerald.com/insight/content/doi/10.1108/JFP-06-2021-0039/full/pdf?title=a-rapid-evidence-assessment-of-psychological-treatment-approaches-for-stalking-behaviour McEwan, T., Galietta, M., & Underwood, A. (2024) Treating stalking: A Practical Guide for Clinicians. John Wiley & Sons Ltd. https://onlinelibrary.wiley.com/doi/book/10.1002/9781119856771 Mayor’s Office for Policing & Crime. (2022a). Building a Safer London: Police & Crime Plan for London 2022-25. https://www.london.gov.uk/publications/building-safer-london Mayor’s Office for Policing & Crime. (2022b). Measuring the things that matter: Insights into police performance (A report produced for the Deputy Mayor for Policing & Crime). References Mayor’s Office for Policing & Crime. (2023) . Strategic Review: MPS High Harm Offender Management: Final report. MOPAC, London. [Internal report- unpublished] Meloy, J. R., Rivers, L. D., Siegel, L., Gothard, S., Naimark, D. & Nicolini, J. R. (2000). A replication study of obsessional followers & offenders with mental disorders. Journal of Forensic Sciences, 45, 147-152 https://drreidmeloy.com/wp-content/uploads/2015/12/2000_AReplicationStu.pdf Monckton-Smith, J., Szymanska, K. & Haile, S. (2017) Exploring the Relationship between Stalking & Homicide. Project Report. University of Gloucestershire in association with Suzy Lamplugh Trust, Cheltenham. https://eprints.glos.ac.uk/4553/1/NSAW%20Report%2004.17%20-%20finalsmall.pdf Mullen, P., Pathé, M., Purcell, R., & Stuart, G. W. (1999). Study of stalkers. American Journal of Psychiatry, 156(8), 1244 – 124. https://ajp.psychiatryonline.org/doi/full/10.1176/ajp.156.8.1244. Mullen, P.E., Pathé, M.T., & Purcell, R. (2001). Stalking: New Constructions of Human Behaviour. Australian & New Zealand Journal of Psychiatry, 35, 16 - 9. https://journals.sagepub.com/doi/abs/10.1046/j.1440-1614.2001.00849.x?journalCode=anpa Nichols, A. J. (2020). Advocacy Responses to Intimate Partner Stalking: Micro, Mezzo, & Macro Level Practices. Journal of Family Violence, 35, 741–753. https://link.springer.com/article/10.1007/s10896-019-00125-4#citeas ONS. (2023). Stalking: findings from the Crime Survey for England & Wales. https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/datasets/stalkingfindingsfromthecrimesurveyforenglandandwales Ostermeyer, B., Friedman, S. H., Sorrentino, R., & Booth, B. D. (2016). Stalking & violence. Psychiatric Clinics, 39(4), 663–673. https://www.psych.theclinics.com/article/S0193-953X(16)30048- X/abstract Parkhill, A. J., Nixon, M., & McEwan, T. E. (2022). A critical analysis of stalking theory & implications for research & practice. Behavioral Sciences & the Law, 40(5), 562–583. https://onlinelibrary.wiley.com/doi/full/10.1002/bsl.2598 Protection of Freedoms Act. (2012). https://www.legislation.gov.uk/ukpga/2012/9/contents/enacted Purcell, R., Pathé, M. & Mullen, P. E. (2001). Women who stalk. American Journal of Psychiatry, 158(12), 1953 – 2107 https://ajp.psychiatryonline.org/doi/epdf/10.1176/appi.ajp.158.12.2056 Richards, L. Fletcher, H, & Jewell, D (2012). Independent Parliamentary Inquiry into Stalking Law Reform: Main Findings & Recommendations. Scott, A.J. (2020). Stalking: How perceptions differ from reality & why these differences matter. In: Ray Bull & Iris Blandon-Gitlin, eds. The Routledge International Handbook of Legal ´ & Investigative Psychology. Abingdon: Routledge. https://research.gold.ac.uk/id/eprint/26591/1/2019%20Scott%20(Book%20chapter).pdf). Sheridan, L., Davies, G. & Boon., J. (2001). The Course & Nature of Stalking: A Victim Perspective. The Howard Journal of Criminal Justice, 40(3), 215 – 234. https://onlinelibrary.wiley.com/doi/epdf/10.1111/1468-2311.00204?saml_referrer Stalking Protection Act. (2019). https://www.legislation.gov.uk/ukpga/2019/9/contents/enacted Stevens, F., Nurse, J. R. C., & Arief, B. (2021). Cyber Stalking, Cyber Harassment, & Adult Mental Health: A Systematic Review. Cyberpsychology, behavior & social networking, 24(6), 367–376. https://kar.kent.ac.uk/83544/ References Storey, J. E., Pina, A., & Williams, C. S. (2023). The Impact of Stalking & Its Predictors: Characterizing the Needs of Stalking Victims. Journal of Interpersonal Violence, 38(21-22), 11569-11594. https://journals.sagepub.com/doi/10.1177/08862605231185303 Suzy Lamplugh Trust. (2021). Unmasking Stalking: A Changing Landscape. https://www.suzylamplugh.org/Handlers/Download.ashx?IDMF=fcfb781a-f614-48c8-adcf-4cfa830c16a7 Suzy Lamplugh Trust. (2022a). Super-complaint on the police response to stalking. https://www.suzylamplugh.org/news/super-complaint Suzy Lamplugh Trust. (2022b). What is Stalking? https://www.suzylamplugh.org/what-is-stalking. Suzy Lamplugh Trust. (2022c). Bridging The Gap: A Stalking Advocate for Every Victim. https://www.suzylamplugh.org/national-stalking-awareness-week-2022-bridging-the-gap Suzy Lamplugh Trust. (2023). ‘I just want this to be over’: Preliminary Findings of Research into Stalking Victims’ Experiences of the CPS, HM Courts & Tribunal Services & the Judiciary. https://www.suzylamplugh.org/Handlers/Download.ashx?IDMF=1c2de425-3272-4d4f-a913-59c45761dcca Taylor-Dunn, H., Bowen, E., & Gilchrist, E. A. (2018). Reporting Harassment & Stalking to the Police: A Qualitative Study of Victims’ Experiences. Journal of Interpersonal Violence, 36(11- 12). https://journals.sagepub.com/doi/10.1177/0886260518811423 Tompson, L., Belur. J. & Jerath, K. (2020). MASIP evaluation final report. UCL Jill Dando Institute of Security & Crime Science. https://discovery.ucl.ac.uk/id/eprint/10097009/1/MASIP%20evaluation%20final%20report%20v2.6.pdf Weller, M., Hope, L., & Sheridan, L. P. (2013). Police & public perceptions of stalking: the role of prior victim-offender relationship. Journal of Interpersonal Violence, 28(2), 320-339. https://journals.sagepub.com/doi/abs/10.1177/0886260512454718 White , E., Longpre, N., & Stefanska, E. (2020). Stalking Behaviors Presented by Ex-Intimate Stalkers: A Victim’s Perspective. Journal of Interpersonal Violence, 37(7-8). https://journals.sagepub.com/doi/abs/10.1177/0886260520934429 Zarling, A., & Russell, D. (2022). A randomized clinical trial of acceptance and commitment therapy and the Duluth Model classes for men court-mandated to a domestic violence program. Journal of consulting and clinical psychology, 90(4), 326. DOI:10.1037/ccp0000722 Notes • MPS stalking perpetrator data refers to those charged or cautioned only. Coded data refers to all suspects (including those not charged or cautioned). This is significant as only a small proportion of stalking suspects are charged or cautioned. In our coded dataset, only 12% of suspects (n=46) were proceeded against. • OASys (Offender Assessment System) Violence Predictor (OVP) estimates the likelihood of non-sexual violent re-offending (e.g., homicide & assault, violent acquisitive & weapon possession offences etc.) over a 24-month period using static (provided by the OGRS) & dynamic risk factors. Offender Management & sentence planning | Prison Reform Trust. • Offender Group Reconviction Scale (OGRS) is a risk assessment tool used to estimate likelihood of re-offending. It uses static factors such as age, gender & criminal history. It gives a score, which shows the likelihood of someone re-offending within a 12 & 24 month period. Offender Management & sentence planning | Prison Reform Trust. • ONS data referred to in the report are not designated as National Statistics & in addition, an error in the survey for year ending March 2023 resulted in missing data for some respondents from October to January 2023. Estimates for stalking victimisation are therefore based on just eight months of interviews. • Police recorded outcomes data referred to here is based on offences that were recorded in the given period. It excludes cases where an outcome was ‘not yet assigned’. • Risk of Serious Harm Assessment (RoSH) is a risk assessment used to assess the risk of ‘serious harm’ to others. The prison and probation services define risk of ‘serious harm’ as ‘a risk which is life-threatening and/or traumatic, & from which recovery, whether physical or psychological, can be expected to be difficult or impossible’. The RoSH assessment includes a variety of risks (risk of serious harm to others, risk of harm to children, risk to self & other risks such as escape & breach of trust) & provides a risk rating ranging from low to very high. • STAC Case data relates to Section 4A stalking reports (from the MPS) reviewed by STAC during the period 03/01/2023 to 31/12/2023. There were 4,558 cases in the original dataset. 25 cases were excluded (3 fell outside the date range & 22 cases contained no data). This left a final sample of 4,533 cases. April 2019 London Gang Exit: Interim Evaluation Summary and Next Steps This document presents a summary of the London Gang Exit evaluation to date, and outlines the plans for the final evaluation currently scheduled for early 2020. It should be read with reference to the suite of interim evaluation reports published alongside it on the MOPAC website. Programme Background London Gang Exit (LGE) is a holistic support service to help individuals involved in gangs or at risk of gang violence or exploitation. Originally jointly commissioned by the Mayor’s Office for Policing And Crime (MOPAC) and the London Community Rehabilitation Company (CRC), the intervention went live in February 2016 and was initially funded until October 2017. MOPAC took over sole funding of the project after this date and LGE is currently funded to March 2020. LGE also receives contributory funding from the Mayor’s Young Londoners Fund between 2018-20. The programme is being delivered by Safer London with key delivery partners Catch-22, Department of Work and Pensions (DWP) and the community CRC. The pan London service is designed to complement and enhance existing local services, matching need to bespoke support across mental health, employment, family support, housing advocacy and specialist support for girls and women. Young people are not compelled to work with LGE, but rather engage on a voluntary basis. The key overarching aims for LGE, measured through a combination of police recorded data and a practitioner/client ‘distance travelled’ assessment tool are: • Reduced / ceased harm from gangs • Reduced / ceased involvement in gangs • Reduced seriousness and frequency of offending Evaluation: Overview of Findings to Date MOPAC Evidence and Insight (E&I) are evaluating LGE in terms of the process of implementation, performance management, impact and value for money. To date, two interim evaluation reports have been completed, as well as three indicative analyses of offending and victimization to inform Payment By Results (PBR) awards over the life of the programme. Safer London routinely provides aggregated performance management information on to inform quarterly contract management meetings (e.g. throughput, service-use demographics, intervention uptake/completion and strategic and operational updates). • Since launch, over 500 young people have been referred onto the programme, with 134 having completed the full six-month intervention (as of Jan 2019). • Referrals have been received from all London boroughs, statutory agencies and self-referrals. • The average age of individuals completing the programme is 19; 70% are male; 89% BAME. Interim Evaluation Report 1: October 2016 The first MOPAC evaluation report presented key performance data and learning around implementation from a practitioner’s survey and interviews. Due to low throughput at this early stage, impact was not assessed. The report highlighted positives in training (upskilling) of practitioners; steady progression with referral numbers; positive engagement with Local Authorities and a shared understanding of current barriers between referral and intervention. Key challenges included staff shortages and inconsistencies in data quality and availability. Communications issues were also highlighted, including the name of the programme, which was considered potentially misleading, raising questions around the clarity of eligibility criteria. Interim Evaluation Report 2: October 2017 The second MOPAC evaluation report drew on key performance data to date as well as process learning from staff surveys and service users. The report also presented early indicative impact on offending and victimization, emphasizing caution in interpreting results at this early stage. In terms of process, staff strongly believed that LGE was providing short and long-term positive outcomes. Staff also indicated that significant barriers in initial delivery design had been overcome (e.g. centralization of referrals). Overall, service users also reported positive experiences, particularly in terms of staff engagement and flexibility of response. Analysis indicated a diverse intake of service users in terms of involvement in group violence, violent offending and victimization, reflective of the broad eligibility criteria. Offending was driven by less than half of the overall cohort, whilst one third were victims. Preliminary impact analysis found tentatively positive outcomes in terms of non-proven offending; there was a decrease in the proportion of individuals offending after the intervention (though not statistically significant). Two Year Performance By Results (PBR): December 2018 The most recent analysis explored police recorded non-proven offending and victimization cases closed between October 2016 and September 2017 inclusive (Year One) and October 2017 and September 2017 (Year Two). Again, findings are indicative only and low case numbers, restricted period of analysis, focus on completed cases only, relatively low prevalence of offending, and before and after methodology mean that robust causal statements as to LGE’s impact on offending behaviour are not possible at this stage of the evaluation. The key findings were as follows: • Across both years overall offending decreased, but not significantly. The decrease was driven by the Year 1 cohort. • Year 2 analysis found no impact on non-proven offending, including violent offending. • Consistent with other periods of analysis, victimisation and violent victimisation decreased both overall, and within Year 1 and Year 2 (all statistically significant). A greater proportion of Year 2 completions had no offending history before or after case opening, across all time-periods explored. Similarly, where offending is recorded, the Year 2 cohort includes less prolific individuals. The Year 2 cohort were more likely to be victims, and particularly victims of violence. This may in part be explained by changes in the types of individuals accepted (i.e. more females at risk of violence/exploitation). This will be explored in detail in the final evaluation report. The Final LGE Evaluation Report: Expected Early 2020 The final report will explore both the process of implementation and any impact of LGE against key outcomes. The report will be informed by interviews with staff, and clients. To allow for a more robust exploration of impact, primarily on police recorded non-offending and victimisation, analysis will utilise a matched control methodology to generate a comparison group. Data access and quality allowing, needs matching and outcomes of bespoke intervention pathways, as well as additional metrics from the distance travelled assessment and delivery partners, will be factored into impact analysis to best understand what works for whom. 1 The London Rape Review A review of cases from 2016 July 2019 2 Acknowledgements MOPAC would like to express their sincere gratitude to the research team from Evidence & Insight: Daniela Wunsch, Barry Charleton and Dr Julia Yesberg. Many thanks also to Dr Anthony Murphy and Dr Ben Hine from the University of West London for their support for this project. MOPAC would also like to thank Dr Paul Dawson and Mat Pickering for their advice throughout the research, as well as Fil Agrestini, Laura Duckworth, Charlotte Keeble, Rachael Parker, Jeannette Woodcock and Julita Siuta for the coding of the data. Many thanks to the Metropolitan Police Service (MPS) for supporting the coding team and providing useful guidance during this review. Finally, thanks to all those partner organisations who contributed to and supported this review with their insightful feedback: Crown Prosecution Service (CPS), Her Majesty’s Courts and Tribunals Services (HMCTS), National Health Service (NHS) England, The Havens, Women and Girls Network, Solace Women’s Aid, Rape and Sexual Abuse Support Centre (RASASC), NIA, Galop, Survivors UK and End Violence Against Women (EVAW). 3 Contents Executive Summary ......................................................................................................... 5 1. Introduction ......................................................................................................................... 7 1.1 Background ....................................................................................................................... 7 1.2 The London Rape Review ................................................................................................ 10 2. Methodology ...................................................................................................................... 11 2.1 Sample ............................................................................................................................. 11 2.2 Coding and data analysis ................................................................................................. 11 2.3 Limitations of the methodology ...................................................................................... 12 3. Results ................................................................................................................................ 13 3.1 Attrition of rape allegations ............................................................................................. 13 3.2 Who reports rape in London? .......................................................................................... 16 3.3 Suspect profile ................................................................................................................. 19 3.4 Types of rape being reported .......................................................................................... 20 3.5 Timeliness and evidential challenges ............................................................................... 23 3.6 Predicting attrition .......................................................................................................... 26 4. Discussion ........................................................................................................................... 31 4.1 Summary of key findings ................................................................................................. 31 4.2 Future research ................................................................................................................ 33 4.3 Conclusion ....................................................................................................................... 34 Appendices ............................................................................................................................. 35 References .............................................................................................................................. 41 4 84% of allegations were classified as a crime by police. 3% resulted in a conviction 58% of victims/survivors withdrew the allegation. 29% of cases the Police decided to take no further action 6% proceeded to trial At a glance … The London Rape Review The Rape Review is a collaborative piece of research between MOPAC and the University of West London seeking to drive insights into the nature of reported rape in London. 501 allegations of rape across London taken from April 2016 146 variables coded covering the entire case progression Almost three in five offences took place in a private or domestic setting. 18 months the average length of time from the date of reporting to trial outcome. 28% of all rape allegations were related to domestic abuse. 7% of cases were perpetrated by a complete stranger. The strongest predictors of victim withdrawal and police No Further Action were procedural characteristics. 5 The London Rape Review: A review of cases from 2016 Executive Summary About the London Rape Review The London Rape Review is a collaborative piece of research between the Mayor’s Office for Policing And Crime (MOPAC) and the University of West London (UWL). The overall aim of the research is to examine reported rape in London, with a specific focus on how key factors related to the victim/survivor, suspect and offence, as well as the subsequent investigation by the police, may affect different outcomes for victims/survivors and the progression of rape cases through the criminal justice system. A total of 501 allegations of rape in April 2016 were examined as part of this review. This research builds on previous reviews of rape in the MPS conducted between 2005 and 2012. Key Findings: Attrition of rape allegations • Victim/survivor withdrawal was the most common form of attrition in the sample of classified cases (58%), followed by no further action by police (29%). • The majority of classified cases (86%) were not submitted to the CPS. Only 9% were charged by the CPS, 6% proceeded to trial and 3% resulted in a conviction. • Victims/survivors who withdrew did so soon after reporting, and the majority within the police investigation stage. Reasons for victim/survivor withdrawal were complex and often interrelated. For example, some withdrew because of the stress and trauma of the police investigation. Key Findings: Who reports rape in London? • Victims/survivors were predominantly female (89%) and of a white ethnic background (57%). Two thirds were under 30 years old at the time of the offence. • There has been an increase in male reporting and an increase in Asian victims/survivors compared to previous years’ reviews. • 95% of victims/survivors had at least one need present (e.g. mental health, under 18, learning difficulty). Cumulative needs were common among victim/survivors, with just over one third of victims/survivors having two needs present, and one in five having three or more needs. • 41% of victims/survivors presented with a mental health issue: a significant increase from previous rape reviews. There was a significant relationship between mental health and previous victimisation. Key Findings: Suspect profile • The profile of suspects has remained reasonably consistent with previous years’ reviews. Almost all suspects were male (99%) and aged between 18 and 49 (77%). • Black suspects (35%) and those aged between 18 and 49 years (77%) were over-represented in the sample. On the other hand, white suspects (43%), those aged under 18 years (12%) and those aged over 50 years (12%) were under-represented. • A minority of suspects were recorded as having a prior police record (29%), or a history of domestic abuse (11%) or sexual assault/rape (9%). 6 • 28% of rapes reviewed as part of this review took place within the context of domestic abuse. Asian victims and suspects were over-represented in domestic abuse cases. • 17% of all rapes were reported in response to the DASH questions asked by police when attending a domestic abuse call. This form of reporting was related to a higher victim withdrawal rate (77%). • In 84% of cases the victim knew the perpetrator and most offences did not involve other forms of violence. • There has been an increase in non-recent reporting with 27% of rapes being reported more than 1 year after the offence took place. • Third parties account for two fifths of all reports and were most commonly support services or family and friends. Key Findings: Timeliness and evidential challenges • On average, case progression from reporting to trial outcome took over 18 months. The length of time between submission to the CPS and decision to charge was 138 days. • There was no relationship between case characteristics and timeliness of case progression. • The lack of forensic evidence was common amongst cases, along with other evidential challenges. There was less evidence present in non-recent cases, and in cases where the victim was over 30 years old. Conclusion The London Rape Review provides a comprehensive picture of reported rape in London and there are clear implications for policy and practice arising from the findings. As the MPS continues to record more incidents of rape than ever before and conviction rates remain low, developing a greater insight into the journey of rape cases through the criminal justice system, as well as identifying key factors affecting case outcomes, is essential. It is hoped the findings from this review will facilitate discourse about how the criminal justice service and its partners can work together to improve experiences and outcomes for the victims/survivors who choose to report their experiences of rape to the police. Key Findings: Types of rape being reported • The strongest predictors of victim withdrawal were procedural characteristics. For example, withdrawal was 6x less likely in cases where the victim/survivor participated in a Video Recorded Interview. • Victim, suspect and offence characteristics were less important in predicting withdrawal. However, being male (3x less likely) and reporting the rape via DASH (3x more likely) were significant predictors. • Procedural characteristics – particularly those related to evidence – were the most important predictors of police NFA. If other evidence cast doubt on the case police NFA was 7x more likely and if there were no forensic opportunities police NFA was 5x more likely. • Cases where the victims/survivors were under 18 years old were significantly less likely to end in police NFA. • Victim/survivor mental health was a significant predictor of police NFA on its own; however, this relationship could be explained by victim/survivor inconsistency in testimony. • Finally, comparing model fit across outcomes showed the case characteristics coded in this review were far better at predicting police decision making than victim decisions to withdraw. Key Findings: Predicting attrition 7 1. Introduction This report outlines the findings from a collaborative research project between the Evidence & Insight Team (E&I) of the Mayor’s Office for Policing And Crime (MOPAC) and the University of West London (UWL). The overall aim was to examine the characteristics of reported rape in London. The focus was on how key factors relating to the offence, those involved, and subsequent responses to it may affect the progression of rape cases through the criminal justice service. The research supports a key ambition in the Police and Crime Plan 2017-2021 (P&CP): to “put victims at the heart of the Criminal Justice Service” (p. 65). i Tackling violence against women and girls (VAWG) is a key priority for the Mayor and the P&CP is clear on the need to better protect those most vulnerable and to tackle high harm crimes. In 2017, Claire Waxman was appointed as London’s first independent Victims’ Commissioner, taking on a key role in making sure victims’ voices are heard and their experiences of services – and of the criminal justice service in particular – are improved. This research contributes to our understanding of victims and their vulnerability, and the findings can be used to drive practical learning and meaningful change. 1.1 Background 1.1.1 The MPS Rape Review 2005-2012 This research updates the methodology used by the Metropolitan Police Service (MPS) Rape Reviews, which examined all rape allegations made to the MPS during the months of April and May from each year between 2005 to 2012.ii Using the MPS Crime Recording Information System (CRIS), each crime report was coded for victim/survivor and suspect characteristics, elements of the offence and, in 2012, elements of the investigation. These rape case reviews proved extremely useful in providing insight into the nature and context of rape and serious sexual assault across the capital. The findings provided both important information on the demographic and situational characteristics of rape offences, and on the trajectories these cases took through the criminal justice process, as well as the associated procedural decisions that occurred throughout investigation and prosecution. Importantly, using inferential analyses, the reviews were able to identify which characteristics had a significant bearing on outcomes. For example, from the 2012 review, important data on the attrition of cases within the sample were identified (e.g. that approximately half of cases ended in victim withdrawal), as were the proportion of cases containing characteristics relevant to operational and investigative challenges in cases of rape (e.g. that 69% of cases involved a delayed report of >24 hours, or that just under 30% of cases involved voluntary consumption of alcohol by the victim). Moreover, accompanying analyses revealed several characteristics acting as significant predictors of case outcome. For example, victim withdrawal was 3x more likely in cases where co-operation or communication problems existed between officer and victim but was 4.5x less likely in cases where the interview was recorded. 8 The 2012 review also highlighted the influence of rape myths1 on the investigative process.iii For example, victims were almost twice as likely to withdraw if they were in an intimate relationship with the suspect, and cases were twice as likely to receive a ‘no further action’ (NFA) classification by officers if the victim had voluntarily consumed alcohol. If the victim had resisted the assault, they were 2x less likely to have their case receive an NFA2 classification. The existence and influence of rape myths on judgements made by the general population is well documented,iv as is their possible impact on police officers’ classification of rape cases, decision making, and case processing.v The previous rape case reviews also provided important information about the trends in rape case progression over time. Figure 1 compares attrition rates across seven previous reviews and reveals a generally consistent trend, with most cases failing to reach the charge stage. There have been some shifts in where attrition occurs in the process. For example, in 2005 and 2006 the greatest dropout was at the allegation to crime stage, whereas in 2010 and 2012 attrition occurred later in case progression when charging a suspect. The 2012 review saw the lowest rate of suspects charged, with just 10% of all allegations made reaching this stage. Figure 1: Attrition rates over time by percentage of all allegations 1 Rape myths are defined as ‘descriptive or prescriptive beliefs about rape (i.e., about its causes, context, consequences, perpetrators, victims, and their interaction) that serve to deny, downplay or justify sexual violence that men commit against women’ 2 Police No Further Action (NFA) refers to when a decision is made by a supervising officer to discontinue an investigation. This can be for a variety of reasons, including victim withdrawal. Police NFA usually occurs when all investigative opportunities have been exhausted and officers feel that the case as it stands does not pass a full code test to pass on to the CPS. Although the investigation is closed at point of NFA, it can be reopened upon discovery or disclosure of other evidence which enables further investigation to take place. 0% 20% 40% 60% 80% 100% Total allegations Crimed Suspect Identified Suspect Arrested Charged 2005 2006 2007 2008 2009 2010 2012 9 1.1.2 Changes and developments since 2012 review Since the 2012 review, several events pertinent to the investigation of rape within London, and the country as a whole, have occurred. These include many high profile and widely publicised sexual abuse scandals and subsequent prosecutions, such as the Jimmy Savile case and Operation Yewtree.vi More recently, cases that have collapsed due to disclosure issuesvii have provided renewed and intensified focus on the procedural, investigatory and prosecutorial decisions taken during the criminal justice process. These events are set within the broader context of the publication of substantial rape investigation reviews, such as the Dame Elish Angiolini Review in 2015, and associated dramatic reforms of police training.viii Alongside - and perhaps triggered by - these events, there has been a rise in recent years of non- recent reporting and an overall increase in reported rape. As Figure 2 shows, for sexual offences recorded by the MPS, there is a clear upward trend both for recorded rape offences and other sexual offences. At the end of 12 months to March 2018, rape offences recorded by the MPS were 140% higher than those recorded in the period to March 2011 (equivalent to over 4,600 more recorded offences). Interestingly, data from the Crime Survey of England and Wales (CSEW) shows little change in the prevalence of rape over the same period. A recent report from MOPAC investigated this disparity (‘Beneath the Numbers’) and concluded that the increases in recorded crime is more likely due to improved crime data integrity and compliance with the National Crime Recording Standards.ix Figure 2: Recorded sexual offences in the MPS year ending March 2011 to March 2018 Because of these developments, both the reporting habits and demographic make-up of rape cases, as well as the associated criminal justice response, may have significantly changed since 0 2000 4000 6000 8000 10000 12000 14000 Mar 2011 Mar 2012 Mar 2013 Mar 2014 Mar 2015 Mar 2016 Mar 2017 Mar 2018 Rape Offs Other Sexual Offences Offs 10 previous in-depth case reviews were conducted. As such, a new research collaboration between E&I and UWL provided the opportunity to conduct an updated review of rape cases reported to the MPS. 1.2 The London Rape Review The focus of this research was on data related to rape allegations made in April 2016. This sample was chosen to ensure comparability with the previous MPS Rape Review data and to enable, as far as possible, any comparisons and observations of changes over time. The decision to focus on 2016 was made because this represented the most recent year where comprehensive outcome information was available to enable a full end-to-end review. The research team worked closely with the independent Victims’ Commissioner for London, MOPAC colleagues, and academic colleagues at UWL, and consulted with a range of external stakeholders (e.g. MPS, CPS, NHS England, Havens, Rape Crisis Centre (RCC) providers, Violence Against Women and Girls (VAWG) service providers) in developing a set of research questions that were reflective of current issues and concerns. As a result, this review provides updated information on the profile of rape cases across London, as well as associated data on the trajectories and outcomes of these cases within the criminal justice service (including analyses on which case characteristics predict case outcomes). It also serves to inform the renewed focus on victims and high harm crimes, vulnerability and safeguarding in the Mayor’s P&CP 2017-2021. Key objectives of the research are: • To contribute to a comprehensive, detailed and up-to-date picture of reported rape in London; • To describe, as far as possible, key characteristics of victims/survivors, perpetrators and offences, as well as subsequent actions and responses; • To better understand the needs and vulnerabilities of those who report rape in London (and how these may impact on outcomes); • To better understand the key characteristics of those perpetrating rape, including offending histories; • To track cases through the criminal justice service and identify key attrition points; • To identify key factors influencing attrition of reported rape in London; and • To make observations on changes over time by comparing findings of the proposed research to those of the MPS Rape Reviews between 2005 and 2012. 11 2. Methodology 2.1 Sample All crime records with an initial classification of rape reported to the MPS in April 2016 were extracted from CRIS for analysis.3 In total, 617 records were extracted. Cases were excluded if: the initial classification was attempted rape; the classification later changed to a non-sexual violence offence; the case was transferred out of the MPS to another police force; or the record was identified as a duplicate of another. Using these criteria, 116 cases were excluded, leaving a final sample of 501. 2.2 Coding and data analysis A team of researchers from E&I, MPS and UWL read each individual CRIS report for the 501 cases and coded them on a total of 146 variables agreed in the coding framework. All researchers were security vetted and followed strict data management and protection protocols. The coding framework was divided into five categories: Victim/survivor characteristics; Suspect characteristics; Offence characteristics (the circumstances of what happened); Procedural characteristics (the police response and investigation); and Outcomes.4 Appendix A presents the variables coded and the descriptive characteristics for the full sample. The coding framework was piloted on 50 randomly selected cases, which were coded by two independent researchers to determine consistency in scoring. From this process, ambiguities over variables were resolved and a detailed and annotated framework was finalised. Throughout the coding process, regular meetings between the researchers addressed any concerns or uncertainties with cases and/or variables. Throughout the results, descriptive statistics are presented on the key variables that were coded5. Relationships between the case characteristics are explored and comparisons are made to previous reviews. Alongside the coding of quantitative variables, researchers took qualitative notes about the reasons given within the CRIS reports for victim withdrawal. Thematic analysis was then conducted on a randomly selected subsample of cases to draw out key themes. To explore whether case characteristics predict case outcomes, a series of logistic regressions were conducted. Each category of case characteristic was examined separately (i.e. victim/survivor, suspect, offence and procedural characteristics) before being combined into an overall model. The two outcomes of interest were victim withdrawal and police no further action, capturing both victim and police decision making. Univariate logistic regressions were first conducted on each case characteristic and the two outcomes.6 Significant univariate predictors 3 Crime records were extracted using Full Business Objects Client. 4 Most of these variables were coded dichotomously (1=yes, 0=no), whereas others were categorical (e.g. ethnicity, offence location, relationship between victim and suspect). 5 Analysis was conducted on the full sample of 501 cases – with the exception of calculations in relation to attrition and case outcome (including prediction of Victim Withdrawal and Police No Further Action), which are based on classified cases only (n=419). 6 Univariate logistic regression explores the relationship between one independent variable (i.e. one case characteristic) and one dependent variable (i.e. one outcome). 12 were retained for inclusion in a multi-variate model for each of the four categories of case characteristics.7 The characteristics that remained significant in each multi-variate model were then retained for inclusion in the final overall model. All analyses were conducted in SPSS version 24. 2.3 Limitations of the methodology Because the coding framework relied solely on the information available on CRIS, there were areas of the journey from offence to trial outcome that were either not captured or were recorded in an inconsistent manner. For example, CPS decisions and the dates on which they occurred were not always detailed and there were variables, such as nationality, which were missing in a substantial number of cases. Furthermore, although the current coding framework captured a significant number of variables, data recorded and held by the CPS, the courts, and victim support services may provide a more complete picture than data available exclusively on CRIS. With regards to the CRIS data, there are further limitations. The system was designed as an investigation record, not a comprehensive data capture tool. Therefore, the quality and detail of the data itself is variable. Although there are specific fields available on the database for many of the variables coded as part of this research, completing them is not always mandatory. Furthermore, most of the information coded is derived from the Details of Investigation: a long, free text log, completed by any number of different police officers and staff throughout the duration of the investigation. There may be details or pieces of information that are of interest which are not recorded during the completion of this log (examples ranging from whether the victim has support of third sector agencies, through to exact dates that case files are submitted to the CPS). Furthermore, many of the variables were coded only for their presence which means we cannot determine whether the absence of a variable is due to omission in data recording or it not being a factor in the case. As such this limits the consistency and reliability of what was coded. Another potential limitation is the sampling of cases. Only rape allegations reported to the MPS during April 2016 were selected for inclusion in this research, to be consistent with previous reviews (all of which used data from April and May in the respective years). These months were previously chosen because they represented ‘average’ months in terms of volume of reported rapes, and thus were thought to be representative of all rape allegations over the year. However, although the volume of allegations is ‘average’ for the year, there may still be sampling bias present (i.e. the allegations for these months being in some way systematically different from cases reported during the rest of the year). A final limitation, which is shared with all research that uses police or official data, is that the cases coded and analysed in this research are only those that were reported to the police. Given the high rate of under-reporting of sexual violence, these cases represent a small proportion of the rapes that may have actually taken place. 7 Multi-variate logistic regression explores the relationship between two or more independent variables and one dependent variable. 13 3. Results 3.1 Attrition of rape allegations A total of 501 allegations of rape were recorded by the MPS in April 2016 and were coded as part of this research. Figure 3 shows the various stages of the investigation, along with the stages of case attrition. Of the 501 allegations of rape made to the MPS, 16% were not classified by the police as a crime. These cases were either recorded as ‘No Crime’ (n=3) or classified as ‘Rape Not Confirmed’ (n=50) or ‘Rape Contradictory Evidence’ (n=29).8 Of the 419 cases that were classified as a crime notifiable to the Home Office, in 67% (n=281) the suspect was identified by the police and in 36% (n=151) the suspect was arrested. In a further 25% (n=103) of cases the suspect, although not formally arrested, was interviewed under caution. The case was submitted to the CPS in 14% (n=60) of cases and in 9% (n=36) the CPS decided to charge (60% of all cases submitted to them). Only 6% (n=23) of all cases proceeded to trial (64% of those charged by the CPS) and only 3% (n=13) resulted in a conviction (61% of those that proceeded to trial). Victim withdrawal was the most common form of attrition for cases in this sample. Over half of victims/survivors (58%, n=243) decided to withdraw their allegation, with the majority withdrawing during the investigation stage (only 8 cases withdrew following CPS submission). Indeed, withdrawal tended to happen early in the process: of those who withdrew, 18% (n=41) did so within 30 days of reporting, 48% (n=111) within 90 days, and 77% (n=179) within 180 days. Qualitative notes taken by the researchers during coding showed that victims/survivors’ 8 The latter two classifications are referred to as N100s. Following a change in Home Office Counting Rules in April 2015, there is now a requirement to create a CRIS report as soon as an incident of rape is logged and to conduct an initial investigation. Rape Not Confirmed (100/1) refers to a case where the victim or third party cannot be identified, and therefore the rape cannot be confirmed. Rape Contradictory Evidence (100/2) is where credible evidence to the contrary exists, indicating that the offence has not taken place. See Appendix B for the Home Office N100 classification codes. Section Summary: Attrition of rape allegations • Victim withdrawal was the most common form of attrition in the sample of classified cases (58%), followed by police no further action (29%). • Out of classified cases, the majority (86%) were not submitted to the CPS. Only 9% were charged by the CPS, 6% proceeded to trial and 3% resulted in a conviction. • Victims/survivors who withdrew did so soon after reporting. Just under half of all withdrawals occurred within 3 months of reporting and 18% within the first 30 days. Reasons for victim withdrawal were complex and often interrelated. 14 reasons for withdrawal were complex and often interrelated. There were typically multiple reasons given for withdrawal, the most common being: • The stress and trauma caused or exacerbated by the investigation, particularly because of having to talk in detail about the incident; • A desire to move on from what had happened, often intensified by feeling surprised and overwhelmed by the process of official police investigation; • Concern for their own safety, or for the perpetrator’s own situation, particularly in cases with a domestic abuse overlap where the victim’s priority often was to put an end to the harmful behaviour, rather than a prosecution; • The act of reporting in and of itself being enough, with reasons for reporting focusing on wanting to get the incident off their chest or seeing reporting as their civic duty (in terms of providing relevant information and intelligence); and • Not having wished to report the rape in the first place, particularly in cases where the report was made within the context of the Domestic Abuse, Stalking and Honour Based Violence (DASH 2009) (see section 3.4.1). Alongside victim withdrawal, the police decided to take no further action in 29% of classified cases (n=123), for reasons such as insufficient evidence, an un-cooperative witness or it not being in the public interest to proceed. In 6% of cases (n=27) the CPS decided to take no further action (45% of all cases submitted to them).9 In 12% of cases (n=62), the investigation determined that no offence had occurred. In 29 cases the victim/survivor stated they had been mistaken and in an additional 19 cases a third party had mistakenly reported a rape. In 14 cases the victim/survivor subsequently stated they had made up the allegation. 9 There were 3 cases which were charged by the CPS but ultimately ended in NFA. 15 Figure 3: Case progression for rape allegations made to the MPS in April 201610 10 Direct comparison between these figures and those of previous rape reviews is problematic due to differences in methodology, as well as changes to crime classification in April 2015. CPS Decision to Charge = 9% (n=36) Proceed to Trial = 6% (n=23) No Crime (1%, n=3), Rape Not Confirmed (10%, n=50), Rape Contradictory Evidence (6%, n=29) Case Classified as a Crime = 84% (n=419) Suspect Identified = 67% (n=281) Sanction Detection Submitted to the CPS = 14% (n=60) Victim Withdrawal = 58% (n=244) Police decision to NFA = 29% (n=123) CPS decision to NFA = 6% (n=27) Suspect Arrested = 36% (n=151) Guilty = 3% (n=14) Not Guilty = 2% (n=8) Allegation made by Victim(s) = 501 16 3.2 Who reports rape in London? 3.2.1 Victim/survivor demographics Overall, the profile of those reporting rape has remained largely consistent when compared to previous years’ reviews. In the 2016 sample, as with previous years, victims/survivors of rape tended to be young, with almost 1 in 3 (n=153) being under 18 years old at the time of the offence and over two-thirds under the age of 30 (n=332; see Table 1 for victim/survivor demographics). Like previous years’ reviews, victims/survivors were predominantly female (89%, n=444) and of white ethnic background (57%, n=465), and black victims/survivors were over- represented compared to the general London population (22% compared to 16% of the London population).11 However, there have been some noticeable shifts in the demographics of those who report, with a small increase in males (from 7% in 2007 to 11% in 2016), a small decrease in black victims/survivors (from 28% in 2010 to 22% in 2016), and an increase in reporting by Asian victims/survivors (from 7% in 2007 to 17% in 2016), with the latter being largely driven by rape allegations made in the context of a domestic abuse investigation (see section 3.4.1). 11 Ethnicity was determined using the six police identity codes (IC). White European (IC1) and Dark European (IC2) were combined to form one category (‘White’). Section Summary: Who reports rape in London? • Victims/survivors were predominantly female and of white ethnic background. Two thirds of victims/survivors were under the age of 30 years at the time of the offence, with almost one third under 18 years. Black victims/survivors were over-represented in the sample. • There has been an increase in male reporting and an increase in Asian victims/survivors compared to previous years’ reviews. • 95% of victims/survivors had at least one need present (e.g. mental health, under 18, learning difficulty). Cumulative needs were common among victim/survivors, with just over one third of victims/survivors having two needs present, and one in five having three or more needs. • 41% of victims/survivors presented with a mental health issue: a significant increase from previous rape reviews. There was a significant relationship between mental health and previous victimisation. 17 Table 1: Demographics of rape victims/survivors 2007 2010 2011 2012 2016 2016 projected London population Gender Female 93% 93% 92% 93% 89% 50% Ethnicity White 66% 60% 58% 57% 57% 57% Black 24% 28% 27% 27% 22% 16% Asian 7% 9% 12% 13% 17% 20% Age at offence Under 18 31% 35% 34% 34% 31% 23% 18 to 29 - 37% 39% 35% 36% 18% 30 to 49 - 25% 23% 28% 27% 32% Over 50 - 4% 5% 3% 6% 27% Note: percentages calculated with missing values excluded 3.2.2 The cumulative needs of victims/survivors Alongside demographic characteristics, the coding attempted to capture other circumstances about the victim/survivor that may add complexity to the case. Some of these needs were measured in previous rape reviews (e.g. mental health), but the 2016 coding attempted to capture more specific detail about the victim/survivor (e.g. whether they were an asylum seeker, missing person, or a sex worker). These needs are presented in Table 2. For some needs, only a small proportion of victims/survivors presented with them (e.g. only 7% of victims/survivors had a learning difficulty and only 2% were an asylum seeker). However, 95% (n=474) of victims/survivors had at least one need present. Over one third of victims/survivors (n=173) had two needs present, and one in five had three or more needs (n=102). It is likely that many of these characteristics are under- reported so the prevalence of these cumulative needs may be underestimated. The most prevalent need amongst the sample was mental health issues (41%, n=205), which was notably higher than observed in previous rape reviews (e.g. 14% in 2007). This increase could be due to better recognition by officers, but could also be due to changes to classifications, namely the introduction of the ‘rape not confirmed’ and ‘rape contradictory evidence’ classifications (with victims noted to have a mental health issue making up 50% of these classifications). 18 The presence of mental health issues was related to prior experiences of victimisation. Victims/survivors with mental health issues were significantly more likely to have been victimised in the preceding 12 months (36% compared to 23% of victims/survivors without mental health issues), and to have been a previous victim of sexual assault or domestic abuse (38% compared to 25% of victim/survivors without mental health issues). Victims/survivors with mental health issues were also significantly more likely to have a prior police record (25%) compared to those without mental health issues (9%). Unfortunately, the coding of mental health in this review precluded us from understanding what types of mental health issues victims/survivors presented with, or whether the issues emerged following the current rape or were pre-existing. Other frequently occurring needs amongst victims/survivors in the sample were being in an intimate relationship with the suspect (35%, n=154), being under the age of 18 years (31%, n=153), and consuming alcohol or drugs prior to the offence (27%, n=123; although this need has reduced in prevalence over time). The high proportion of victims presenting with multiple needs – as well as the overlap between mental health and previous victimisation and offending – not only represents a unique challenge for police investigating these rape allegations but increases the difficulty of the victims/survivors’ journey. Table 2: Needs of rape victims/survivors Need 2007 2010 2011 2012 2016 Mental health issue 14% 14% 17% 18% 41% Intimate relationship with perpetrator 35% 28% 32% 27% 35% Under 18 31% 35% 34% 34% 31% Consumed alcohol/drugs 35% 35% 38% 32% 27% English not first language - - - - 16% Deaf or has physical disability - - - - 10% Learning difficulty 7% Missing person or homeless - - - 4% 7% Sex worker - - - 2% 3% Asylum seeker - - - - 2% Cumulative needs No needs 5% One need 40% Two needs 35% Three or more needs 21% Note: percentages calculated with missing values excluded 19 3.3 Suspect profile The profile of suspects has remained reasonably consistent with previous years’ reviews (see Table 3). Almost all were male (99%) and aged between 18 and 49 (77%). The majority of suspects were either white (43%) or black (35%). There was an over-representation of black suspects (35% compared to 16% of the London population) and an under-representation of white suspects (43% compared to 60% of the London population). There was a significant association between suspect and victim/survivor ethnicity, with suspects typically offending against their own ethnic group.12 The age profile for suspects was slightly older than for victims/survivors, with a smaller proportion of suspects under the age of 18 (12% compared to 31% of victims/survivors). Only a minority of suspects were recorded as having a history of previous offending, including sexual assault or rape. Less than one third (29%, n=143) were recorded as having a prior police record, which represents a decrease from 2012 where 38% were recorded as having a prior police record. Only 11% (n=57) of suspects in 2016 were recorded as having a history of domestic abuse, and 7% (n=34) a history of sexual assault or rape.13 12 These associations were statistically significant for all ethnic groups at the p<.05 level according to the results of chi-square tests. 13 A history of domestic abuse or sexual assault/rape was coded as present if, during intelligence checks by police, the suspect had previously been recorded as a ‘suspect’ or ‘accused’ in an offence relating to sexual assault/rape or domestic abuse, or if they were previously convicted of such an offence and it was shown on their PNC record. Section Summary: Suspect profile • The profile of suspects has remained reasonably consistent with previous years’ reviews. Almost all were male (99%) and aged between 18 and 49 (77%). • Black suspects and those aged between 18 and 49 years were over- represented in the sample, compared to the London population. On the other hand, white suspects, those aged under 18 years and those aged over 50 years were under-represented. • A minority of suspects were recorded a having a history of previous offending, including sexual assault or rape. 20 Table 3: Demographics of suspects 2012 2016 2016 projected London population Gender Male 99% 99% 50% Ethnicity White 35% 43% 57% Black 44% 35% 16% Asian 17% 19% 20% Age at offence Under 18 16% 12% 23% 18 to 29 40% 35% 18% 30 to 49 38% 42% 32% Over 50 6% 12% 27% Note: percentages calculated with missing values excluded 3.4 Types of rape being reported Section Summary: Types of rape being reported • Rape frequently took place within the context of domestic abuse. Asian victims/survivors and suspects were over-represented in cases involving domestic abuse. • 17% of all rapes were reported in response to the DASH questions asked by police when attending a domestic abuse call. This form of reporting was related to a higher victim withdrawal rate. • In the majority of cases the victim/survivor knew the perpetrator and most offences did not involve serious violence. • Compared to previous reviews, there has been an increase in non- recent reporting with 27% of rapes being reported more than 1 year after the offence. • Third parties accounted for two fifths of all reports and were most commonly support services or family and friends. These reports were associated with lower rates of victim withdrawal. 21 3.4.1 Overlap between rape and domestic abuse There was an overlap between rape and domestic abuse: 28% (n=142) of all allegations of rape were related to domestic abuse.14 Victims/survivors in cases involving domestic abuse were more likely to be female compared to victims/survivors in non-domestic abuse cases (94% vs. 87%), less likely to be white (46% vs. 62%) and more likely to be Asian (31% vs. 11%). A smaller proportion of victims/survivors in cases related to domestic abuse had mental health issues (34% vs. 48%). However, a larger proportion of victim/survivors in domestic abuse cases had been previously victimised in the last 12 months (39% vs. 24%) and had been a previous victim of sexual assault or domestic abuse (47% vs. 24%).15 Suspects in cases involving domestic abuse were more likely to be Asian compared to suspects in non-domestic abuse cases (33% vs. 19%). They were also more likely to have a prior police record (49% vs. 21%), and a history of domestic abuse (34% vs. 3%), sexual assault/rape (14% vs. 4%), and other types offending (39% vs. 14%). Consistent with the finding for victims/survivors, suspects in domestic abuse cases were less likely to have mental health issues present than suspects in non-domestic abuse cases (14% vs. 4%).16 There were also some significant differences in characteristics of the offence between domestic and non-domestic abuse cases. In domestic abuse cases, the victim/survivor was significantly less likely to have consumed alcohol prior to the offence (11% vs. 31%) and were significantly more likely to have physically or verbally resisted (52% vs. 35%). The DASH model was implemented across all police services in the UK from March 2009 and requires police to use a common checklist for identifying, assessing and managing risk in domestic abuse cases17. In this sample, 17% (n=86) of all rapes were reported in response to the DASH questions asked by police when attending a domestic abuse incident. Of note, those rapes that were reported via the DASH questions were far more likely to be withdrawn by the victim/survivor, compared to other reporting methods (77% victim withdrawal rate compared to 53% for cases not reported via DASH).18 14 Domestic abuse was coded as present if a Domestic Violence flag had been attached to a CRIS report, or if other information in CRIS made it clear the case was related to domestic abuse. 15 All differences were statistically significant at the p<.05 level according to the results of chi-square tests. 16 All differences were statistically significant at the p<.05 level according to the results of chi-square tests. 17 The Domestic Abuse Stalking and Honour-based violence (DASH 2009) questions are a risk identification, assessment and management model used by police services in the UK. A bank of questions is asked when attending what appears to be any domestic incident to identify, assess and manage ongoing risk. The questions include those asking whether any sexual incidents have taken place as part of abuse. 18 This difference is statistically significant according to the results of a chi-square test: 2 (1) = 14.85, p<.001. Rape DA 22 3.4.2 Rape misconceptions not supported by the data It is often assumed that rape occurs suddenly, at night, by an aggressive stranger with a weapon, and involves visible victim resistance and physical injury.x However, these misconceptions or ‘rape myths’ were not supported by the data. In most cases, the victim/survivor knew the perpetrator. As Table 4 shows, the most common relationship was an intimate partner (35%), followed by an acquaintance or friend (34%). In a further 12% of cases the perpetrator was a family member, and in 2% a professional or carer. In only 7% of cases was the rape perpetrated by a complete stranger. Table 4: Relationship between victim/survivor and suspect Relationship 2012 2016 Stranger 1* 8% 7% Stranger 2 21% 11% Intimate 27% 35% Acquaintance / friend 33% 34% Professional / carer 2% 2% Familial 8% 12% Note: percentages calculated with missing values excluded *Stranger 1 is a complete stranger whereas Stranger 2 is someone the victim met a short time before the rape Consistent with the high level of known perpetrators, almost three in five offences took place in a private or domestic setting: 24% (n=106) at the victim’s home address, 24% (n=108) at the suspect’s home address, and 15% (n=69) at the shared address of the victim/survivor and suspect.19 Further, and beyond the intrinsic violence of the offence itself, most of the offences did not involve serious violence or the use of a weapon. Just 23% (n=113) of victims/survivors were recorded as having an injury (76% of which were recorded as ‘minor’) and in only 2% (n=11) of cases the perpetrator used a weapon. Furthermore, in only two fifths of cases (40%, n=198) was it recorded that the victim/survivor verbally or physically resisted. 3.4.3 Increase in non-recent reporting There has been an increase in non-recent reporting since the 2012 review (see Table 5). In the 2016 sample, same day reporting accounted for only 29% of all allegations made, compared to 40% in 2012. Rapes reported more than one year after the incident increased from 18% in 2012 to 27% in 2016. This rise is likely due to the aftermath of high-profile cases in the media, such as those investigated by Operation Yewtree. 19 Percentages calculated with missing values excluded (offence location was missing in 52 cases). 23 Table 5: Time between offence and report Time between offence and report 2012 2016 Same Day 40% 29% 1 – 7 days 20% 16% 1 week – 1 month 10% 12% 1 month – 1 year 13% 17% More than 1 year 18% 27% Note: percentages calculated with missing values excluded 3.4.4 Third party reporting Overall, two fifths of allegations coded were reported by a third party (42%, n=209). Third parties were most commonly support services, including Havens, Rape Crisis, and counselling services (29%, n=58), family (26%, n=51), friends or colleagues (15%, n=30) or medical staff (11%, n=21). In cases where the rape was reported by a third party, victim withdrawal rates were significantly lower (52% compared to 62% when the victim/survivor reported the crime).20 Where the third party was a family member or friend, withdrawal rates were the lowest (48%). 3.5 Timeliness and evidential challenges 3.5.1 Timeliness of case progression On average, case progression took over 18 months from the date of reporting to trial outcome (see Figure 4). The average length of time between a rape being reported and the police making the decision to take no further action was 213 days (range 0 to 795 days, n=123). The average time from reporting to CPS submission was 345 days (range 1 to 972 days, n=45), indicating that 20 According to the results of a chi-square test: 2 (1) = 4.44, p<.05. Section Summary: Timeliness and evidential challenges • On average, case progression from reporting to trial outcome took over 18 months. The length of time between submission to the CPS and decision to charge was 138 days. • There was no relationship between case characteristics and timeliness of case progression. • A lack of forensic evidence was common amongst cases, along with other evidential challenges. A number of case characteristics predicted the presence of evidence in a case including the length of time from offence to report and victim age. 24 cases submitted to the CPS take over 100 more days, on average, of police investigative time before being ready for submission. The average length of time between CPS submission and decision to charge was 138 days (range 0 to 604 days, n=40): higher than the average length for all rape cases in London in 2016/17 (which was 95 days) and the whole of England and Wales in the same year (67 days).xi There was an average of 133 days (range 23 to 325 days, n=10) between CPS decision to charge and trial, and trials lasted, on average, for 11 days (range 1 to 46 days, n=11)21. Interestingly, there was no relationship between characteristics of a case and how long it took police to decide to take no further action, or to submit the case to the CPS. Figure 4: Timeliness of case progression 3.5.3 Evidential challenges Table 6 shows how common the lack of forensic evidence is, as well as the prevalence of other evidential challenges. In three fifths of cases there were no forensic opportunities22. Not surprisingly, having no forensic opportunities was more likely in non-recent cases (85% of cases reported more than one month after the rape had no forensic opportunities), compared to those reported within 24 hours of the incident (36%). Witnesses were available in only 24% of cases and in only 17% of cases was technological evidence available. An Early Evidence Kit (EEK) was administered in one fifth of cases, the victim/survivor attended a Haven in one fifth of cases and a Video Recorded Interview (VRI)23 was completed with the victim/survivor in just over one third of cases. These proportions do not necessarily reflect a lack of police effort. An EEK is not always possible or appropriate, for example, if the victim/survivor refuses or is unable to provide a sample, is deemed unable to consent to the EEK (e.g. if under the influence of drink or drugs at the time of report, the presence of learning difficulties or mental health issues, or if the victim/survivor is under the age of 16 at the time of the report), or if administering the EEK would delay medical attention or a forensic medical examination. In many 21 Differences may be due to different inclusion parameters being used between the rape review sample and overall official figures. The current study likely had stricter criteria, eliminating any duplicate cases from the sample and only including cases with an initial classification of rape, a current classification of rape or another type of sexual offence. 22 This variable was coded based on the investigating officer explicitly stating within the CRIS report that there were ‘no forensic opportunities’ present. 23 Video Recorded Interviews are an evidential record to be presented in court as an Achieving Best Evidence (ABE) form of interview. Report Police NFA CPS Submission Decision to charge Trial Outcome 213 days 345 days 138 days 133 days 11 days 25 cases where Havens attendance was deemed appropriate, the victim/survivor refused or did not attend (n=54). Similarly, the research team noted many cases where appointments for a VRI were repeatedly scheduled but the victim/survivor chose not to participate. Table 6: Evidential challenges in case progression Evidence All cases n=501 Physical evidence Havens attendance 19% (n=97) Early Evidence Kit administered 21% (n=103) No forensic opportunities 61% (n=305) Victim based Video Recorded Interview completed 34% (n=172) Technological evidence Victim or suspect technology seized 17% (n=85) Other evidence Witnesses 24% (n=118) Social networking sites 13% (n=64) Other evidence casts doubt 15% (n=77) This next section explored whether characteristics about the offence, as well as those involved, were able to predict the presence or absence of evidence in a case. A total ‘presence of evidence scale’ was created by combining the evidential variables in Table 6. Total scores on this scale ranged from -2 to 5.24 Table 7 shows the variables that were significant predictors of the presence of evidence in a case.25 Younger victims/survivors were significantly more likely to have evidence present in their case compared to victims/survivors over 30 years old. In cases where the suspect had a prior police record or where English was not their first language, evidence was significantly more likely to be present. In cases where social networking sites were involved, if the victim/survivor had been drinking prior to the offence, and if the victim/survivor resisted either physically or verbally during the offence, evidence was significantly more likely to be present. Finally, the further the date of reporting was from the incident, the less likely there was to be evidence present in the case. 24 The ‘positive’ evidence (e.g. witnesses) was summed together and the ‘negative’ evidence (e.g. no forensic opportunities) was subtracted from this total to give a final score. 25 Simple linear regressions were conducted between each case characteristic and the presence of evidence, followed by multivariate linear regression. Only variables that were significant in the multivariate model are presented in Table 7. 26 Table 7: Case characteristics predicting presence of evidence Characteristic B p Victim under 18* 0.63 .001 Victim 18 to 29* 0.60 .001 Suspect has prior police record 0.76 <.001 Suspect English not first language 0.76 <.001 Social networking sites 0.53 .007 Victim had been drinking 0.55 .005 Victim resisted 0.66 <.001 Reported 1 week – 1 month after** -0.96 <.001 Reported 1 month – 1 year after** -1.27 <.001 Reported >1 year after** -1.21 <.001 * Compared to victims over 30 ** Compared to same-day reports 3.6 Predicting attrition Section Summary: Predicting attrition • The strongest predictors of victim withdrawal were procedural characteristics. For example, if the victim/survivor participated in a Video Recorded Interview withdrawal was 6x less likely and if an Early Evidence Kit was administered withdrawal was 2x less likely. • Victim, suspect and offence characteristics were less important in predicting withdrawal. However, being male (3x less likely) and reporting the rape via DASH (3x more likely) were significant predictors of withdrawal. • Like victim withdrawal, procedural characteristics – particularly those related to evidence – were the most important predictors of police NFA. If other evidence cast doubt on the case police NFA was 7x more likely and if there were no forensic opportunities police NFA was 5x more likely. • Cases where the victims/survivors were under 18 years old were significantly less likely to end in police NFA. • Victim/survivor mental health was a significant predictor of police NFA on its own; however, when controlling for victim/survivor inconsistent account, it lost significance, suggesting the relationship between victim/survivor mental health and police NFA was due to inconsistency in testimony. 27 This final section looks at whether we can predict attrition using the characteristics that were coded as part of the review. The focus here being on the two main forms of attrition in this sample: victim withdrawal and police decision to take no further action (NFA). It was not possible to conduct predictive analysis on the likelihood of CPS NFA because of the low numbers of cases at this stage (n=60). Predictive analysis in crime classification was not included because the introduction of stringent guidelines and criteria around ‘no criming’ and the introduction of new classifications (e.g. ‘Rape Not Confirmed) removes police discretion from this outcome. Table 8 shows the case characteristics that were significant predictors of victim withdrawal and police NFA. Case characteristics were split into four categories: victim/survivor, suspect, offence, and procedural. The four categories were first examined in separate models before being combined into an overall model.26 This process allowed the relative contribution of each category to be determined. Each case characteristic presented in the table below was a significant predictor in the separate models. The characteristics in bold represent the variables that remained significant in the overall model (i.e. these variables were predictive of outcomes after controlling for the characteristics in the other categories). 3.6.1 What factors predict victim withdrawal? The only victim/survivor characteristic that predicted victim withdrawal was gender: male victims/survivors were significantly less likely (x3) to withdraw compared to female victims/survivors. Although it may seem surprising that no other victim/survivor characteristic predicted withdrawal, the qualitative findings showed that victims/survivors chose to withdraw for many different reasons. One reason was not having intended to report rape in the first place. Indeed, as Table 8 shows, victims/survivors who reported their rape in response to the DASH questions asked when police attended a domestic abuse call were significantly more likely (x3) to withdraw than victims/survivors who reported by other means. In terms of suspect characteristics, in cases where the suspect had a prior police record (x2) or was a family member (x4) the victim/survivor was significantly less likely to withdraw. The presence of witnesses made victim withdrawal significantly less likely (x2), along with cases where the victim/survivor sustained an injury (x2). Several procedural characteristics were significant predictors of victim withdrawal. If an Early Evidence Kit was administered (x2) and if the victim/survivor participated in a video recorded interview (x6), withdrawal was significantly less likely. It often took many attempts for the VRI to be completed and victims/survivors commonly withdrew before it took place, suggesting a potential barrier to case progression. Interestingly, if the suspect denied intercourse or sexual contact had occurred, victim withdrawal was significantly less likely (x6). 26 Univariate logistic regressions were first conducted on each case characteristic and the two outcomes. Significant univariate predictors were retained for inclusion in a multi-variate model for each of the four categories of case characteristics. The characteristics that remained significant in each multi-variate model were then retained for inclusion in the final overall model. All predictive analyses exclude cases that were found to be false or mistaken allegations. When predicting victim withdrawal, cases that were not classified as a crime were excluded. When predicting police no further action, cases that were withdrawn by the victim were also excluded. 28 Table 8: Case characteristics predicting victim withdrawal and police NFA Victim Withdrawal Police NFA Victim/Survivor Characteristics Male x3 less Mental health issues x2 more Under 18 years x3 less Model fit 0.03 0.10 Suspect Characteristics Prior police record x2 less 18 to 29 years x3 less Family member x4 less Family member x3 less Model fit 0.07 0.12 Offence Characteristics DASH x3 more Witnesses x2 less Injury sustained x2 less Offence location private x2 less Witnesses x2 less Model fit 0.10 0.08 Procedural Characteristics Multiple OICs involved x8 more Evidence casts doubt x7 more Early Evidence Kit x2 less Inconsistent victim account x6 more Video recorded interview x6 less Police perception of chance of conviction x5 more Suspect denies rape x6 less No forensic opportunities x5 more Evidence casts doubt x8 less Suspect denies rape x3 less Suspect involved in another rape x8 less Early advice from CPS x10 less Suspect arrested x15 less Model fit 0.43 0.64 Full Model Fit 0.47 0.70 Note: model fit was assessed using Nagelkerke R square which determines how much variation in the outcome is explained by the model. For example, a figure of 0.47 means the model explains 47% of the variance in outcome. Odds ratio were used to determine how much more or less likely an outcome was, given the constant effect of a predictor variable. Having multiple officers in charge (OIC) of the case predicted a higher likelihood (x8) of victim withdrawal. Although these findings could suggest having one OIC in charge of a case may protect against victim withdrawal (perhaps by leading the victim/survivor to feel more supported in the process), having one OIC was very rare in this sample (n=30) and victims/survivors typically liaised with the Sexual Offences Investigative Trained (SOIT) officer rather than the OIC throughout the investigation. One additional finding is shown in the table: if evidence cast doubt on the case, the victim/survivor was significantly less likely to withdraw (x8). The direction of this finding is counter-intuitive given the importance of corroborating evidence already identified, but it could 29 simply be due to the fact this variable led another outcome to become more likely, namely police no further action. The analysis so far has examined each category of case characteristic separately. The last model looked at the four categories simultaneously to determine the relative contribution of each category to predicting victim withdrawal. As can be seen by the characteristics in bold in the table, the strongest predictors of victim withdrawal were the procedural characteristics: all five remained significant in the overall model. Suspect and offence characteristics (aside from reporting via DASH) were no longer significant predictors after taking the procedural characteristics into account, suggesting these variables were less important in predicting victim withdrawal. 3.6.2 What factors predict police no further action? When predicting police decisions to take no further action on a case, two victim/survivor characteristics were of note. First, if the victim/survivor had mental health issues police were significantly more likely to NFA the case (x2). However, the coding of this variable precluded us from understanding what types of mental health issues victims/survivors presented with, or whether the issues emerged following the rape or were pre-existing. The second victim/survivor characteristic that predicted police NFA was age: victims/survivors who were aged under 18 years at the time of the offence were significantly less likely to have their case end in police NFA (x3). Related to victim/survivor age, cases where the suspect was a family member were significantly less likely to be NFA’d by police (x3).27 Cases where the suspect was aged 18 to 29 (compared to over 30) were significantly less likely to be result in police NFA (x3). Offences that occurred in the private dwelling of either the victim/survivor or suspect were significantly less likely to end in police NFA compared to offences that occurred in a public place (x2). Evidential factors were consistent predictors of police decisions to NFA. The first stage of the Full Code test specifies that police must be satisfied there is enough evidence in the case to provide a realistic prospect of conviction. If there were witnesses available (x2) and if the suspect was involved in another rape case (x8), police were significantly less likely to NFA the case. On the other hand, if evidence cast doubt (x7), if the victim/survivor gave an inconsistent account (x6) and if there were no forensic opportunities available (x5) police were significantly more likely to NFA the case. Consistent with the presence of evidence weighing heavily on police decision making, cases where the police perceived the chance of conviction by the courts to be low were significantly more likely to end in police NFA (x5). If police had sought early advice from the CPS, they were significantly less likely to NFA the case (x10), again most likely because police only contact the CPS about cases they think have a chance of being successfully prosecuted. Finally, if the suspect had been arrested, the case was significantly less likely to end in police NFA (x15). 27 More victims under 18 were offended against by a family member (26%) compared to older victims (2%). This difference was statistically significant according to a chi-square test: 2 (1) = 67.48, p<.001. 30 When all four categories of case characteristics were included in the same model, victim/survivor mental health was no longer a significant predictor of police NFA. Victim/survivor mental health was significantly related to the victim/survivor giving an inconsistent account.28 This variable remained significant in the overall model, suggesting that the relationship between victim/survivor mental health and police NFA was due to inconsistency in victim/survivor accounts. Like victim withdrawal, almost all procedural characteristics remained significant in the overall model, suggesting these variables were most important to predicting police NFA compared to characteristics about the suspect or offence. However, controlling for procedural characteristics, victim/survivor age remained significant in the overall model. This finding suggests that regardless of what procedural characteristics were present (many of which related to the presence or absence of evidence), victims/survivors under 18 years were significantly less likely to have their case end in NFA. Finally, comparing model fit across outcomes showed the case characteristics coded as part of this review were far better at predicting police decision making than victim decisions to withdraw. 28 35% of victims with mental health issues gave an inconsistent account compared to 16% of victims without mental health issues. This difference was statistically significant according to a chi-square test: 2 (1) = 19.55, p<.001. 31 4. Discussion The London Rape Review used an innovative and in-depth approach to analysing data on reported rape in London. Through the coding of over 500 crime reports across 146 variables, the resulting data set provides the most up to date and comprehensive picture of reported rape in London, allowing for the tracking of cases within the criminal justice service, and analysis of factors influencing case outcomes. The research serves to inform the renewed focus on victims and high harm crimes, vulnerability and safeguarding in the Mayor of London’s Police & Crime Plan 2017-2021. 4.1 Summary of key findings Consistent with previous reviews, and indeed numerous other studies of recorded rape, the London Rape Review found steep attrition of cases following initial reporting: only 14% of all cases were submitted to the CPS for a charging decision, 9% were charged by the CPS, 6% proceeded to trial, and just 3% were convicted. These figures represent a continuation of trends in attrition patterns observed in previous reviews, with the 2016 sample recording the lowest proportion of cases charged and convicted. Victim withdrawal was the most common form of attrition in the sample. Over half of all rape allegations (58%) were withdrawn by the victim/survivor. Those who withdrew usually did so soon after reporting, and all but eight during the police investigation stage. Alongside withdrawal, the police decided to take no further action in a further 29% of cases. Despite the increase in recording over recent years, including a rise in non-recent offences, the profiles for those who report rape in London – and those suspected of perpetrating rape – have remained broadly consistent across time. Victims tend to be female and young (with a substantial proportion below the age of 18 at the time of the offence), whilst suspects are overwhelmingly male and with an age profile that tends to be slightly older than victims. For both victims and suspects, those with a black ethnic background are over-represented compared to their overall proportions in the London population, although the proportion of black victim/survivors has decreased slightly over time. Interestingly, compared to previous years, the 2016 dataset saw an increase in Asian victims and a slight increase in male victims. The data also showed a high prevalence of cumulative needs amongst victims/survivors, with 95% presenting with one or more needs and over half presenting with at least two needs. The most common needs were mental health, being in an intimate relationship with the perpetrator, and being a child or teenager at the time of the offence. Other needs captured as part of this review were less prevalent amongst the sample; for example, being homeless, an asylum seeker or a sex worker. This low prevalence may point to the particularly high rates of underreporting (or indeed non-reporting) by such vulnerable groups. Nevertheless, the high proportion of cumulative needs adds to the complexity of cases, both in terms of carrying out the police investigation and attending to victims/survivors’ support needs. Two in five allegations were by a victim/survivor noted to have a mental health issue: a considerable increase from previous rape reviews. Victims/survivors with mental health issues 32 were significantly more likely to have been previously victimised and to have a prior police record. These findings point to the significant demand placed on first responders and investigating officers around mental health and the implications this has on the training and skills required to progress the investigation in a trauma-informed way. Although sample sizes precluded us from drawing any definitive conclusions, some additional analyses (not reported in the results) showed that where victims/survivors with mental health issues received support (from an RCC, ISVA or Haven), they were more likely to have their cases proceed than where they did not receive support. There is certainly merit in further research aimed at evidencing the impact of support; however, there are related considerations around the extent to which current provisions in London are equipped to consistently and expediently make support available to victims/survivors following reporting.xii One of the most important aspects of this review was being able to identify which case characteristics influence different attrition outcomes. Procedural characteristics emerged as the most important predictors of both victim withdrawal and police NFA. Victims/survivors who participated in a Video Recorded Interview (VRI) and who were administered an Early Evidence Kit (EEK) had significantly lower rates of withdrawal. Within the case files, participation in a VRI emerged as a significant barrier to case progression, with victims/survivors frequently expressing concern or apprehension about the process. This finding raises questions around how each victim can be supported to give the best evidence possible, despite existing complexities and challenges. There was a substantial overlap between rape and domestic abuse. Reporting via the DASH was a significant predictor of victim withdrawal, with many noting they never intended to report the rape in the first place. At the same time, perpetrators in offences with a domestic abuse link showed a higher prevalence of previous offending. These findings suggest a separate rape investigation may not always be reflective of victims/survivors wishes; however, it may be the perpetrators in these cases that represent the greatest risk of continued harm. Victim/survivor input and autonomy may be essential in these cases to decide whether a separate investigation is the most appropriate course of action, or a focus primarily on safeguarding and harm reduction. Relatedly, victims/survivors frequently mentioned their reason for reporting was related to safety and the discontinuation of risk (i.e. wanting to ensure this does not happen to someone else), rather than seeking a prosecution. Again, this finding suggests that for some rape cases, a criminal justice resolution is not the desired outcome for the victim/survivor. Although some victims/survivors may require additional support to help them continue the criminal justice journey, both Stern (2010) and Angiolini (2015) suggest the exploration of ‘alternative outcomes’ or ‘other measures of success’.xiii Where the victim does not withdraw, procedural characteristics – particularly those related to the presence of evidence – emerged as key predictors of police decisions to take no further action. Police NFA was significantly more likely if there were no forensic opportunities available in a case and if other evidence cast doubt on the case. These findings are unsurprising given the first stage of the Full Code test specifies that police must be satisfied there is enough evidence in the case to provide a realistic prospect of conviction. Indeed, those cases where police sought early advice from the CPS were significantly less likely to end in NFA, suggesting police know what makes a ‘good case’. 33 Angiolini (2015) acknowledged in her review ‘the evidential opportunities presented by electronic and digital communications and social media’. More recent developments and cases involving disclosures have raised considerable concerns about how intrusive this may be for the victim/survivor and whether evidence gained in this way is always used in a proportionate and appropriate way.xiv In the cases reviewed, social media was involved in 13% of cases and victim or suspect technology was seized in 17%. These variables were not predictive in the multi-variate models; however, both were significantly related to police NFA on their own. As these findings are a reflection of a sample from 2016, analysis on more recent data is needed in order to contribute robust evidence to the current debate around the use of victim disclosure and technology evidence, and the impact of this on victims and on case outcomes. Compared to previous reviews, victim, suspect and offence characteristics were less influential in predicting attrition. The 2012 review found evidence for some ‘rape myths’ impacting on police decision making; for example, victims who physically resisted the attack were significantly less like to have their case end in police NFA.xv However, the 2016 review found these kinds of ‘real rape’ stereotypes did not have a significant bearing on outcomes, suggesting there may be some shifts in how investigating officers are influenced in their decision-making. Yet, consistent with the 2012 review, we found some evidence that being a ‘credible victim’ was important for case progression. Cases where there were inconsistencies in victims/survivors’ accounts were significantly more likely to be given an NFA decision. However, research shows that victim accounts of traumatic events are likely to be inconsistent.xvi Therefore, there is a risk that by filtering out these cases, some victims/survivors could be denied access to justice. Finally, comparing model fit across outcomes showed the case characteristics coded in this review were far better at predicting police decision making than victim decisions to withdraw. The overall model predicting police NFA explained 70% of the variance in outcomes, compared to victim withdrawal which explained 47%. Police case files can only tell the story from the perspective of the police officers involved in the case, so it is likely we are missing out on important information about victims/survivors’ experiences of the investigative process which may prove to be stronger predictors of their decision to withdraw. 4.2 Future Research Although the findings contained in this report provide a comprehensive overview of reported rape in London, there are many avenues for future research using this, and other, data sources. Initial follow-up to the current London Rape Review is underway, which will include more in-depth qualitative analysis around victim withdrawal reasons to better understand the key themes and factors driving attrition at this point. Key partners such as the CPS are being engaged to gather more detail around outcomes, police-CPS contact and correspondence, and decision-making, to build a more complete picture. Finally, although this review represents the most up-to-date picture of reported rape in London, there have been additional developments over the past three years which makes repeating this review with a more recent sample of cases a vital next step. 34 4.3 Conclusion The MPS is receiving more allegations of rape than ever before and conviction rates are at their lowest since these in-depth rape reviews started in 2005. Within this context, victims/survivors are presenting with more and more complex needs, and more stringent guidance requires police to record and thoroughly investigate all complaints of rape. Developing a greater insight into the journey of rape cases through the criminal justice system, as well as identifying key factors affecting case outcomes and attrition, is essential for shaping policy and practice. It is hoped the findings from this review will facilitate discourse about how the criminal justice service and its partners can work together to improve experiences and outcomes for the victims/survivors who choose to report their experiences of rape to the police. 35 Appendices Appendix A: Case characteristics included in the coding framework Variable N % of sample Outcome Classified as a crime 419 84% Victim withdrawal (excl. no crime) 244 58% Police ‘no further action’ (excl. no crime and withdrawals) 123 29% Submitted to CPS (excl. no crime and withdrawals) 60 14% Victim/survivor characteristics Age at time of offence Under 18 years 153 31% 18-29 years 179 36% 30-49 years 134 27% Over 50 years 30 6% Age at time of reporting Under 18 years 113 23% 18-29 years 181 36% 30-49 years 156 31% Over 50 years 47 9% Gender Female 444 89% Male 55 11% Ethnicity White 266 57% Black 101 22% Asian 78 17% Other 20 4% English not first language 80 16% Requires interpreter 44 9% Asylum seeker 9 2% Recognised as vulnerable or intimidated 295 59% Deaf or has physical disability 49 10% Has mental health issue 205 41% Has learning difficulty 31 6% Is a sex worker 14 3% Is a missing person or homeless 37 7% Has links to gangs 3 1% No recourse to public funds 4 1% Previously made false allegation sexual assault 32 6% Previously made false allegation other 19 4% Has PNC ID 78 16% Known as a suspect 85 17% Previous victimisation last 12 months 141 28% Previous victimisation of sexual assault/DA 151 30% History of consensual sex with perpetrator 131 26% Suspect characteristics Age at time of offence Under 18 years 45 12% 18-29 years 126 35% 30-49 years 152 42% Over 50 years 42 12% Ethnicity White 153 43% Black 127 35% 36 Asian 68 19% Other 12 3% Gender Female 7 2% Male 461 99% Deceased 7 1% English not first language 62 12% Requires interpreter 6 1% Asylum seeker 1 0% Disability 9 2% Learning difficulty 5 1% Has mental health issue 33 7% Has a PNC ID 143 29% Links to gangs 5 1% Previous history of domestic abuse 57 11% Previous history of sexual assault/rape 34 7% Previous history of other offending 106 21% Relationship to victim/survivor Stranger 78 16% Intimate Partner 154 31% Friend or Acquaintance 150 30% Familial 52 10% Professional or Carer 9 2% Offence characteristics Location Victim’s dwelling 106 24% Suspect’s dwelling 108 24% Shared dwelling of victim and suspect 69 15% Not domestic 166 37% Drug facilitated 22 4% Domestic violence related (DV Flag) 142 28% DV primary issue (DASH) 86 17% Linked to other offence 160 32% Witnesses 118 24% Social networking sites 64 13% Offence recorded or photographed 20 4% Injury sustained 113 23% Weapon used 11 2% Victim/survivor had been drinking 126 25% Victim/survivor had taken drugs 33 7% Victim/survivor believes they were drugged 36 7% Victim/survivor verbally resisted 184 37% Victim/survivor physically resisted 71 14% Perpetrator had been drinking 94 19% Perpetrator had taken drugs 30 6% Procedural characteristics Time taken to report Same day 138 29% 1 – 7 days 77 16% 1 week – 1 month 58 12% 1 month – 1 year 80 17% More than 1 year 131 27% Reported by Victim/survivor 292 58% Third party 209 42% Number of SOITs involved in case One 269 55% 37 Multiple 84 17% None 132 27% Number of OICs One 36 7% Multiple 465 93% Victim/survivor attended Havens 97 19% Video recorded interview 172 34% Police reference delays due to workload 112 22% Victim/survivor is difficult to contact 167 33% Early investigative advice sought from CPS 25 5% Victim/survivor unsure if offence took place 53 11% Victim/survivor unsure where offence took place 59 12% Victim/survivor lack of understanding of consent 40 8% Victim/survivor gives inconsistent account 127 25% Victim/survivor identifies perpetrator 369 74% Victim/survivor gives detailed description of perpetrator 256 51% Early Evidence Kit administered 103 21% No forensic opportunities 305 61% Body Worn Video footage 12 2% Request for third party material made 179 36% Possible technological evidence referred to 133 27% Victim/survivor technology seized by police 54 11% Suspect technology seized by police 66 13% Technological evidence Supports victim's case 12 7% Supports suspect's case 21 12% Supports neither case 84 47% Not stated 63 35% Perpetrator claims consent 106 21% Perpetrator denies intercourse or sexual contact 65 13% Other evidence casts doubt 77 15% Perpetrator involved in another rape case 23 5% Police perceptions chance of conviction 61 12% Police perceptions victim credibility/reliability 100 20% OIC expresses doubt about case 78 16% Victim/survivor receives ISVA support 67 13% Victim/survivor receives support through RCC 44 9% Note: Percentages calculated with missing valued excluded 38 Appendix B: N100 classification codesxvii N100 classification codes Following a change in Home Office Counting Rules in April 2015, there is now a requirement to create a CRIS report as soon as an incident of rape is logged. Every CAD opened or closed as a rape will require a CRIS report. Unless already recorded as a crime, every report of a rape, suspected or possible rape must be taken at face value and recorded, irrespective of the source of the report. This includes reports made by third parties, through partnership working or from other agencies. The reported incident of rape will be in addition to any other record, e.g. CAD N100 codes are designed to be used in the first stages of an allegation. Once the crime has been investigated by Sapphire a 'no crime' application must be made in the usual way (see How do I change reports to 'no crime' in the Supervision and Review Toolkit). Categories 100/1 – Victim or in some cases a *Third party (see below) has not confirmed the crime and/or cannot be traced. Additionally, this code can be used where a crime is not confirmed but further investigation is required (old CRI) 100/2 – Credible evidence to the contrary exists 100/3 – Transferred or committed in another force area N100/1 – Cannot be confirmed Following a satisfactory initial investigation where every effort has been made to identify the victim or the third party without success, a 100/1 report is recorded. The minimum requirement to finalise these entries would be: That a CID officer and the local Duty Officer have been informed & they are satisfied that no crime has been committed. The CID officer and their own DI are to be named on the dets of the report. The local Duty officer should also make a formal entry on the dets of the report to confirm that he or she is satisfied with the classification. *IT IS IMPORTANT TO NOTE THAT IF THE THIRD PARTY IS A PARENT, A PROFESSIONAL OR A CARER N100/1 SHOULD NOT BE USED AND THE REPORT MUST BE RECORDED AS A CRIME IRRESPECTIVE OF WHETHER CONFIRMATION IS RECEIVED – see 'Third Party Information' below. 39 N100/2 – Evidence to the contrary First responders must conduct a thorough initial investigation, supporting victims and, in line with the Golden Hour principle, secure and preserve evidence whilst arresting offenders at the earliest opportunity. Following the initial investigation, if credible evidence has been obtained indicating that the offence has not taken place, then the N100/2 classification should be used. Perceived or even diagnosed Mental Ill Health or a propensity to make false claims are not credible evidence that a rape has not occurred. Officers are reminded of the impact of trauma on victims and the effect this may have on the consistency of their account. A victim might not always be willing to repeat an allegation or engage with the police from the outset of an investigation. The credible evidence must be thoroughly documented on the CRIS, e.g. if CCTV is present and allegedly contradicts the allegation, then it must be viewed by the officer. The Toolkit and associated Policy must be adhered to when a CRIS report is completed for rape; The matter must be referred to a CID officer & the Duty Officer must be informed of the circumstances. Each of their details must be listed in the body of the DETS. A formal entry must be made by the Duty Officer once the evidence gathered by the initial investigator has been reviewed. N100/3 – Transferred to another force This classification is reserved for records of rape allegations where the offence has occurred outside the Met. The CRIS reports will be processed by the Transfer Crime Team (TP CRIB) as before, but will be 'classified' as N100/3 as opposed to confirmed crimes. Third Party Information In April 2015, the Home Office made amendments to the Crime recording requirements for reports made by third parties. It is beyond the scope of this document to detail these fully here. The complete detail can be obtained from the relevant editions of the National Crime Recording Standard (NCRS) and Home Office Counting Rules (HOCR) by reading Counting rules general rules. (Navigate, firstly, to NCRS Paragraph 3.6(ii) and, secondly, to Section A Whether and When to Record (3 of 7) and scroll down to the box 'Reports of Rape via 'Third Parties'....') The brief extract below indicates that many crimes reported by parents, carers and professional third parties must be recorded as crime irrespective of whether the victim provides confirmation: 'ii. Parents, Carers and Professional 'Third' Party Reports Crimes are often reported by individuals acting on behalf of victims. These may be referred to as 'Third Party' reports and commonly such reports include the following: a) Persons acting in a professional capacity e.g. doctors, nurses, social workers and teachers reporting crimes, (often of a safeguarding nature), on behalf of victims of any age. 40 b) Parents or Carers acting as a guardian or responsible adult, ('in loco parentis'), reporting crime in the best interests of and/or to ensure that a child or young person has appropriate access to police services. When such persons report crimes, they should always be regarded as acting on behalf of a victim. Where there is no doubt as to their status and/or position or the veracity of their report, those reports must be recorded as crimes. Such recording must occur regardless of whether the victim has given their permission for the reporting individual to speak to the police and irrespective of whether the victim subsequently confirms that a crime has been committed. Other 'Third Party' reports from persons acting on behalf of victims should be treated on their individual merit and in line with guidance at paragraph 2.2 and 3.6 i within the Standard.' 41 References i MOPAC (2017). A Safer City for All Londoners: Police and Crime Plan 2017-2021. Available from: https://www.london.gov.uk/sites/default/files/mopac_police_and_crime_plan_2017-2021.pdf ii MPS (2005). A review of rape investigations in the MPS; MPS (2007). The attrition of rape allegations in London: A review; Hohl, K. and Stanko, E. (2015) Complaints of rape and the criminal justice system: fresh evidence on the attrition problem in England and Wales. European Journal of Criminology, 12(3): 324-341. iii Bohner, G. (1998). Vergewaltigungsmythem [rape myths]. Landau, Germany: Verlag Empirische Pädagogik. iv Grubb, A., & Harrower, J. (2008). Attribution of blame in cases of rape: An analysis of participant gender, type of rape and perceived similarity to the victim. Aggression and Violent Behavior, 13, 396–405. Grubb, A., & Harrower, J. (2009). Understanding attribution of blame in cases of rape: An analysis of participant gender, type of rape and perceived similarity to the victim. Journal of Sexual Aggression, 15, 63–81. Grubb, A., & Turner, E. (2012). Attribution of blame in rape cases: A review of the impact of rape myth acceptance, gender role conformity and substance use on victim blaming. Aggression and Violent Behavior, 17, 443–452. v Hine, B., & Murphy, A. (2017). The impact of victim-perpetrator relationship, reputation and initial point of resistance on officers' responsibility and authenticity ratings towards hypothetical rape cases. Journal of Criminal Justice, 49, 1-13. Hine, B., & Murphy, A. (2019). The influence of ‘High’ vs. ‘Low’ rape myth acceptance on police officers' judgements of victim and perpetrator responsibility, and rape authenticity. Journal of Criminal Justice, 60, 100- 107. Murphy, A., & Hine, B. (2019). Investigating the demographic and attitudinal predictors of rape myth acceptance in UK Police officers: developing an evidence-base for training and professional development. Psychology, Crime & Law, 25(1), 69-89. For review, see Sleath, E., & Bull, R. (2017). Police perceptions of rape victims and the impact on case decision making: A systematic review. Aggression and Violent Behavior, 34, 102–112 vi https://www.telegraph.co.uk/news/uknews/crime/11393099/Operation-Yewtree-The-successes-and- failures.html vii https://www.theguardian.com/uk-news/2018/jan/30/met-police-and-cps-apologise-to-man-after-collapse- of-case viii Angiolini, E. (2015). Report of the independent review into the investigation and prosecution of rape in London. Retrieved from: https://www.cps.gov.uk/publication/report-independent-review-investigation-and-prosecution- rape-london-rt-hon-dame-elish ix MOPAC Evidence and Insight (2019). Beneath the Numbers: An exploration of the increases of recorded Domestic Abuse and Sexual Offences. Retrieved from: https://www.london.gov.uk/sites/default/files/201902_beneath_the_numbers_- _an_exploration_of_the_increases_of_recorded_domestic_abuse_and_sexual_offences_v1.pdf x Hine, B. and Murphy, A. (2019). The influence of ‘High’ vs. ‘Low’ rape myth acceptance on police officers’ judgements of victim and perpetrator responsibility, and rape authenticity. Journal of Criminal Justice, 60, 100- 107. xi CPS (2017). Violence Against Women and Girls Data. Retrieved from: https://www.cps.gov.uk/violence-against- women-and-girls-data xii https://www.independent.co.uk/news/uk/politics/sexual-assault-rape-victims-counselling-waiting-list- support-violence-women-girls-appg-report-a8679456.html xiii Stern, V. (2010). Stern review of rape reporting in England and Wales. London: Home Office. Angiolini, E. (2015). Report of the independent review into the investigation and prosecution of rape in London. Retrieved from: https://www.cps.gov.uk/publication/report-independent-review-investigation-and-prosecution- rape-london-rt-hon-dame-elish xiv https://www.theguardian.com/society/2018/sep/25/revealed-uk-police-demanding-access-data-potential- rape-victims xv Hohl, K. and Stanko, E. (2015) Complaints of rape and the criminal justice system: fresh evidence on the attrition problem in England and Wales. European Journal of Criminology, 12(3): 324-341. xvi Hopper, J. (2018). How reliable are the memories of sexual assault victims? Scientific American. Retrieved from: https://blogs.scientificamerican.com/observations/how-reliable-are-the-memories-of-sexual-assault-victims/ xvii MPS intranet (2019). First responder – Rape and serious sexual assault policy. 1 The London Survivors Gateway Pilot: A 2-year evaluation Final Report October 2020 Rachael Parker, Charlotte Keeble & Daniela Wunsch MOPAC Evidence and Insight 2 Contents Acknowledgments ................................................................................................................................... 3 Executive Summary ................................................................................................................................. 4 Introduction ............................................................................................................................................ 7 Methodology ......................................................................................................................................... 13 Results: Learning from year 2 of the London Survivors Gateway Pilot ................................................ 14 The Role of the Gateway Navigator ...................................................................................................... 17 Process Learning from Year 2 of London Survivors Gateway ............................................................... 20 Results: Learning from year two of the Complex Needs ISVA service .................................................. 26 The Experience of Complex Needs ISVAs during Year 2 ....................................................................... 29 Conclusion ............................................................................................................................................. 34 References ............................................................................................................................................ 37 Appendices ............................................................................................................................................ 38 3 Acknowledgments MOPAC Evidence & Insight would like to thank Laura Duckworth for her research support, Dr Melissa Pepper for her work on the Year 1 evaluation of the pilot and Sophie Geoghegan-Fittall for her work during the Year 2 evaluation. MOPAC would also like to thank Dr Paul Dawson for his research advice and Tara Poore for her support and assistance in undertaking the evaluation. Many thanks also go to Lucy Hayton and colleagues from Women and Girls Network (WGN) for their support and assistance. Finally, MOPAC would like to thank the practitioners, stakeholders and service users who gave up their time to share their views and perspectives with the evaluation team. 4 Executive Summary The London Survivors Gateway addresses sexual violence by simplifying access routes into services, ensuring consistency of support and improving outcomes for victim-survivors. The project brings together the Mayor’s Office for Policing And Crime (MOPAC), NHS England (NHSE) and commissioned services: Women and Girls Network (WGN) - West London Rape Crisis and lead operational partner; NIA (East London Rape Crisis), Solace Women’s Aid (North London Rape Crisis), RASASC (South London Rape Crisis), GALOP (LGBT and anti-violence charity) and Survivors UK (male rape and sexual violence charity). The project is further supported by the Havens1 and Kings College Hospital NHS Foundation Trust. In addition to an online and telephone central point of access for sexual violence services across London (the London Survivors Gateway), the project also funded six complex needs Independent Sexual Violence Advisors (ISVAs) to work with survivors who require specialised support, and undertook a London-wide mapping of sexual violence services. The initial pilot, funded through the Home Office Transformation Fund together with contributions from MOPAC and NHSE, ran from June 2018 until March 20202. Evidence and Insight (E&I) - MOPAC’s in-house social research and analytical team - were commissioned to undertake an evaluation of the pilot of the London Survivors Gateway. The 18-month evaluation examined two distinct areas: monitoring the performance of the service through the routine capture of core project metrics; and generating in-depth understanding of the processes - from design through implementation of the service. This final evaluation report presents the second year of performance data and process learning; firstly, exploring findings from the Gateway, followed by findings from the ISVA service. Key Findings Project throughput and activity of the Gateway • Over the duration of the project (between October 2018 and March 2020) a total of 9,843 attempted calls were made to the Gateway phone line and a total of 2,988 case clients were registered on the Gateway system, the majority of which were referred online. • In year 2 of the pilot (April 2019 to March 2020), 6,534 attempted calls were made to the London Survivors Gateway. The majority of calls were made during the Gateway’s opening hours (87%, n=5,660), of which 57% (n=3,208) were successfully connected to a Navigator. • The proportion of calls successfully connected to a Navigator increased during year 2 (i.e., 57% versus 47% in year one) which is positive, although there remains room for further improvement. • Also in the second year, 2,352 survivors became registered as ‘case clients’ on the Gateway system; the majority were online referrals to the London Survivors Gateway website (91%, n=2,146), and almost half of all referrals were from the police (47%, n=1,103). • The majority of case clients were female (93%, n=2,189) and most were aged between 18-34 (62%, n=1,467). There were more referrals for White case clients (44%, n=1,043) compared to clients of Black, Asian and Minority Ethnic backgrounds (BAME) (Black: 16%; Asian: 10%, Other: 11%); these proportions are in line with those seen in the overall London population. • For Year 2 case clients, a total of 2,456 outcomes were recorded (including referrals into services; client declining referral; or no service available), and 172 clients were recorded as having ‘entered’ the 1 The Havens is network of 24/7 specialist sexual assault referral centres (SARCs) located across London for people who have been raped or sexually assaulted. 2 Following a ‘soft’ launch in June 2018 to coincide with the start date of some of the complex needs ISVA work, the Gateway central point of access went live in October 2018. 5 Gateway more than once (therefore more than one outcome was recorded). Two in five outcomes were referrals to other services (n=1086, 44%) predominantly within the Gateway partnership (including into waiting lists). For one in ten (10%, n=261) there was no suitable service available. Process findings from the Gateway • Consistent with findings from Year 1, there was perceived to be a high demand to the Gateway service, with Navigators now ‘holding’ large numbers of cases to gradually ‘funnel’ case clients into the partnership services. • Finding ways to meet and manage this demand was an ongoing challenge for the service and various adjustments were made throughout the life of the pilot. This included the recruitment of additional staff, the introduction of the Gateway coordinator role, a revision to the Gateway Navigators’ shift patterns as well as a temporary hold on awareness raising. These led to some improvements (i.e., more calls answered in year 2), although the issue was not fully resolved. The project would likely have benefitted from a more systematic approach at the beginning, in setting out expectations around the volume and demand to the service, with clear decision-making processes on designing the resource required to meet the expected demand. • The Gateway has been well-received by both service users and referring agencies, and feedback suggests that, following first use, referrers continue to refer and recommend the service to colleagues. • Service users gave particularly positive feedback on their interactions with Navigators, although they reported that subsequent waiting lists into services can have a detrimental impact on their well-being and recovery. • Positively, some of the challenges in getting different partners to work together - seen during the set up in Year 1 - have since been resolved. Partners acknowledge that relationships need to be continuously nurtured at all levels, from senior management through to practitioners. The performance data also indicated that Navigators are predominantly referring into agencies within the Gateway partnership, which also points to positive relationships. • The relationship between practitioners and the police (who were the largest referrers) was brought to the fore in this research. Both Navigators and ISVAs highlighted having to challenge police at times on the use of language with sexual violence survivors and the use of rape myth terminology. There were also some problems highlighted by both police and Navigators around the referral criteria to the Gateway and managing referrals for survivors who are also experiencing domestic violence. Project throughput and activity for the complex needs ISVA service • Over the duration of the funding period for the additional complex needs ISVAs (mid-August 2018- March 2020), a total of 293 individual survivors were supported. • In year two of the pilot, ISVAs supported 273 clients and of these, 63% (n=171) were new clients and 37% (n=102) were existing clients to the service (who were referred in Year 1 of the pilot). Most of these clients had self-referred (35%, n=95) or had been referred via police (32%, n=87); these were also the most common referral routes reported in Year 1. • Over half of clients who engaged with the ISVA services were female (55%, n=149) and 32% (n=88) male; these proportions are similar to Year 1. • The most common age group among ISVA clients in year 1 was 25-34 years (25%, n=69), with the second largest age category split between 18-24 (22%, n=60). For clients where ethnicity was known (n=237), there was a very similar split between non-BAME (n=121, 51%) and BAME clients (n=116, 49%). Compared to Year 1 there is a higher proportion of BAME survivors supported in Year 2. • Seventy per cent of ISVA clients (n=141/201) had experienced multiple incidents of sexual violence (a higher proportion than the 61% in Year 1), perpetrated by either the same person over a period of time and/or by difference perpetrators at different times, and the majority required ongoing ISVA support (93%, n=254), as opposed to a single incident of support. 6 • Most clients reported experiencing at least one form of disadvantage, disempowerment or discrimination. Twenty-nine percent (n=78) had experienced multiple forms of disadvantage, disempowerment, or discrimination. • In terms of self-reported client outcomes, the biggest improvements were seen in areas of ‘more able to access further support’, where 70% (n=152) reported improvements, and ‘more able to assert their rights’ (with 75%, n=161 reporting improvements). Process findings from the complex needs ISVAs • During year 2, ISVAs reported perceiving an increasing range of complex needs in clients, particularly in supporting clients with legal issues, including civil and family court; assisting in victim’s right to review (VRR) and issues relating to immigration, asylum and no recourse to public funds. Additionally, ISVAs mentioned an increase in clients presenting with mental health needs. • These increasing types of needs were reflected in the areas of further need for training the ISVAs identified. This included better understanding of family court issues and supporting clients with complex mental health issues. • ISVAs reported facing many challenges in their role, some of which were similar to those reported by Navigators. They also highlighted concerns for supporting clients who have experienced delays with the criminal justice system – an issue that had been exacerbated by the Covid-19 pandemic. Conclusions • The Gateway has been very well received - by service users, referring agencies and by stakeholders. Overall, it is seen to have made access easier for survivors and referral pathways more straightforward for referrers. The role and the work of the Navigators have been particularly well received. • Managing demand has been the Gateway’s biggest challenge and over the course of the pilot various changes to how the Gateway has been run have been implemented in order to respond to this. However, demand challenges continued and it is likely that balancing awareness-raising with readying for an increase in demand will be an ongoing task for the Gateway to manage. • In addition to managing incoming demand, the Gateway also faced a challenge in managing onward demand. Referrals into support services often were into waiting lists, whilst for one in ten no suitable service was available. This raises a wider potential imbalance between demand and availability of support services – with staff highlighting the need for more finances to address this. This situation impacted on the role of the Navigator, which shifted from the initial vision of a triage role to conducting lengthy assessments over the phone with survivors, providing first-hand emotional support and ‘holding’ cases to attempt to funnel onward referrals into support services – further clarity on this role would be beneficial. • The Gateway has played, and continues to play, an important role in bringing partners together across London. This is an important aspect to further build on to continue to facilitate information-sharing and knowledge exchange. Keeping on top of an often-changing landscape of service availability (due to fluctuations in services closing and opening their waiting lists) will require ongoing monitoring and regular communication between partners - to ensure the online mapping tool remains accurate and up to date. It is therefore recommended that the mapping becomes a regular exercise and that resource is allocated to maintain it. • The findings highlighted that there is scope to improve the mutual understanding and partnership working, especially between support services and criminal justice agencies, such as the police. This could take the form of awareness-raising and knowledge exchange, but also joint training, and would need to cover use of language with sexual violence survivors and rape myth terminology, but also an understanding of the different support roles, such as ISVAs. 7 Introduction Background Tackling sexual violence and providing support to ensure better outcomes for victim-survivors are priority areas in the London Mayor’s Police and Crime Plan 2017-2021 and the Violence Against Women and Girls (VAWG) Strategy 2018-2021. To illustrate, the Crime Survey for England and Wales (CSEW)3 estimated 3.1 per cent of women (510,000) and 0.8 per cent of men (138,000) aged 16 to 59 experienced sexual assault in year ending March 2017. Using the Mid- Year 2018 population estimate for London4 and the London level CSEW results5, we can estimate that approximately 151,000 adults aged 16 to 59 years who live in London experienced any form of sexual assault abuse in the year prior to completing the survey. Women were more likely to have experienced sexual assault than men (4.1% compared with 1.3%). This equates to an estimated 114,000 women and 37,000 men. In the year to March 2017, the MPS recorded 17,608 sexual assault offence, an increase of 10 per cent from the previous year. A number of studies have highlighted the broad and complex needs that victim-survivors of sexual abuse often present, which require support across a range of specialist services, including physical and mental health, substance misuse, disability, insecure housing, unemployment, gang association, and experiences of early life sexual abuse6. Furthermore, sexual violence is often embedded within ongoing gender-based violence - domestic abuse, forced marriage and trafficking – with crimes often experienced on a repeat basis and perpetrated by people known to the victim-survivor, including intimate partners7. Sexual violence is also a significantly underreported crime, often linked to feelings of shame, blame, or fear, with additional barriers for Black Asian or Minority Ethnic (BAME), disabled, or Lesbian, Gay, Bisexual, and Transgender (LGBT) victim-survivors, or those involved in sex work8. This points to the importance of multi-faceted service provision and synergistic relationships between partners that acknowledge the often-complex nature of sexual violence and can deliver specialist support that shifts to the needs of the victim-survivor9. The tailored practical and emotional support offered by Independent Sexual Violence Advisors (ISVAs) is a critical part in helping victim-survivors to cope, recover, and move forward with their life, including (where appropriate) their journey through the Criminal Justice System (CJS)10. Indeed, in a review of rape 3 The CSEW is the preferred measure of trends in the prevalence of sexual assault as it is a self-report survey of people living in households in England and Wales and therefore unaffected by changes in police activity, recording practices, and propensity of victim-survivors to report crimes to the police. 4https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/dataset s/censusoutputareaestimatesinthelondonregionofengland 5 Crime Survey for England and Wales, Office for National Statistics: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/adhocs/008805crimesurveyenglandan dwalesestimatesofsexualassaultanddomesticabuseexperiencedbyadultsaged16to59 6 Lea et al., 2015; Smith et al., 2015; NatCen, 2015; Campbell, 2007; Ullman, 2016; Allen et al., 2004. 7 Lovett and Kelly, 2009; Krug et al., 2002. 8 ONS, 2018; Ceelen et al, 2016; Zinzow and Thompson, 2011; James and Lee, 2015; Ullman and Townsend, 2007; Sigurvinsdottir and Ullman, 2015. 9 Astbury, 2006; Hester and Lilley, 2018. 10 Robinson et al., 2009; Home Office, 2017. 8 complaint handling in England and Wales, one author commented on the significant effect an ISVA can have on the development of criminal cases by the police and prosecution, and successful delivery of justice by statutory bodies11, while another reported improved professional standards of agencies that victim-survivors interacted with when an ISVA is involved.12 The London Sexual Violence Needs Assessment A 2016 Sexual Violence Needs Assessment of London, commissioned by MOPAC and NHS England, highlighted the importance of specialist support for those who have experienced sexual violence, including ISVAs and services that meet specific needs of LGBT victim-survivors, those with learning difficulties, and BAME women and men. These services were recognised as vital in terms of improving outcomes for victim-survivors and were found to impact positively on attrition rates in the CJS. However, the review also pointed to limited and uneven access to ISVA support across London, and the funding pressures, demanding caseloads, and long waiting lists that many services were facing – particularly smaller community-based organisations, those that target specific equalities groups, preventive work, and health and wellbeing services. Many victim-survivors (62% of those who responded to a survey as part of the needs assessment) do not seek immediate support, with some waiting a number of years before they do so, often related to fear of not being believed, or concern about attitudes of those who they report to – particularly the police. Furthermore, victim- survivors often have very limited awareness of services available and may struggle to access support in their local area13. Following on from the Sexual Violence Needs Assessment, MOPAC worked with NHS England (London) and MOPAC’s commissioned/funded services to develop a new sexual violence model that would address the reports’ findings, ensuring a consistent offer of support and improving outcomes for victim-survivors through high quality care from the point of entry to exit from services. It was this work that led to the development of the London Survivors Gateway. The London Survivors Gateway Pilot The London Survivors Gateway sought to address sexual violence by simplifying access routes into services and increasing support for the complex needs that victim-survivors often present. Based on learning from the Essex based Synergy Model, the overarching aim of the project is to ensure consistency of support offer to all victim-survivors, with improved outcomes for victim- survivors from the point of entry to their exit from services. The London Survivors Gateway pilot - funded through the Home Office VAWG Transformation Fund, together with contributions from MOPAC and NHS England (NHSE) - was launched in June 2018, with the initial pilot funding period lasting until March 202014. 11 Stern (2010). 12 Robinson et al. (2009). 13 MBARC, 2016. 14 Following a ‘soft’ launch in June 2018 to coincide with the start date of some of the complex needs ISVA work, the London Gateway central point of access went live in October 2018. 9 The pilot brought together MOPAC, NHSE and commissioned sexual violence support services: Women and Girls Network (WGN) - West London Rape Crisis and lead operational partner in this pilot; NIA (East London Rape Crisis), Solace Women’s Aid (North London Rape Crisis), RASASC (South London Rape Crisis), GALOP (LGBT and anti-violence charity) and Survivors UK (male rape and sexual violence charity). Further support came from the Havens (a network of 24/7 specialist sexual assault referral centres (SARCs) located across London for people who have been raped or sexually assaulted) and Kings College Hospital NHS Foundation Trust. The Gateway sought not to replace existing working practices, but rather to enhance service provision by building partnerships across the London sexual violence landscape and establishing consistency for all victim-survivors. The Evidence and Insight (E&I) Unit - MOPAC’s in-house social research and analytical team - were commissioned to undertake an evaluation of the London Survivors Gateway. The two-year evaluation of the initial pilot period (June 2018 – March 2020), examines two distinct areas: monitoring the performance of the service through the routine capture of core project metrics; and, generating in-depth understanding of the processes - from design through implementation of the service15. This final evaluation report focuses on the second year of performance data and process learning, outlining where there have been key changes and developments compared to year one. A summary of the findings of year one of the evaluation - which explored the initial set-up and implementation of the London Survivors Gateway and additional complex needs ISVAs - can be found in Appendix A. Key Elements of the London Survivors Gateway The London Survivors Gateway Pilot comprises three key features: a central point of access for sexual violence services across London (the London Survivors Gateway or ‘the Gateway’ for short), six additional complex needs ISVAs to work with survivors who require multiple levels of intervention and specialised support, and a London wide mapping exercise of sexual violence. The London Survivors Gateway Based on learning from the Essex Synergy Model - The Gateway is a universal access and referral portal for London, providing support to survivors via a pan-London telephone line (open Monday to Friday, 10am to 4pm, with calls made outside of these times directed to the 24/7 Havens telephone line). The Gateway is managed by Women and Girls Network (WGN; West London Rape Crisis) and is staffed by several Navigators and who are employed by WGN to provide a first contact response to victim-survivors that helps them to make informed decisions about their care and access to support services. 15 The feasibility of an impact evaluation was explored in the initial stages of the evaluation design. However, it was not possible to identify a suitable control group, or comparable baseline data (due to the Gateway being a new pan- London service uplift; and different data collection methods previously being used across the different support services). 10 Gateway Navigators offer person-centred, trauma-informed information and support to victim- survivors, conduct an initial needs assessment (covering areas related to informed choice, forensic medical examinations, and safeguarding), make direct referrals to the most relevant and appropriate sexual violence service in London and, where appropriate, assist survivors in accessing those services. Referrals to the Gateway service come directly from the victim-survivor or via other agencies that they are in contact with (e.g., the police). The Gateway telephone line also receives calls directed from the Havens within opening hours. The phone line is supported by a website (see Appendix B) that offers victim-survivors and referrers a call back within two days from a navigator after submitting a short online referral form. The anticipated outcomes of the Gateway were that: • Survivors find it easy to access information and advice on sexual violence services in London (via the Gateway service); • Survivors who access the Gateway feel believed and understood; and • Those accessing the Gateway have an increased awareness of sexual violence services in London. Appendix C sets out the Gateway process. Updates to Gateway processes and resources over the course of the pilot Throughout the lifetime of the pilot, staffing levels have varied and several changes were made to some of the internal processes at the Gateway. The pilot first went live with three Navigator posts in place - reflecting the amount of funding that had been secured at the time16. Concerns were raised by partners that this would likely not be enough to manage demand. Indeed, two additional Navigators were recruited during Year 1; one of whom took on a Gateway Co-ordinator role in November 2019 (but due to capacity and recruitment issues still managed a full caseload until March 2020). From January 2020 there were only two Navigators and one Co-ordinator due to staff moving onto different roles within WGN. This was the case until March 2020, when three additional Navigators were appointed in March. Furthermore, it is understood that WGN are continuing to recruit more staff for the Gateway and at the time of writing are expecting to welcome a further three members of staff (with one covering a year’s maternity post and to replace another member of staff). Therefore, capacity has increased to six Navigators overall. In addition to the recruitment of more staff, the Gateway have made several changes to their ways of working to manage demand and capacity. This included the introduction of a shift system 16 The Gateway pilot was awarded £1m of Home Office funding, which was increased to a total of £1.36 million through MOPAC and NHS match funding, though remained lower than the original funding bid for £1.6 million (which included a MOPAC Contribution). 11 to allow some navigators to focus on answering calls while others carry out other tasks associated with the role including following up call backs after initial referral, making enquiries on the survivors’ behalf, onward referrals, and addressing safeguarding concerns. Whilst outside of the evaluation period, it is understood that from March 2020 further changes to the Gateway processes and working structures were being made. This included allocating specific days to Navigators for conducting the different aspects of their roles (e.g. making first contact with new referrals, conducting previously booked needs assessments, and making follow up calls), but also capping the number of weekly new referrals per navigator to 10. Alongside the most recent uplift in staff numbers, these changes were considered to be sufficient to meet the current demand, provided the levels of referrals remained the same. Complex Needs ISVAs In addition to the Gateway itself, six complex needs ISVAs are located across the partner agencies (WGN, NIA, Solace, RASAC, GALOP and Survivors UK). The ISVAs work with victim-survivors (across the service – not just those via the Gateway) who present with more complex needs and require tailored, intensive support. The ISVAs also deliver awareness raising activities to ensure the service is accessible to a range of survivors in diverse communities. The anticipated outcomes of the ISVAs are a reduction in the impact of sexual violence for survivors with complex needs, including: • Survivors feel more in control of their lives; • Survivors have better health and well-being, and an increase in positive coping strategies; • Survivors feel more able to access further support; • Survivors feel more able to develop and maintain positive relationships with those who matter to them; and • Survivors feel more able to assert their rights. Pan-London Mapping Exercise The final aspect of the pilot is a London-wide mapping exercise focused on identifying gaps in sexual violence service provision in London to inform future work and commissioning decisions. During the pilot, an online map was produced by WGN which depicts several sexual violence services in London. The map is located on the Gateway website17 which can be accessed by agencies and the general public. Additionally, in Summer 2020, a service dashboard for the partnership agencies was developed on the website. This is accessed through a ‘partner log in’ tab on the website and provides a space for agencies to regularly update details of their own services (including referral criteria and waiting list information). The aim of this is for the Gateway and other agencies to have up to date information on service provision in the partnership. 17 https://survivorsgateway.london/service-map/ 12 The anticipated outcome of the mapping exercise is that MOPAC and agencies within the partnership have an improved awareness of available services in London and the need and demand. 13 Methodology This final evaluation report outlines learning from Year 2 and builds on interim findings from the first-year report (see Appendix A), which focused on the initial learning generated during year one of the London Survivors Gateway (1st July 2018 to 31st March 2019). Year 2 learning is based on: • Performance data capturing the activity of the London Survivors Gateway Pilot - including the work of the Complex Needs ISVAs - between 1st April 2019 and 31st March 2020 (provided by WGN). • Telephone and online survey data capturing service user feedback on the Gateway (conducted and provided by WGN). Respondents were those survivors who were referred and received a service from the Navigators at the Gateway. There were 151 respondents to the online survey and 251 respondents to the telephone survey; the two groups of respondents likely overlap with each other.18 • Fieldwork was conducted by E&I between February and June 2020. This includes: o An online survey of individuals from both statutory and charity agencies who had made a referral into the London Gateway during year 2 (the survey was sent to a list of 350 referrers, provided WGN; and received a 16% response rate (n=56 responses were received); o A virtual focus group with the six complex needs ISVAs in April 202019; o Seven semi-structured face to face practitioner interviews (including Gateway navigators and Haven staff)20; o Ten semi-structured interviews (face-to-face and telephone) with key stakeholders involved in the set up and delivery of the Gateway pilot; and o Three semi-structured telephone interviews with London Gateway service users. The report will first explore the performance and process learning from the London Survivors Gateway, followed by the performance and process learning of the complex needs ISVA service. 18 It is not known by how much the two groups overlap. The samples do not include survivors who may have attempted to reach the Gateway but weren’t connected, or survivors who were referred but unable to be reached. 19 This was conducted during the Covid-19 lockdown but focussed on the ISVAs experiences of the role prior to the lockdown. 20 One of the Navigator interviews was designed to draw out the various scenarios and draw upon some anonymous case studies from their role. Whilst the interviewee drew on some real-life scenarios in their answers, no personally identifiable information about clients were disclosed. 14 Results: Learning from year 2 of the London Survivors Gateway Pilot This section will present findings from year 2 of the London Survivors Gateway, including performance analysis of the Gateway throughput, activities, and demographics of service users. It will also include a detailed overview of the Gateway Navigator role, followed by process learning from interviews with stakeholders, practitioners, referrers and service users. Gateway Performance Analysis This section presents year two performance data in relation to the London Survivors Gateway (1st April 2019 to 31st March 2020) and comprises of three sub-sections: 1) an overview of referrals into and calls made to the London Gateway; 2) an overview of service user demographics; and 3) an outline of onward referrals and outcomes. Referrals and calls into the London Survivors Gateway • Over the duration of the entire project (between October 2018 and March 2020) a total of 9,843 attempted calls were made to the Gateway. • In year 2 (between 1st April 2019 and 31st March 2020), 6,534 attempted calls were made to the Gateway21. This is consistent with first year findings where 3,309 attempted calls were made in a 6-month period, suggesting attempted call volumes have remained stable. • Most of the attempted calls were made during Gateway opening hours (87%, n=5,660), of which only 57% (n=3,208/5,660) were successfully connected to a navigator - an average of 9 calls per day. This is an increase from the proportion of calls that were successfully connected during the first year of service (47%) – which may reflect some of the changes made to staffing levels and processes that were outlined previously. However, even with the positive increase in year 2, there remains further room for improvement in successfully connecting to callers. • The remainder of attempted calls made during opening hours were missed because the caller hung up at the busy message (n=2,309) or because the call was abandoned by the caller (n=143). 21 This included all incoming calls to the Gateway from survivors and professionals. Current data recording practices mean it is not possible to disaggregate this data to identify callers (i.e., survivor or professional). The number of calls does not necessarily equate to the number of individuals calling (i.e., one individual may call multiple times). 6534 Attempted calls to the Gateway 1592 onward referrals to support services 2352 Case clients 2146 online referrals 15 • For calls that were made to the Gateway during closed hours (n=874), 778 callers hung up at the closed message, 78 were successfully transferred to the Havens, whilst for 18 the transfer to the Havens was unsuccessful. • There is no data available to track callers or indicate whether callers who were unsuccessful tried calling the Gateway again. There was some feedback from survivors that the Gateway telephone line automatically hanging up when busy, rather than putting callers on hold, was perceived as a ‘significant barrier’. Case client demographics • ‘Case clients’ for the Gateway are defined as those who have been referred to the service (either via self-referral or third party) and whose contact details are registered on the system from the referral form. This does not mean the case client automatically receives the full service from the Gateway (e.g., in cases where the client becomes uncontactable following referral). A non-case client is someone who comes in via the Havens phoneline, or who asks a query to the Gateway, but is not registering for the Gateway service. • Over the duration of the entire project (between October 2018 and March 2020) a total of 2,988 case clients were registered on the system. • During year two of the pilot, 2,352 individuals were registered as ‘case clients’ on the Gateway system - an average of 196 per month. This is an increase on the number of case clients in Year 1 where there was an average of 127 per month.22 Consistent with findings from the first-year report, most case clients were referred online via the London Survivors Gateway website23 (n=2146), with a further 183 being telephone referrals and 23 from email referrals. • Consistent with previous findings, almost half of all referrals came from the police (47%, n=1103/2352), followed by the survivor themselves (28% n=670), another agency (such as Education, Social Services, Health services; 22% n=515) or a third party (2%, n=36). • The majority (93%, n=2,189) of case clients were female, with the next largest category male (n=139). Thirteen clients identified as non-binary, seven as Transgender Female, and four as Transgender Male. The majority of case clients were aged between 18-34 (62%, n=1467). 168 case clients were aged 13-17 years. • There were slightly more referrals for White case clients (44%, n=1043) compared to clients of Black Minority Ethnic backgrounds (BAME) (Black: 16%; Asian: 10%, Other: 11%). The remainder were unknown/preferred not to say. Comparing this to the ethnicity 22 It was not possible to break down data month by month. 23 Gateway staff anticipated from the outset that the majority of case clients would derive from online referrals due to professionals opting for this referral route, and the anonymity it provided to survivors for self-referral. 16 of the London population, it shows that the proportion of White and Black survivors are in line with the overall population, whilst Asian survivors are underrepresented among the case client group (the proportion of White residents in London is 45%, proportion of Black residents is 13%, and proportion of Asian residents is 18%)24. Overall there is very little difference in demographics compared to case clients seen in the first year of service. (See Appendix E for full breakdown of demographics of case clients). • Just over half of case clients sought support from the Gateway in relation to rape (56%, n=1326), followed by other sexual violence (25%, n=595), and child sexual abuse (13%, n=298).25 • Almost seven out of ten survivors knew the perpetrator (n=1616, 69%), most commonly as an acquaintance (23%, n=545) or a current or former intimate partner (22%, n=516). In 338 (14%) cases, the perpetrator was a stranger26. The high proportion of cases in which the perpetrator was known to the survivor has been reflected in other studies, including the recent MOPAC London Rape Review in which the perpetrator was known to the survivor in 84% of rape cases reviewed (MOPAC, 2019). Case client outcomes • When a survivor ‘enters’ the Gateway through either the online referral form, or via telephone, an outcome of that interaction will be recorded on the system. Each outcome recorded represents a different time that a survivor enters the Gateway service; for example, a survivor may enter the service for the first time, but no service is available to refer them onto, but then they may enter the service a second time on another date and successfully be referred to another service. • Across all year 2 case clients a total of 2,456 outcomes was recorded. More than one interaction outcome was recorded for 172 clients, suggesting the majority entered the Gateway just once. • Of these outcomes, 44% (n=1086) were referrals to other services - predominantly within the Gateway partnership (n=965), though it is worth mentioning that this includes referrals onto waiting lists.27 For a quarter of registered case clients (26%, n=630), the Gateway was not able to make contact following the initial referral.28 For one in ten outcomes (10%) no suitable services were available. 24 2011 Census. 25 Data on survivor needs was not available. 26 This included cases of stranger whom victim-survivor had no prior contact with or where there was only a brief interaction between victim-survivor and perpetrator (defined by the police as ‘Stranger 1’, n=195) and cases where the victim-survivor and perpetrator are briefly known to one another (defined by the police as ‘Stranger 2’, n=124). 27 There is no data on outcomes from the service that the survivor was referred onto by the Gateway. 28 This means that when the referral has been made, and the Gateway attempt to contact the survivor to initiate contact and an initial needs assessment; the survivor is uncontactable. This could be for a number of unknown reasons, and the survivor may have changed their mind. 17 Table 1. Types of case client outcomes Case client outcome types Number of outcomes Percentage Referred to service in partnership 965 39.3% Unable to make contact 630 25.7% No services available 261 10.6% Declined service29 236 9.6% Inappropriate referrals30 172 7.0% Referred to other sexual violence service 86 3.5% Signposted to service in partnership 42 1.7% Referred to non-sexual violence service 35 1.4% Signposted to other sexual violence service 14 0.6% Signposted to non-SV service 13 0.5% Awaiting outcome 1 0.04% Transferred for FME 1 0.04% Total 2456 Types of referrals into other services • Each successful referral by the Gateway into another service is recorded. There are occasions where survivors are referred into one service but engage with more than one ‘activity’ within that service (for example a client may receive both a counselling service and an ISVA service from the same agency). • Data shows that there were 1592 total onward referral activities in Year 2 which includes multiple activities for some clients. Most referrals were to one of the pilot partnership organisations including West London Rape Crisis (21%, n=328), North London Rape Crisis (n=292, 18%), South London Rape Crisis (15%, n=234), East London Rape Crisis (12%, n=185), Havens (7%, n=115), Survivors UK (5%, n=81) and GALOP (3%, n=42). The remainder were made to organisations outside of the partnership (e.g., Ashiana, Gaia, Asian Women’s Resource Centre, Respond) (n=315). The Role of the Gateway Navigator The purpose of the Navigator role at the Gateway is to provide a pan-London, initial response to survivors of sexual violence. The role was initially conceived as a way to triage survivors out to sexual violence services across the capital. However, over the course of the pilot, the Gateway 29 A survivor may decline a service for a number of reasons. For example they may not have been aware of the referral (if they were referred by a third party), or they may have had poor experiences with services before and are reluctant to engage again, or they may have changed their minds/their situation could have changed.. 30 An inappropriate referral is usually one where there are immediate safeguarding risks, for example the survivor may be homeless, or they may be at risk of domestic violence. The Gateway will advise the referring agency that the safeguarding need should be met first, before ongoing work can begin on supporting them with trauma from sexual violence experience. 18 service has developed, resulting in changes in the role of the Navigator, which will be discussed in the next section. The Navigator Approach The Navigators described several key aspects to their overall approach when working with survivors (see figure below). This approach was seen as a direct response to the high level of need they often encountered31. Indeed, over the course of the pilot the Navigator role had extended beyond the initial triage role into one that provided its own level of support – something that had not been anticipated or designed at the outset of the pilot. The offer of individualised support is central to the role and Navigators described their approach as holistic and as recognising every aspect of the survivor’s life. This includes Navigators creating a bespoke ‘package’ for each client based on the needs assessment they conduct (see Appendix D for Needs Assessment scenarios): “we will figure out what their support needs will be, we will look at what we can give in terms of resources, we will connect that client with helplines, both practical and emotional support…”. To support high-risk clients or during circumstances where survivors cannot immediately receive support from other services in the partnership (i.e., due to waiting lists, or closed services), Navigators report that they often advocate for the client to their GP and request that they provide support to them and continue to conduct regular welfare checks with survivors in the interim. These additional responsibilities were considered time consuming but seen to be demonstrating the caring nature of the role. Navigators also described the therapeutic and psychoeducational approach they take, often involving a form of emotional support and acknowledging a survivor’s response to trauma. Their initial interaction with each survivor was seen to be unfolding in stages, ending the 31 Needs data was not available, however anecdotally the Navigators reported there to be highly vulnerable survivors. 19 communication with a strength-based approach which focuses on empowering the survivor through recognising their courage and tenacity. In this way, Navigators considered the Gateway as a forum for survivors to be heard: “To come through, to be acknowledged, to be met, to be recognised, not just in your victimhood, but also to be celebrated in your survivor hood”. Feedback on the Navigators’ approach from survivors was very positive. Almost all of the survivors surveyed (99%, n=148/150 from the online survey; and 94% (n=233/249 from the telephone survey)32 said they felt ‘believed and understood’, with comments that the Navigators were experienced as ‘warm’ and ‘caring’. Whilst this suggests that the Gateway has achieved one of its key anticipated outcomes (that of survivors feeling believed and understood), it should be noted that the feedback gathered through the surveys does not include the views of those who have been able to make full contact with the Gateway. When asked about improvement suggestions, two out of five surveyed survivors (n=54/135) thought there could be no improvements. The most popular suggestions were ‘longer opening hours’ (n=35), ‘provide a webchat’ (n=28), and ‘more ongoing support available’ (n=26). Feedback from more in-depth telephone interviews with survivors was similarly positive, although for some interviewees it was difficult to differentiate between the Gateway and other services they had received support from (e.g., services they were referred into, or a previous service they had received). After the Navigators have conducted their initial work with survivors over the phone, and depending on the outcome of the needs assessment, the Navigators will then decide which service(s) are most appropriate for each survivor to be referred onto. The Navigators discussed their lack of continuity with survivors as a challenge to their job, and that it may be “nourishing to get more of a whole sense of a person”. They highlighted that their role meant: “we come at crisis and we leave at crisis…” and suggested that following a survivor through the entirety of their journey to recovery could improve job satisfaction and create more sustainability in the role. This issue was reflected in the stakeholder interviews where it was acknowledged that vicarious trauma is a risk for the Navigators, as they are “dealing with the worst moments of a survivor’s life as part of their day to day work”. Overall, the Navigator position was seen to be a specialist role reliant on skills, expertise and previous experience. Both Navigators who were interviewed had pre-existing work histories and expertise within the sexual violence sector, and as a result had joined the Gateway with established partner relationships and the confidence to directly converse with survivors to give appropriate support and decipher often complex and varied information. Despite their wealth of experience, the Navigators suggested that they would still benefit from mental health training to 32 Both surveys were anonymous. It is possible that there is some overlap between those who have completed both surveys. “I've done a lot of mental health assessments before, and they've all felt very clinical. Some have been better at being patient when I'm stressed and struggling to answer, some less patient. But my phone conversation was different from both of these. I felt more understood than I did at any other mental health service. I was even offered new perspectives that reassured me quite a lot.” - Survivor 20 help obtain all the necessary information from survivors in a way that is most appropriate and useful. Process Learning from Year 2 of the London Survivors Gateway This section considers the process learning generated through the year 2 evaluation fieldwork, structured under four thematic headings: Development of the London Gateway pilot; Resource and Capacity; Nurturing Successful Partnerships; Success for the future. This section utilises findings from interviews, focus groups, and surveys conducted with key stakeholders across the partnership, the Navigators, survivors and referring agencies. Development of the London Gateway Pilot during Year 2 Stakeholders, when asked about any changes to the pilot model within the second year, mostly agreed that there had been no significant changes to the core design but acknowledged that the pilot is ‘maturing’ and becoming more ‘embedded’ within sexual violence referral pathways.33 It was noted that there had been a particular increase in the number of self-referrals (this is reflected by the performance data which showed an increase of self-referrals from 11% to 28% in the year April 2019-March 2020, compared to the first 6 months). Despite some changes being made in Year 2 (the uplift in staff and a shift system for Navigators), the nature of the Gateway telephone service – in particular the assessment time (reported to take on average around 40 minutes) to provide a high quality, bespoke service to each survivor – and the multiple contacts often required for each referral, means that it is likely there will always be times when lines are engaged. As touched upon in the previous section, the Navigator role in particular has developed over time, and it was noted by interviewees that the role had become quite different to its original description. At the core of this is a shift from triage or ‘helpline’ to ‘holding’ cases and providing more in-depth support over the phone - which was formally recognised by amending the Navigator job specification. The expansion of the Navigator role in this way was felt to be a way to manage the demand for support services for sexual violence survivors, resulting in Navigators acting as a ‘funnel’ into these services: “it makes it simpler for services because we can hold things and process them and give a stopping point for the clients so it doesn’t flow out as much, but we are also giving out more clients than people have had before”. Resource and capacity 33 This feedback was taken before changes were made to the Navigators’ processes in March 2020. “When it started it was understood like a triage service[…] then it was understood as a helpline, but it’s not that either, it’s like a weird temporary crisis case management where sometimes you have people who are currently unsafe and sometimes you have people calling for the Havens who are in danger, it’s a very multi-faceted role and there is a lot to it.” 21 Evidently, the pilot and the awareness of it developed fast and “became a success very quickly at the front end”, which has been demonstrated by the large number of calls and referrals to the service, and overall positive feedback from service users and referrers. Again, despite the various amendments and uplift in staffing that were made to accommodate the demand, there was still strong concern expressed from stakeholders that further resource is needed to enable the Gateway to respond to it. Partners also reported that the uncertainty around future funding adds further complication in their abilities to plan for new resource. The issue of resource and capacity was seen as a key challenge for the Navigators who report to be managing large caseloads (it was anecdotally reported by a Navigator that at the time that they were each managing around 60- 70 cases at the Gateway), including many cases with complex needs.34 It was remarked in stakeholder interviews that the Gateway has been a ‘victim of its own success’, as it quickly became known and attracted referrals, however with its current level of resource it is difficult to meet the levels of referrals. The Gateway appears to be putting a spotlight on the level of need and issues around service provision for sexual violence survivors in London. A considerable amount of work went into awareness raising and training in the initial stages of the programme, including sessions with referring agencies35. Stakeholders acknowledged that further awareness raising needs to and should be carried out to reach wider communities and agencies but noted that this needs to be balanced against the resource available. At present, the resource and capacity available has meant that no further awareness raising has been carried out for the Gateway. Year 2 findings suggest that resource and capacity challenges - highlighted in the first year evaluation report and prior to that in the London Sexual Violence Needs Assessment – are an ongoing issue, particularly when it comes to specialist service provision for BAME, male, LGBT, and disabled survivors and those for whom English is not a first language. Due to capacity of services, even when survivors are referred into another service, for the most part it is into the service’s waiting lists. Service user feedback was clear that despite receiving a positive response from the Gateway, the subsequent waiting lists for services can have a damaging impact on their wellbeing, including their recovery taking far longer and a deterioration of mental health. Nurturing Successful Partnerships Relationships within the Partnership 34 This is anecdotal and data on survivor needs was not available. 35 Police, housing sector, local authorities, NHS, youth workers, psychologists, sexual health workers, student union representatives and other voluntary organisations. “Well the waiting was excruciating because I was desperate because I was having to go to work and explain myself and me being a fighter I did go to work, so I obviously felt like anyone else would the sooner the better when you’ve had a trauma like that, but obviously the resources aren’t brilliant, so that’s no finger pointing” - Survivor interview “if awareness increases, there needs to be something on the other side, otherwise it’s an unhelpful thing to call if you call in a desperate state and there is nothing to be referred to.” 22 As highlighted in the first evaluation report, partnership working underpins the Gateway pilot. The pilot has allowed agencies to build and develop strong connections with each other, enabling them to capitalise on shared knowledge, maximise resources and begin the building of a joined- up London service. When implementing a new multi-agency programme there are naturally certain difficulties and tensions that can arise, an issue that has been evidenced in previous MOPAC evaluations of multi-agency initiatives36. The first-year evaluation found that there were some tensions between partners during project consultation, design and set up. These were related to both theoretical (e.g., the ethos or approach underpinning agencies) and practical (e.g., opening hours, response times) differences between partner agencies across statutory and voluntary sectors, with some concerned about maintaining and protecting their independence, identity, and ‘ways of working’. It was also noted that there were sometimes misunderstandings around the types of referrals from the Gateway into the partner agencies, and that perhaps the Gateway were trying to ‘fit’ survivors into services that were not wholly appropriate. However, year 2 interviews with stakeholders acknowledged that these were “all-natural tensions which can be ironed out in a constructive way”, and there was a suggestion that improving the partners’ awareness of each other’s differences has helped to reduce these tensions. This improved awareness appears to be in part due to the regular partnership meetings that are held; with stakeholders reporting that they find these meetings useful in facilitating positive interaction between them. Stakeholders were reportedly grateful to be involved in the partnership, acknowledging that the Gateway has given them a “seat around the table”. The regular partnership meetings enable relationships to be routinely developed and nurtured. Internal training - in particular with Navigators and ISVAs - was conducted jointly across the partnership during the initial stages of the project, drawing on the skills and expertise of colleagues from the different agencies involved. This offered opportunities to share knowledge around supporting survivors but was also important in terms of developing an understanding of each other’s work, which helped to address some of the concerns that arose during the planning stages of the pilot. The shared training was well received by stakeholders who noted that this helped build positive relationships at the start and provided a forum to discuss shared issues. It was noted by partners that staff would benefit from continued joint training to encourage more shared learning as the programme progresses. 36 Such as The Lighthouse, and The Persistent Offender Programme. https://www.london.gov.uk/what-we- do/mayors-office-policing-and-crime-mopac/data-and-statistics/academic-research “I think it’s about how we continue the spirit of how we have incredibly successfully collaborated, to what that means for the future as well” “We were invited when we first started to go out to every agency and see them, and that’s so nice because you get so much more excited and you can say to the survivor ‘I’m sending you off to these people and I trust them and it’s going to be really good” 23 A further suggestion for improving partnership relationships was for Navigators to meet with practitioners and partners face to face, so that each could continue to learn more about their respective services and roles. This was noted to be particularly important for new staff to build initial understanding and awareness of other services. In their own interviews, the Navigators also expressed the importance of knowing the professional to whom they were referring a client onto. For the Havens in particular, it was reported that the Gateway has helped to improve communication between the Havens and the four Rape Crisis Centres, with partners from the Havens now feeling more understood by the other agencies, and vice versa. In the initial stages of the pilot, there had reportedly been some misunderstandings about the Havens referrals and their processes, and that inappropriate referrals had been made to the Havens from the Gateway. These were discussed and resolved at the time by the agencies but again, the continued face to face interaction between Navigators and practitioners on the ground are suggested to further improve upon these processes. Online Mapping Service The online mapping service was intended to be a key tool for better communication and partnership working, and to provide agencies with an improved awareness of available services in London. The online map37 was developed by WGN and is located on the Gateway website which is available to the public and can be accessed by agencies and survivors themselves. In their interviews, stakeholders were asked about their knowledge and awareness of the online map. There was agreement from interviewees that the map could be useful, as the sexual violence service landscape is continuously changing. However, many stakeholders reported to not have seen and/or used the mapping service. They reported to understand the difficulties in doing such an exercise; that it’s resource intensive and would require a dedicated role to complete it fully. There is a particular challenge reported for services, and the Gateway Navigators, in keeping up to date with each other’s waiting lists - and when they have been closed and reopened. The Gateway have since developed a dashboard on their website for individual partners to access and regularly update on whether their lists are open/closed.38 Relationships between referring agencies and the Gateway Referrers to the Gateway were asked to participate in a short survey to report on their experiences of working with the Gateway. Mirroring the findings from the performance data, most referrers who completed the survey worked within the police (93%, n=52/56). Other sectors included VAWG third sector organisations (n=2), Health (n=1) and ‘Other’ (n=1). The nature of the feedback received was mostly positive, and there was a suggestion from findings that once referrers became aware of the Gateway, they went onto make continued and frequent use of the service. Indeed, half of respondents reported referring on a monthly basis (50%, n=28/56) and 29% (n=16/56) reported referring weekly. 37 https://survivorsgateway.london/service-map/ 38 This is a recent development, and one which occurred after the stakeholder interviews were conducted. 24 Perceptions of referrers’ relationships with Navigators were largely positive: n=18/56 reported the relationship as ‘satisfactory’, n=17 as ‘good’ and n=16 as ‘very good’. However, it was strongly reported by respondents in several parts of the survey that they would like to receive some form of feedback from the Gateway after making a referral (n=44 thought it was important to receive feedback); even a short confirmation email would be welcomed by some. It is not a current requirement of the Navigators to provide feedback to referrers (and would likely impact on their capacity). Despite this, more than half of respondents did report to have received feedback (n=20 sometimes; n=7 always; and n=5 often), with only one third (n=21) of respondents reporting having never received feedback. Overall feedback on the Gateway service was overwhelmingly positive: the majority of referrers responding to the survey felt that the Gateway had made referring into services easier (n=48/56), and in open ended responses noted that they found the overall referral process and the referral form simple to use. The Gateway as a single point of contact ‘hub’ was widely praised and seen to be mitigating the need for referrers to seek out available support, and that it provides a positive alternative for survivors who are concerned about engaging with police, and the Navigators’ specialist knowledge was also praised. There was a minority (one referrer and one partnership stakeholder) who held concern that the Gateway provides an additional ‘layer of repetition’ for survivors and as such there is duplication of work and the survivor having to tell their story multiple times. This point was further reflected during two interviews with Gateway service users, who were not always certain which services they had received or how they had been referred to them. Promisingly, most referrers ‘agreed’ that they had encouraged other colleagues to refer to the Gateway as well (86%, 48/56). This aligns with previous findings from stakeholders reporting the Gateway being a ‘victim of its own success’ in terms of the volume of referrals received. However, whilst the pilot has put a hold on further awareness raising activities, word of mouth among referrers may still be occurring. This is a positive reflection on the Gateway itself but poses a risk to managing capacity and responsiveness when put in the context of the ongoing resourcing issues in the partnership. Whilst the Gateway was highlighted as relieving the pressure on police, there were some challenges noted by both police and Navigators in relation to referral processes. Navigators reported sometimes receiving ‘inappropriate’ referrals in regard to high risk domestic violence survivors. A Navigator gave an example of receiving a police referral of a survivor who did not want to discuss their experience of sexual violence but required practical assistance in leaving a risky domestic violence situation. In this circumstance, the Navigator reportedly advised the referrer that the responsibility for statutory services to safeguard the survivor from the immediate threat of domestic violence was more urgent than the sexual violence support need; and that safeguarding for domestic violence is beyond the remit of the Gateway. Police fed back that they are also aware of the issue with referrals and it was noted that “there seems to be a bit of an argument between Police and Gateway staff on when a client requires an ISVA or an IDVA. I have had several discussions about this, being asked to make separate referrals only to be told my client requires what I had initially asked for.” Despite police being the largest referrers, they highlighted a further challenge for referring, where gathering information for the referral form can be problematic with survivors for whom a lack of trust in the police is a barrier. 25 Navigators also described some issues relating to police language and attitudes towards survivors of sexual violence, and that some police officers were still using language to describe sexual violence in a way that reinforces rape myths. For example, one Navigator explained how “there is still a lot of issues in terms of attitudes towards survivors and ideas, like belief and phrasing in the way they write about things” and another reported specific language issues: “we see repeatedly throughout the day ‘he had sex with her” (as opposed to using the word rape). The potential for Gateway staff to train SOIT officers in sexual violence was expressed as an option to be explored. Success for the future Interviewees were asked to reflect on the future and for the next phase of the Gateway. Expansion, funding and resourcing played a large role in this conversation. Stakeholders agreed that for future success and sustainability of the programme, additional resource was required at all levels: from the Navigators handling the calls to the practitioners working within the partner agencies. It was noted that the uncertainty of the funding landscape hinders the ability to forward plan and grow the programme: “otherwise it’s very reactionary and it’s just firefighting”. As previously discussed, partners also noted they would like to see additional agencies involved in the partnership, to cover the wide range of individuals who they support. Positively, since these interviews were conducted, the charity ‘Respond’39 have now officially joined the partnership. It has been clear from the interviews that partners have found the Gateway to be a positive model in bringing sexual violence support services together and putting services on the map. However, as is noted by one interviewee, the Gateway now faces the further challenge of maintaining this knowledge of all appropriate sexual violence services and expanding on it - with a view to replacing the responsibility for other agencies (such as GP surgeries) to hold this information). Some stakeholders expressed concerns for the sustainability of the Navigator role, and their conducting lengthy assessments on the phone, and discussed how it could or should change. Some partners are of the belief that the Navigator role should be reduced, in that they should not conduct long assessments with clients over the phone, with concern that clients could become attached to the Navigators, and it would also mitigate the risk of information being missed. Instead, some partners believe the Navigator calls should be ‘information-led’ and provide a signposting service only. On the future development of their own role, Navigators suggested that it would be useful to ‘tighten up’ the incoming referral form in terms of structure and they emphasised the importance of having solid data processing systems in place from the outset. When asked to provide advice to others setting up a similar service in a different location, Navigators focussed on three key areas to consider prior to delivery: the importance of meeting partners face to face; Gateway staff should receive full training; and the demand for the service should be overestimated, with one Navigator advising, “it’s going to be busier than you think it’s going to be”. 39 https://respond.org.uk/ 26 Results: Learning from year two of the Complex Needs ISVA service This section moves on to present findings from Year 2 of the complex needs ISVA service, which included six ISVAs from the partnership organisations. The section first explores findings from the performance analysis of year 2 the ISVA service (1st April 2019 to 31st March 2020), followed by process learning on the experience of the complex needs ISVAs. Complex Needs ISVA Performance Analysis The performance analysis comprises three sub-sections: 1) an overview of service user throughput, demographics and needs and demographics; 2) an outline of case outcomes; and 3) an overview of training and awareness activities conducted by ISVAs. Service User Throughput, Demographics and Needs Additional Complex Needs ISVAs (data from 1st April 2019 to 31st March 2020) • Over the duration of the funding for the additional complex needs ISVAs (mid-August 2018-March 2020), a total of 293 individual survivors were supported. • Complex needs ISVAs supported a total of 273 clients during Year 2 of the pilot. Of these, 171 (63%) were new clients and 102 (37%) were existing clients who were referred in Year 1. The rate of referrals for new clients in year 2 was lower than in year 1 (monthly average of 14 and 20 respectively) – however they have continued to support most Year 1 clients into year 2. • The number of year 2 referrals are not evenly split among the six ISVAs, and Survivors UK received the highest number of referrals to their complex needs ISVA (n=77, 28%), 273 ISVA service clients 93% Required ongoing support 70% Experienced multiple SV incidents 23% Completed the CJS process 27 followed by RASASC (n=59, 22%; see table 2 below for a full breakdown). The length of caseloads is not recorded, but anecdotally it was reported that the length of time that ISVAs may support an individual client can vary widely: between three months to four years depending on the needs and the length of the criminal justice processes. Table 2. Number of referrals per agency in Year 2 • Most clients in Year 2 self-referred (35%, n=95) or were referred via police (32%, n=87). These were also the most common referral routes in Year 1. • Over half of clients who engaged with the ISVA services were female (55%, n=149) and 32% (n=88) were male; these proportions are similar to Year 1. There is a larger proportion of male clients compared to Gateway referrals, due to Survivors UK ISVA receiving the largest number of referrals. Eleven survivors out of the total year 2 sample identified as non-binary, and one survivor identified as intersex40. Twenty-two clients out of the total sample identified as trans or as having had a trans history. • The largest age group of the ISVA clients was between 25-34 years (25%, n=69), with the second largest age category being 18-24 (22%, n=60). Compared to Year 1 the age splits are still broadly comparable. For clients where ethnicity was known (n=237), there was a very similar split between non-BAME (n=121, 51%) and BAME (n=116, 49%) (see Appendix F for full client demographics). Compared to Year 1 there is a higher proportion of BAME survivors being supported in Year 2. • Over half of clients (54%, n=147) had experienced rape and almost one third had experienced Child Sexual Abuse (29%, n=78). Compared to Year 1, this is a lower proportion of those who experienced rape (63%) and a slightly higher proportion who experienced child sexual abuse (23%). Other offences were related to other sexual assault (n=25), sexual exploitation (n=7) and trafficking/prostitution (n=2). Seventy per cent of ISVA clients (n=141/201) had experienced multiple incidents of sexual violence (a higher 40 Gender was not recorded for all clients. Agency Number of referrals Proportion Survivors UK 77 28% RASASC (South London Rape Crisis) 59 22% Galop 49 18% SOLACE (North London Rape Crisis) 37 14% NIA (East London Rape Crisis) 27 10% WGN (West London Rape Crisis) 24 9% Total 273 28 proportion than the 61% in Year 1), perpetrated by either the same person over a period of time and/or by difference perpetrators at different times, and the majority required ongoing ISVA support (93%, n=254), as opposed to a single incident of support. • Most clients were recorded as having experienced some form of disadvantage, disempowerment or discrimination (n=220), and 78 of these had experienced multiple forms. Consistent with Year 1 findings, long term or complex mental health issues were still the most prominent (n=100, 37%), followed by repeat victimisation (n=50, 18%), physical disabilities (n=26, 10%), and problematic substance misuse (n=35, 9%) (see table 3). Table 3: Disadvantage, disempowerment, and discrimination experienced by ISVA clients Type of disadvantage, disempowerment, discrimination Number of survivors with type of disadvantage (out of 273) Proportion Long term or complex mental health issues 100 36.6% Repeat victimisation 50 18.3% Physical disabilities 26 9.5% Problematic substance use 25 9.2% Learning disabilities 20 7.3% Under 18 20 7.3% Homophobia, biphobia or transphobia 19 7.0% Immigration issues 17 6.2% Homelessness 11 4.0% ‘Honour’ based violence 9 3.3% Harmful practices 3 1.1% Leaving care 3 1.1% modern day slavery 2 0.7% Children's services involvement 1 0.4% Long term health condition 1 0.4% Perpetrator is police officer 1 0.4% Learning disabilities 1 0.4% Effects of CSE 1 0.4% Outcomes for ISVA clients • Criminal justice information was recorded for 266 of the ISVA clients. At the time of reporting, almost half (45%, n=120) were ongoing cases, for 18% (n=49) the outcome was unknown or the information was not obtained, whilst for 12% (n=33) the crime was not reported to police. • For the 64 criminal justice cases where there was a recorded outcome; the majority (n=44, 69%) were police NFA (no further action) and a further 11% (n=7) were CPS (Crown 29 Prosecution Service) NFA. The remaining outcomes were: 8% (n=5) unsuccessful prosecution; 6% (n=4) conviction; 2% (n=1) acquitted at court; and 2 clients withdrew from the criminal justice process. Comparing these outcomes to other research such as the MOPAC London Rape Review (2019)41, the ISVAs’ clients have a considerably higher proportion of cases with an NFA outcome (compared to 29% in the Rape Review), and a considerably lower proportion of cases where the client withdrew (compared to 58%). Another study42 which examined rape cases in two police forces in South West and North East England, also had significantly fewer cases with NFA outcomes (56%) and significantly higher cases of victim withdrawal (34%) compared to the ISVA clients.43 • Each host organisation captured outcome measures, either based on client questionnaire answers or worker assessments. Outcome data was provided for 216 ISVA clients.44 Overall, this showed many improvements, particularly in relation to service users feeling more able to access further support and to assert their rights – a finding that is similar to year 1. Changes were less pronounced in other categories, particularly in relation to survivors feeling able to develop and maintain positive relationships, and regarding better health and wellbeing (see table 4). Table 4: Outcomes for ISVA clients Outcome Deteriorated No change Improvement Support ongoing Total More in control of lives 0% (n=1) 38% (n=81) 56% (n=119) 5% (n=11) 212 Better Health & Wellbeing 4% (n=9) 43% (n=93) 44% (n=95) 9% (n=19) 216 More able to access further support 1% (n=2) 24% (n=51) 70% (n=152) 5% (n=11) 216 More able to develop & maintain positive relationships 0% (n=0) 57% (n=123) 35% (n=76) 8% (n=17) 216 More able to assert their rights 0% (n=0) 21% (n=46) 75% (n=161) 4% (n=9) 216 The Experience of Complex Needs ISVAs during Year 2 This section moves on to consider the learning generated from focus groups that were conducted with the complex needs ISVAs. To set some context, in focus groups during year 1 of the pilot, ISVAs described their services to survivors as being flexible, trauma-informed, and person-centred. ISVAs hold a holistic view of survivors’ complex needs and ensure that these are recognised and addressed in their interactions 41 MOPAC, 2019 42 Walker et al. (2019) 43 These were statistically significant differences with a 95% confidence level. It should be noted that these comparisons were for rape cases, and ISVA clients supported survivors of rape and other forms of sexual violence. 44 Outcomes data was not provided for all clients, likely due to the fact that many were still being supported and hadn’t yet reached the stage of being able to assess their outcomes. 30 with other services. ISVAs often defined their role in terms of an advocate, ally and representative for survivors of sexual violence, whose needs are not only multiple but compound, magnifying each other and deepening vulnerability. ISVAs also described that they act as an ‘interpreter’ or ‘translator’, communicating survivors’ needs to ensure that professionals in other agencies acknowledge and respond appropriately. Supporting survivors with complex needs Reflecting on the development of their roles during year 2 of the pilot, the ISVAs agreed that their role had expanded, in the sense that they were seeing an increase in the types of needs that clients presented with. The wide range of the ‘complex needs’ that form a core aspect of their work has featured heavily in previous focus groups, where ISVAs talked about how these needs can heighten the barriers that survivors face in accessing services, navigating systems, and achieving positive outcomes (which may, or may not, be criminal justice in nature), particularly where their circumstances are characterised by insecurity or risk. This can include immigration status, language barriers, inadequate housing, lack of funds to travel to appointments, and deep chronic trauma, which requires emotional support before survivors feel able to engage with services. During year 2, ISVAs reported that they have increasingly been supporting clients dealing with legal issues; including civil and family court, Victim’s Right to Review (VRR), immigration, asylum, and no recourse to public funds. Additionally, ISVAs mentioned an increase in clients presenting with mental health needs, especially when their mental health can affect how they present their story or testimony, like schizophrenia. Housing concerns are a recurring issue for the ISVAs in supporting many of their clients. However, it was discussed that there are also gender differences in survivors being able to access some support. One of the ISVAs supports male survivors and it was reported that there are particular issues for male survivors in obtaining certain types of support which women are more able to access. For example, the ISVA spoke of a male client who was in a domestic violence relationship, but who was told by homeless shelters in this instance that he had to be a street sleeper in order for him to be supported into housing. This is further compounded by there being no emergency accommodation or DV beds for male survivors. Training requirements 31 The ISVAs felt that they required further training to be able to address and keep up with the ever increasing, multiple complex needs outlined above - to be in a better position to advise or signpost to appropriate support This included family court issues where ISVAs expressed the need to better understand the processes involved, but also supporting clients with complex mental health issues, particularly ensuring they are appropriately supported so that they can give their best evidence at court, and finding different ways of working with clients with learning difficulties that felt ‘creative’ and not ‘reductive’. Challenges and barriers As discussed earlier in the report; Navigators described some issues they experience relating to police language and attitudes towards survivors of sexual violence. Many similar views emerged from speaking to the complex needs ISVAs. From the first focus group in September 2018 through to the most recent in April 2020, ISVAs have repeatedly reported enduring myths about sexual violence that are still held by some parts of the criminal justice system (i.e., stereotypical or misled beliefs sometimes related to issues such as alcohol, levels of physical violence, and relationships between survivors and perpetrators). Indeed, ISVAs described one of the main barriers in their role as the lack of understanding around sexual violence by other agencies, and how it can have a “tragic impact” on the survivors. The ISVAs reported that there should be more awareness raising and training across all sectors that come into contact with survivors (particularly the acutely marginalised such as asylum seekers or rough sleepers), both on the nuances of sexual violence and the professionals that support them. Indeed, ISVAs highlighted situations where they felt they had to explain or justify their own role to police officers or court staff. In the focus groups early on in the pilot, ISVAs told that they were met with negative reactions and spoke about police officers believing they were ‘on the opposite side’ and ‘there to confront them’. In the most recent focus group it was evident that a perceived lack of respect from some other professionals still remains a key barrier, where CJS professionals haven’t recognised the role of the ISVA: “I’ve personally had the experience where the solicitor has refused to acknowledge my presence and was speaking directly to the [survivor], even though the woman said she didn’t want to speak to them”. Another ISVA expanded on the impact of these attitudes towards the ISVA role: “she said ‘I don’t know what an ISVA stands for, so I’ll call it a support worker.’ It’s an “I feel like the systemic barriers and then barriers depending on who you are – your identity and mental health needs – all of those things will be used to attack your credibility and all of those things are massive barriers for those survivors we support, but it also impacts us doing the work because we constantly have to push back” “In terms of learning difficulties and a greater awareness of how that can impact, the discrimination that the survivor is facing and also sometimes ways of working that better meet their needs, so maybe creative ways of working - I’ve got a client who works better with diagrams and timelines – so creative ways of working with people with learning differences” 32 accredited role, so of course we’re not going to be taken seriously if the people we are meant to be working most in partnership with – the CJS – doesn’t take it seriously.” The ISVAs discussed concerns for supporting clients who have experienced delays with the criminal justice system, and it’s a “real pressure for survivors who have come forward to stay in the CJS for that long”, which has an effect on the case work that the ISVAs do with them. The ISVAs also discussed this issue in light of the Covid-19 pandemic who anticipate that the delays in the system will be ‘tremendous’ and will have significant demand on the survivors, particularly those with complex mental health issues, and the knock-on effect that could have on mental health services. There was also strong concern expressed of the anticipated expanding caseload of supporting existing clients experiencing delays in their cases, whilst taking on new clients. At the time of conducting this focus group there was a Covid-19 lockdown and the ISVAs described being particularly concerned for victims being ‘locked in’ with their abuser or very difficult family members. Overcoming challenges and finding successes Whilst there is much room for improvement in terms of the ISVAs interactions with other professionals, it was acknowledged that there are many professionals with whom they have positive interactions within the criminal justice system. One ISVA noted that the longer they have been in the role, they notice the more meaningful connections they make with certain professionals which ultimately should have a positive impact on the client’s outcomes. Evidence has also shown that ISVAs play a key role in awareness raising of the London Survivors Gateway pilot to other agencies and engaged with professionals from various sectors.45 Being an ISVA was reported to be challenging work – most notably in terms of supporting traumatised survivors, but also managing sometimes difficult and defensive relationships with other professionals – underlining the importance of clinical supervision and peer support for ISVAs. Despite the challenges of their work, the ISVAs reported take strength from some of the personal successes that they see and experience with clients. The ISVAs gave some anecdotes of types of success they have come across, which were often focussed on the criminal justice experiences and outcomes for their clients; from having positive relationships with some SOIT officers and seeing the positive impact that has on the client, to witnessing some 45 Training and awareness raising had been delivered to 2,172 professionals from a range of agencies including the Metropolitan Police Service (MPS), British Transport Police, housing/homeless sector, local authority, NHS, youth workers, psychologists, sexual health workers, education, and voluntary sector organisations. There were also many awareness raising sessions with the aim of engaging with potential service users; which had a total estimated audience of 2,800 individuals in locations including sexual health clinics, youth groups and events such as Pride, Domestic Violence Awareness days and for women involved in prostitution. “At the beginning communication was difficult and they couldn’t understand our role and that is still the case for a lot, but there are some really positive examples where they have seen this approach where we are supporting clients with lots of different needs and they are actually pointing out the difficulties from the clients perspective to engage with the CJS.” 33 clients’ cases reach trial with a guilty verdict, and seeing the positive effects of this on the clients when they feel ‘empowered’. 34 Conclusion From its ‘going live’ date in October 2018 to the end of the evaluation period in March 2020, 9,843 attempted calls made to the Gateway and a total of 2,988 survivors of sexual violence were registered as case clients for the Gateway service. Furthermore, in the period between July 2018 and March 2020, the six additional complex needs ISVAs have supported 293 survivors. There is also now a pan-London map on the Gateway website which provides access to information about sexual violence service availability in London, which has not previously been available in this format. The Gateway has been well received by service users, referring agencies and stakeholders. Overall, user feedback suggests that it has made access easier for survivors and referral paths more straightforward for referrers – and in this way the Gateway largely appears to have met one of its key aims: making it easier for survivors to access support and advice. However, it should be noted that 43% of calls to the Gateway were not successfully connected, and whilst some of those individuals may have gotten through on a subsequent attempt, it is not possible to obtain their feedback on this type of experience of the service. Despite the overall positive feedback there were a minority who raised the risk of the Gateway becoming one more layer within an already complex and ever-changing landscape of services. This is something that should be continuously reviewed to ensure the Gateway continues to position itself as that central point that brings partners and services together and eases referral pathways. Continuing to find ways to nurture relationships with existing partners as well as expanding the partnership will play an important role here. Demand and resourcing have been key challenges for the Gateway throughout the pilot. These types of issues for sexual violence services are not new and were highlighted in the Sexual Violence Needs Assessment, where it was described that services are struggling to meet demand with limited resources. For the Gateway, a similar story emerges where the number of Navigators in place at the start were too few to meet the demand and the volume of calls that came through - reflecting some of the early concerns raised by partners. Over the course of the pilot, various changes were made to how the Gateway was run, including recruitment of additional staff and amendments to some processes. However, demand challenges continued and it is likely that balancing awareness-raising with readying for an increase in demand will be an ongoing task for the Gateway to manage. Evidently, the resourcing issues from the start quickly had an impact on the Navigator role. Its original design was to triage survivors into suitable sexual violence services, after conducting needs assessments over the phone. In practice, the role evolved to encompass more than this - it quickly shifted to conducting lengthy assessments over the phone with survivors and providing emotional support as they were often a first point of support for that survivor. An additional issue is that the sexual violence services were still struggling with their own capacity, and unable to take on new referrals at the rate that they were coming through to the Gateway. As such, it was reported Navigators began to ‘hold’ cases themselves, to funnel them into support services. Indeed, Navigators were able to refer just under half of registered case clients to other support services - and it is worth noting this does often means a referral onto a waiting list. For one in 35 ten case clients there was no suitable service available, something that survivor feedback suggested had a detrimental impact on their recovery, mental health and emotional well-being. There is a risk to both the quality of support that can be given to vulnerable survivors, but also the wellbeing of Navigators if caseloads are too high and demand unmanageable. Given how the Navigator role in particular had to evolve, it may now be time to review this crucial role and how it can be enabled. For example, in order to be able to return to how the role of the Navigator, and the remit of the Gateway, was initially conceived - that of a triage and signposting service - a more robust and reliable service landscape would be needed. Indeed, there was a clear indication made by stakeholders that extra funding would need to be secured to continue to meet the demand. Moreover, the Gateway had to put their awareness-raising work on hold to prevent further increases in volume. This was a difficult, but necessary decision to keep the integrity of the service but means that there are currently survivors who are not being reached and are not being able to benefit from the support offered by the Gateway service. Good partnership working is key in implementing a multi-agency pan London approach and the Gateway pilot has highlighted - unsurprisingly - the importance of good partnership working in supporting sexual violence survivors across the capital. This relates to services providing support for survivors, both those already within the Gateway partnership or those the Gateway refers into, but also wider working between partners whose work brings them into contact with survivors. Positively, lots of issues that were seen at the start for partnership working have since improved and the Gateway clearly has played - and continues to play - an important role in bringing partners together across London. This is an important aspect to continue and further build on in order to nurture relationships and to continue to facilitate good communication, information- sharing and knowledge exchange. Indeed, it was reportedly important for partners to regularly meet with each other, ideally face to face; this is particularly important for new members of staff, or new services joining the partnership. It is therefore a recommendation that these ‘meet and greets’ are mandatory for new staff joining the partnership. Evidently, knowledge of service provision is a key factor in the success of the Gateway and it is therefore recommended that regular knowledge and information sharing activities should be prioritised. There is also potential to improve mutual understanding and working together between agencies, especially between support services and criminal justice agencies, such as the police. It was reported by all practitioners interviewed that there are still issues with some parts of the police and criminal justice services in how they speak to survivors of sexual violence, particularly the language and rape myth terminology that is still being used. Therefore, further awareness-raising and knowledge exchange is recommended, but also joint training, and would need to cover use of language with sexual violence survivors and rape myth terminology, but also an understanding of the different support roles such as ISVAs. This is clearly important to address for the survivors and their experiences, but to also improve partnership working and address some of the challenges in the relationship between the Gateway and particularly the police (as the largest referrers) that were highlighted in this research. Finally, WGN’s achievement to collate data across different Gateway services and different partners should be acknowledged. This has been time consuming yet important work, ensuring 36 that the various activities and elements of the pilot could be documented and evidenced. There is scope to build upon this by further streamlining and reviewing data collection processes across agencies and consider where the collection of more granular information may enable a more comprehensive picture of the service, the needs of service users as well as demand and capacity issues. For example, collecting data on the duration of phone calls or average case lengths would help to gauge capacity and resource issues in more detail, whilst the recording of multiple referrals for one survivor would to help tell the story of capacity challenges more widely in sexual violence services. 37 References Allen, N.E., Bybee, D.I. and Sullivan, C.M. (2004) ‘Battered Women’s Multitude of Needs: Evidence Supporting the Need for Comprehensive Advocacy’ Violence Against Women 10 (9): 1015-1035 Astbury, J. (2006) Services for Victim/Survivors of Sexual Assault: Identifying Needs, Interventions and Provision of Services in Australia Melbourne: Australian Institute of Family Studies Campbell, L., Keegan, A., Cybulska, B. and Forster, G. (2007) ‘Prevalence of Mental Health Problems and Deliberate Self-Harm in Complainants of Sexual Violence’ Journal of Forensic and Legal Medicine 14 (2): 75-78 Ceelen, M., Dorn, T., van Huis, F.S. and Reijnders, U.J. (2016) ‘Characteristics and Post- Decision Attitudes of Non-Reporting Sexual Violence Victims’ Journal of Interpersonal Violence 34 (9): 1961-1977 Dawson, P. and Stanko. B. (2013) Implementation, Implementation, Implementation: Insights from Offender Management Evaluations Policing: A Journal of Policy and Practice 7 (3): 289- 298 Krug, E. G., Dahlberg, L.L., Mercy, J.A., Zwi, A.B and Lozano, R. (Eds.) (2002) World Report on Violence and Health Geneva: World Health Organization Hester, M. and Lilley, S.J. (2018) ‘More Than Support to Court: Rape Victims and Specialist Sexual Violence Services’ International Review of Victimology 24(3): 313-328 Home Office (2017) The Role of the Independent Sexual Violence Advisor: Essential Elements London: Home Office James, V.J. and Lee, D.R. (2015) ‘Through the Looking Glass: Exploring how College Students’ Perceptions of the Police Influence Sexual Assault Victimization Reporting’ Journal of Interpersonal Violence 30(14): 2447-2469 Lea, S., Falcone, M., Doyle, K. and Reardon, S. (2015) An Audit of Independent Sexual Violence Advisors (ISVAs) in England and Wales London: King’s College London and Lime Culture Lovett, J. and Kelly, L. (2009) Different systems, similar outcomes? Tracking attrition in reported rape cases across Europe. London: CWASU MBARC (2016) Sexual Violence: The London Sexual Violence Needs Assessment for London for MOPAC and NHS England (London) London: MBARC 38 MOPAC (2019) The London Rape Review: A Review of Cases from 2016 London: MOPAC Natcen (2015) Violence, Abuse and Mental Health in England (REVA Briefing 1) London: Natcen, Truth Consulting and CWASU, London Metropolitan University Robinson, A. (2009) Independent Sexual Violence Advisors: A Process Evaluation Research Report 20 London: Home Office Sigurvinsdottir, R. and Ullman, S.E. (2015) The Role of Sexual Orientation in the Victimization and Recovery of Sexual Assault Survivors Violence and Victims 30(4): 636-648 Smith, N., Dogaru, C. and Ellis, F. (2015) Hear Me. Believe Me. Respect Me: A Survey of Adult Survivors of Child Sexual Abuse and their Experiences of Support Services Suffolk: University of Suffolk Stern, V. (2010) The Stern Review: A Report by Baroness Vivien Stern CBE of an Independent Review into how Rape Complaints are Handled by Public Authorities in England and Wales London: Government Equalities Office/Home Office Ullman, S.E. and Townsend, S.M. (2007) ‘Barriers to Working with Sexual Assault Survivors: A Qualitative Study of Rape Crisis Center Workers’ Violence Against Women 13(4): 412-443 Ullman, S.E. (2016) ‘Sexual Revictimization, PTSD, and Problem Drinking in Sexual Assault Survivors’ Addictive Behaviours 53: 7-10 Walker, S.L, Hester, M., McPhee, D., Patsios, D., Williams, A., Bates, L. and Rumney, P. (2019) ‘Rape, inequality and the criminal justice response in England: The importance of age and gender’ Criminology & Criminal Justice 1-19 Zinzow, H.M. and Thompson, M. (2011) ‘Barriers to Reporting Sexual victimization: Prevalence and Correlates among Undergraduate Women’ Journal of Aggression, Maltreatment and Trauma 20 (7): 711-725 Appendices Appendix A: Key findings from the interim report, July 2019 The interim report reflected on initial learning generated in year one of the London Sexual Violence Triage Pilot. The report presented the ‘activity’ of the project through performance data and reflected on process learning generated throughout fieldwork conducted. Key Findings: 39 • There were 3,309 attempted calls made to the Gateway (in addition to 563 online referrals) over the reporting period (1st October 2018 to 31st March 2019) with just under half of calls made within opening hours (47%, n=1,448/3,077) successfully connected to a navigator. While stakeholders and practitioners interviewed welcomed the service as a timely and straightforward pathway to support, this suggests demand for the service that outweighs current resources. • There were 636 case client referrals to the Gateway, with 398 outward referrals for support - most of which were to agencies within the pilot partnership. This demonstrates the value of a strong partnership approach to supporting survivors of sexual violence. Indeed, partner agencies were keen to build on existing relationships and develop new ones; however, outlined some tensions between partners during pilot consultation, design and set up. • Complex needs ISVAs supported 122 clients (between mid-August 2018 and mid-February 2019), all of whom experienced some form of disadvantage disempowerment, or discrimination, most notably long-term mental health illness. This underlines the importance of specialist support for survivors, particularly given levels of under reporting to the police. • ISVAs and navigators who took part in evaluation fieldwork highlighted frustrations around availability of support services, additional barriers that survivors with complex needs face, and the negative effects of myths about sexual violence (i.e., stereotypical or misled beliefs sometimes related to issues such as alcohol, levels of physical violence, and relationships between survivors and perpetrators) that still exist amongst some parts of the police service (the valuable support offered to survivors by many police officers was also acknowledged). • Looking to the future, there was concern about the short term and insecure nature of pilots, and the effect this may have on efforts to support chronically disempowered survivors which requires significant time – in some cases longer than the two years of this pilot. Appendix B: London Survivors Gateway website home page (https://survivorsgateway.london/) 40 41 Appendix C: Gateway Process Source: WGN presentation to VAWG Transformation Event, January 2019 42 Appendix D: Needs Assessment Scenarios Pathway of supporting a survivor from referral, through to needs assessment, to onward referral. Examples of Navigator scenarios 43 Appendix E: Demographics of case clients Gender Gender Number Female 2189 44 Male 139 Non-binary 13 Trans Gender Female 7 Trans Gender Male 4 Total 2352 Age Age Band Number Up to 12 years 1 13 to 17 years 168 18 to 24 years 660 25 to 34 years 807 35 to 44 years 374 45 to 54 years 219 55 to 64 years 87 65 to 74 years 16 75 to 84 years 6 85 and over 2 N/A 12 Total 2352 Ethnicity Ethnicity Number White 1043 Prefer not to say 433 Black or Black British 376 Asian or Asian British 227 Mixed or dual 147 Any other ethnic group 112 Chinese 14 Total 2352 Disability Disability Number of clients Yes 650 No 1260 Undisclosed 2 Prefer not to say 440 Total 2352 45 Sexuality Sexuality Number of clients Heterosexual 1368 Prefer not to say 441 Undisclosed 290 Bisexual 115 Unsure/Questioning 46 Lesbian 31 Gay man 22 Other 14 Pansexual 11 Queer 8 Asexual 3 Celibate 1 Not appropriate 2 Total 2352 Appendix F: Demographics of complex needs ISVA clients Gender Number of survivors Percentage Female 149 55% Male 88 32% Unknown 23 8% Non-binary 6 2% Non-binary/static 5 2% Intersex 1 0% Prefer not to say 1 0% Total 273 46 Age range Number of survivors Percentage 13 – 17 24 9% 18 -24 60 22% 25 -34 69 25% 35 - 44 48 18% 45 - 54 53 19% 55 - 64 13 5% 65 and over 6 2% Total 273 Ethnicity Number of survivors Percentage White 121 44% Black 55 20% Asian 33 12% Other 28 10% Prefer not to say 19 7% Unknown 17 6% Total 273 Cost of living research and crime in London Magdalena Dominguez and Tom Kirchmaier February 2024 Purpose and outcomes of this study The aims are to: • Predict what impact the cost-of-living crisis has across a range of crime types. • Explore where the impact may be felt the most (or on which groups). • Explore the wider harms of the crisis, both in the short and longer-term. Findings suggest: • A 10% increase in cost of living is associated with an 2.4% decrease in police demand and an 8% increase in crimes. • The most affected areas are the very city center and eastern outer areas. • Most impact will take place in 2023 (25.6% increase in crime). In 2024 milder impacts are expected (0.6%). MOPAC commissioned analytics to explore the impact of the Cost-Of-Living Crisis on Londoners Setup The cost-of-living crisis: an overview • The cost of living has been at worrying high levels during the past two years in the United Kingdom. • Inflation reached its highest value in 40 years in October 2022 (annual increase of 11.1%). • The latest measure (July 2023) is of 6.8%. • Prices have grown more than wages, reducing real income and purchasing power of households. • 50% of adults reported that they were worried about the cost of energy and food (Office for National Statistics, Opinions and Lifestyle Survey). • This mostly affects low-income households as they spend a higher share of their income on energy and food, the prices of which rose most. The latest hike in inflation. 2022 to date The graphs highlight the trends in inflation. While the left figure plots the variation in consumer price index from 2004, the graph on the right side expands on the latest registries. They portray a substantial increase in inflation since 2022. Consumer Price Index at highest growth rate in 40 years: Differences in inflation • Inflation impacts different types of households in different ways. • Not all households experience the same average rate of inflation because of the difference in their underlying consumption basket. • For example, richer households experienced a lower impact of inflation than those at the bottom of income distribution. • Consumption basket differences: food and energy (prices that have grown the most) weight differently in both groups. Differences in inflation, graphical evidence The situation has been worse for low-income households and for renters The graphs highlight the trends in inflation depending on housing tenure (left panel) and income levels (right panel). The left panel shows that inflation is higher for subsidised renters as compared to owner-occupiers, and the right one shows that inflation is higher for the lower income quintiles. Thus, they highlight differential effects for different housing groups. Does this matter for police agencies? • Academic literature has shown that economic hardship relates to crime. • Economic incentives (Draca and Machin, 2015), Deprivation (Kawachi et al., 1999), Recessions (Bell, Bindler and Machin, 2018), Welfare cuts (Melander and Miotto, 2022). • The relationship is complex and multifaceted. • Economic incentives: acquisitive crime (burglary, theft, fraud/scams). • Psychological effects: interpersonal violence (violence against the person, domestic violence). • Social disorganization: gang recruitment. • How applicable are previous research findings to the current cost of living crisis situation? Data and Model A detailed COLC measure We construct a Cost-Of-Living Crisis (COLC) measure that varies over time and space, using: • Consumer Price Indices (“CPI”, monthly variation) – It indicates the inflation level at a time. It is taken from the ONS. • Household composition (“n”, spatial variation) - It indicates the composition of the population, as it will have different consumption baskets. Taken from the 2021 Census. A detailed COLC measure • Result: COLC measure is a “weighted average” CPI at the OA-month level. 𝐶𝑂𝐿𝐶𝑖𝑡= ෍ 𝑔=1 𝐺 𝐶𝑃𝐼𝑔𝑡. 𝑛𝑖𝑔 𝑛𝑖 In simple terms, COLCit provides a very detailed inflation measure, for a specific point in time (a month) and a specific area in London (an output area). Inflation for population group “g” at month “t” Relevance of population group “g” in area “i” 1. Output Areas (OAs) are the lowest level of socially homogeneous geographical area for census statistics. They have a resident population between 100 and 625 persons. A police demand measure • The outcome variable to analyze from changes in COLC is police demand. • We proxy it with Computer Aided Dispatch (CAD)1 call data from 2019-2022 for London. • Approximately 10.3 M registries • Date --> Year and Month • Geographical coordinates --> Output Area (OA) • Opening code --> Call Category CAD Call Category Average per OA-Month % Total All Calls 8.29 100 Public Safety 2.41 29 Crime 2.00 24 Anti-Social Behaviour 0.95 11 Administration 2.35 28 1. CAD is a crime management system to manage the reporting of crimes and allocate police forces accordingly. Every time a call comes in, a call handler at the back end records the details of the crime incident on CAD, and further flags it with appropriate required response. During the process, a sequential CAD or incident log number is automatically generated. This facilitates easy follow-up actions for either party to serve the interests. COLC and CAD Calls: geographical variation COLC measure – 2022 average CAD Calls – All – 2022 average The maps represent the COLC measure (on left) and the volume of CAD calls (right) distributed across output areas in Greater London in 2022. Darker shades of red indicate higher values, and the distribution gives us an idea about the geographical variation of COLC and calls. Analysis setup: Correlational study Estimating equation: 𝐶𝑎𝑙𝑙𝑠𝑖𝑡= 𝜷. 𝐶𝑂𝐿𝐶𝑖𝑡+ 𝜆𝑖+ 𝛿𝑡+ 𝜀𝑖𝑡 • The fixed effect for Output Areas 𝜆𝑖is introduced in the model to control for variation in OAs across London. • The fixed effect for calendar months 𝛿𝑡is introduced in the model to control for common cyclical variables that are common to all areas. • Our left-hand side or dependent variable reflects the registered CAD call data for each OA-month pair. • 𝜷captures the average correlation between the cost-of-living measure and police demand. Side note: Confidence Intervals • From our estimation, we will obtain the impact of cost-of-living on CAD calls on average. This is referred to as a point estimate. • It is also relevant to provide ‘Confidence Intervals’ to these estimates. This will provide a within which the actual impact from 2019- 2022 is likely to fall. • Confidence Interval will also be used to determine the uncertainty associated with point estimates of predicted effects for 2023 and 2024. • The lower bound and upper bound of the confidence interval of point estimates is calculated as: Lower-bound of CI = point estimate - ( z * standard error) Upper-bound of CI = point estimate + ( z * standard error) 1. The lower bound of Confidence Interval indicates the lowest value of the range within which our point estimate is likely to fall. 2. The upper bound of Confidence Interval indicates the highest value of the range within which our point estimate is likely to fall. 3. ‘z’ indicates the level of confidence we have in the intervals created from the standard normal distribution. In simple words, it indicates how likely it is that the interval we have constructed contains the actual point estimate. We have conducted our analysis at 95% of the Confidence Interval Main Findings Main findings – Correlational study Estimating equation: 𝐶𝑎𝑙𝑙𝑠𝑖𝑡= 𝜷. 𝐶𝑂𝐿𝐶𝑖𝑡+ 𝜆𝑖+ 𝛿𝑡+ 𝜀𝑖𝑡 Findings: • A 10% increase in the cost of living (an intuitive point of reference that has a similar value to the recent largest increase in inflation) is associated with a decrease of all calls by 2.4% with respect to their mean. This is for all calls across all the sample period (2019-2022). • It is associated with a decrease on Anti-Social Behaviour and Administration calls (approx. 27% and 13% respectively). Decreases in ASB Nuisances and Police Generated Resource Activity calls. • It is associated with an increase on Crimes and Public safety calls (approx. 8% for both). Increases in Burglaries, Thefts, Violence Against The Person, Abandoned Calls and Wanted -Police/Court Order/Bail. Correlational study- All Categories Note: The table represents the coefficient 𝜷obtained from regressing COLC measure on the large call categories we already presented using Equation 1. ***, **, * indicate significance at the 1%, 5% and 10% respectively. All Calls Public Safety and Welfare Crime Anti-Social Behavior Administration Violent COLC -0.013*** 0.011*** 0.010*** -0.015*** -0.019*** 0.006*** (0.002) (0.002) (0.001) (0.001) (0.001) (0.000) Observations 1,242,816 1,242,816 1,242,816 1,242,816 1,242,816 1,265,712 R-squared 0.895 0.738 0.752 0.516 0.908 0.681 Correlational study- Crime Categories Note: The table represents the coefficient 𝜷obtained from regressing COLC measure on 18 different categories of crime using Equation 1. We find an increase in acquisitive crimes (theft, burglary , shoplifting), violence against the person and drug offences. We don’t find an association between COLC measure and sexual offences. ***, **, * indicate significance at the 1%, 5% and 10% respectively. Sub-Categories for Crime Abduction Kidnap Bomb Threat Burglary Dwelling Burglary Other Than A Dwelling Criminal Damage Drugs Offence COLC -0.000 0.000** 0.002*** 0.001*** 0.000 0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Fraud And Forgery Harassment Act Offences Malicious Communications Robbery Sexual Offences Sexual Offences Rape COLC -0.001*** -0.000* -0.001*** 0.000** -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Theft Shoplifting Theft Of Motor Vehicle Theft From Motor Vehicle Theft Other Unlisted Crime Violence Against The Person COLC 0.001*** 0.001*** 0.001*** 0.001*** -0.000 0.003*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Correlational study- Public Safety Categories Note: The table represents the coefficient 𝜷obtained from regressing COLC measure on 18 different categories of public safety using Equation 1. We find an increase in abandoned calls and wanted calls. ***, **, * indicate significance at the 1%, 5% and 10% respectively. Sub-Categories for Public Safety Abandoned Call Absconders AWOL Alarm Police Installed Alarm Central Station Alarm Premises Audible Only Animals Pets Domesticated COLC 0.007*** 0.000 0.000 0.001*** 0.000** 0.000** (0.001) (0.000) (0.000) (0.000) (0.000) (0.000) Civil Disputes Collapse Illness Injured Trapped Concern For Safety Domestic Incident Firearms Hoax Call To Emergency Services COLC 0.000*** 0.001*** -0.001 -0.001*** 0.000 0.001*** (0.000) (0.000) (0.001) (0.000) (0.000) (0.000) Immigration Industrial Incident Accident Insecure Premises vehs Licensing Missing Person Natural Disaster Incident Warn COLC 0.001*** -0.000** 0.000 -0.000*** 0.001 -0.000*** (0.000) (0.000) (0.000) (0.000) (0.001) (0.000) Protest Demonstration Sudden Death Suspicious Circumstances Suspicious Package Object Truancy Wanted Pol Crt Order Bail COLC -0.000** 0.000*** -0.001*** 0.000*** 0.000*** 0.002*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Correlational study- Violence Categories Note: The table represents the coefficient 𝜷obtained from regressing COLC measure on different categories of violent crime using Equation 1. ***, **, * indicate significance at the 1%, 5% and 10% respectively. Sub-Categories for Violent Crime Violence Against The Person Sexual Offences Burglary Dwelling Burglary Other Than A Dwelling COLC 0.003*** -0.000 0.002*** 0.001*** (0.000) (0.000) (0.000) (0.000) Robbery Bomb Threat Abduction Kidnap Sexual Offences Rape COLC 0.000** 0.000** -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) Correlation is not causation • Our current findings indicate a significant correlation between the cost of living and CAD calls. • However, they do not imply causation (i.e. an increase in cost of living causes a change in CAD calls). Further Findings Further Evidence Temporal Estimation: We now allow the effect of COLC to vary over time. 𝐶𝑟𝑖𝑚𝑒𝑖𝑡= 𝜷𝒕. 𝐶𝑂𝐿𝐶𝑖𝑡× 𝑇𝑖𝑚𝑒𝑡+ 𝜆𝑖+𝛿𝑡+ 𝜀𝑖𝑡 where, • 𝑇𝑖𝑚𝑒𝑡ranges from Jan 2019 to Nov 2022, with Dec 2022 as base. • 𝜷𝒕allows us to estimate the effect of COLC on crime for different points in time. • We do not allow the effect to vary over space too (aggregate measure: London). Variation over time (𝛽𝑡): 2021 onward Panel A: Calls for violent crimes Panel B: Calls for crimes Panel C: All calls • The graphs show the point estimates of the correlation between COLC and CAD calls over time. Each dot is the estimated correlation for that month. The vertical line around each point shows its confidence interval. • While the correlation weakens (gets closer to zero) from 2021 to early 2022, we observe a larger negative correlation of the COLC measure on all calls, crimes and violent crimes for the last few months in the sample. Further Evidence Spatial Estimation: We now allow the effect of COLC to vary over space. 𝐶𝑟𝑖𝑚𝑒𝑖𝑡= 𝜷𝒈. 𝐶𝑂𝐿𝐶𝑖𝑡× 𝐺𝑟𝑜𝑢𝑝𝑔+ 𝜆𝑖+𝛿𝑡+ 𝜀𝑖𝑡 where, • 𝐺𝑟𝑜𝑢𝑝𝑔includes sub-divided groups of Output Areas, using an ONS classification (link). • 𝜷𝒈allows us to estimate the effect of COLC on crime for each of the groups in the spatial classification. • We do not allow the effect to vary over time too (aggregate measure 2019-2022). Spatial Variation: all calls Note: The map highlights the areas with higher correlation between COLC and all CAD calls • The correlation between COLC and all CAD calls seems bimodal. • The effect of the COLC measure on calls is larger in two spatial clusters: one in central London, and another in outskirts London. • The underlying factors might differ between these two. Spatial Variation: crime calls Note: The map highlights the areas with higher correlation between COLC and crime CAD calls • The effect of the COLC measure on crime calls is largest in central London. • It is also present and distributed across patches of outer areas of London. Spatial Variation: violent crime calls Note: The map highlights the areas with higher correlation between COLC and violent CAD calls • The effect of the COLC measure on violent crime calls is larger in central London. • We also find increase in violent crime calls in some fringes of outer London. Spatial Variation: acquisitive crime calls Note: The map highlights the areas with higher correlation between COLC and violent CAD calls • The effect of the COLC measure on acquisitive crime calls is larger in outer east London. • We also find increase in violent crime calls in some fringes of west London. Spatial Variation: Analysis by Borough All CAD Calls Crime Calls Violent Crime Calls ASB Calls Public Safety Calls Public Attitude Survey Has the COLC changed perceptions? Yes. • Boroughs with higher increases in cost-of-living (Havering, Bexley, Bromley) show: • More worries on their area (crime, knife crime, gun crime, gangs, and ASB). • Lower beliefs the police in their area are doing a good job or listen to concerns of local people. • Less trust in the Metropolitan Police Service. Predictions Predicting We now turn to understand how the forecasted cost-of-living in London in 2023 and 2024 will impact police demand. This will assist MOPAC to deploy resources and manage crime. How is that done? 1. We forecast the COLC variable using: 1. OBR’s price forecast for goods and services. 2. ONS’ weights of those goods and services for each household group consumption basket. 2. We predict how much variation in CAD calls would be associated with the forecasted COLC in 2023 and 2024. 1. Point estimates and confidence intervals. 2. Predicted changes by year, semester, and quarter. Main predicted variations 2023 and 2024 • Overall, CAD calls are expected to mildly increase in 2023 with respect to 2022. • Calls related to crimes and violent crimes experience important rises in 2023 with respect to 2022. • In 2024, we predict stability with respect to 2023, in line with a lower expected increase in the Cost-Of-Living measure. D2.8% 0.2% D25.6% D0.6% D27.2% D0.7% D38% D0.8% 100 105 110 115 120 125 130 135 140 2022 2023 2024 All Calls Crime Calls Violent Calls Acquisitive Spatial predicted variations - 2024 COLC is to mostly affect police demand for violent crimes in the outer east. ONS classifies most affected areas as “hard pressed living” migration and churn: middle-aged households with high unemployment and mixed ethnic composition. All calls Violent crime calls Acquisitive crime calls Visualization: interactive maps Knowledge of differential impacts of the Cost-Of-Living Crisis across space and call categories will ensure best evidence is fed into MOPAC to support the office in future decision making on resource allocation. Visualization: interactive maps 1. Cost of Living 2. All Calls 3. Crime Calls 4. Violent Crime Calls 5. Acquisitive Crime Calls Cost of living research and crime in London Magdalena Dominguez and Tom Kirchmaier February 2024 1 London Victim and Witness Service: An update on key learning Sophie Geoghegan-Fittall, Valerie Forrester & Ashley Herron MOPAC Evidence and Insight 2 Executive Summary The London Victim and Witness Service (LVWS) commissioned by the Mayor’s Office for Policing And Crime (MOPAC), went live in April 2019. The service provides support for London residents affected by or witness to crime. The Evidence and Insight (E&I) Unit were commissioned to undertake an evaluation of the service. The two-year evaluation examines two distinct areas: monitoring of the performance of the service through the routine capture of core project metrics; and, generating in-depth understanding of the processes - from design through implementation – of those working to deliver the service. Year one of the evaluation was designed to explore implementation of the service and early service delivery. Limitations, as a result of the global COVID-19 pandemic and delays encountered in information sharing, somewhat limit insight generation and should be noted when reviewing the report. Performance • LVWS is a high-volume service receiving on average 13,900 referrals per month to the victim service and providing support to 3,200 victims each month on average. • Providing support to victims of Domestic Abuse (DA) is a substantial component of service delivery; although, domestic-abuse related referrals accounted for less than a quarter (23%) of total referrals into the service, 44% of DA victims referred to the service went on to take-up the service. • The take-up rate was higher for referrals deemed higher risk or having a greater level of need (high risk DA and enhanced priority referrals). • In Year one, triage IVAs supported the majority of service users (79%, n= 30,299), providing support to an average of approximately 2,500 cases per month. Further exploration of case allocation and service user journey data would enable understanding of service delivery ‘pinch’ points. • In Year one, 3,785 referrals were made to the witness pre-trial and outreach service, with 56% (n= 2,121) of referrals taking-up the offer of support. In order to accurately reflect the operational activity in this area, changes to the recording of pre-trial visits - and other aspects of the witness service – are afoot. • MOPAC User Satisfaction Survey1 (USS) data from quarter one 2020/21, indicates that less than half (44%) of victims who were surveyed reported being offered the services of LVWS and of those, 18% took up the service offered. Of those victims who were supported by LVWS, 83% were satisfied with the service provided2. Amongst those who reported that they were not offered a referral to LVWS, 30% felt they may have benefitted from this service, which points towards further training requirements and awareness raising about the service, and how it can support victims. 1 The USS is a telephone survey asking 9,600 victims each year about their experience of reporting a single crime incident to the police. Questions cover initial contact, the response they got and how they were treated by police. Victims of Residential Burglary; Assault; Personal Robbery and Hate Crime are interviewed 6-12 weeks after the report of their incident. Exclusion criteria: Under 18; Domestic Violence; Sexual offences and Police Officers assaulted on duty. 2 Of those contacted by LVWS 3 • Data availability and quality issues in relation to service user demographic information should be addressed to enable a fuller understanding of service access and engagement. • To develop our understanding of how the service works, data relating to case closures, service user journeys, service user needs and outcomes is required. Implementation • Staff feedback suggests that early implementation of the service presented challenges, particularly in relation to the required mobilisation timescales, and staff resignations. • Key learning can be taken for future MOPAC commissioned services, to ensure sufficient time for planning and preparation prior to service delivery. • Several factors were identified which may be contributing to high caseloads and resource pressures reported by staff: staff turnover and reduced volunteer numbers; the new service requiring staff to ‘juggle’ different work types which may lead to ineffective ways of working; gaps in training particularly in relation to new roles; and, the Case Management System - which is particularly cumbersome - may lead to inefficiencies. • Despite some early implementation challenges, staff feedback suggests that the principles of a Pan-London, integrated service work well, and staff were broadly positive about the service user experience. • Staff consider that delivery partners bring useful specialisms but partnership working requires further development. • Overall, IDVA co-locations were viewed positively with new co-locations reportedly seeing increasing referral numbers. Most challenges discussed by participants were practical in nature including: space to work; access to computers and Wi-Fi; an environment to conduct sensitive work. Going forward • Staff identified several areas of focus for the service going forward: staff wellbeing and morale; staff turnover and retention; service quality and consistency; and, development of partnerships with delivery partners. • There is key learning to be taken for future MOPAC commissioned services, to ensure sufficient time for planning and preparation prior to service delivery. • Workforce capacity and resourcing issues need to be addressed to ensure sustainable ways of working for those delivering the service. • Exploring the use of volunteers in the service should be considered, as well as reviewing the roles and tasks performed by LVWS staff, to ensure efficient and effective ways of working. • As LVWS progresses into Year two of delivery, opportunities to improve data recording practices and systems should be explored to develop understanding of how the service works. This should include increasing the amount and quality of data captured in relation to take-up of service and case closure. • A key area of focus for Year two is the development and redesign of the service user satisfaction survey. Work is underway to explore viability of an online survey, where appropriate, to yield more robust and reliable results and enhance our understanding of service user experience. 4 • The ability to evaluate Year two of LVWS will depend on the quality and quantity of data available. One area that requires focus is an improvement in data sharing between Victim Support and MOPAC E&I – this will create opportunities to understand the service user experience more fully; how the service works and for which groups. Without this data the evaluation will be limited. 5 Introduction The London Victim and Witness Service (LVWS) commissioned by the Mayor’s Office for Policing And Crime (MOPAC), is a support provision for London residents, who are affected by or witness to crime. The new service, which went live in April 2019, is delivered through a number of specialist partnerships, led by Victim Support. A key aspect of the new service is the provision of end-to-end, dedicated and specialist support throughout victims and witnesses’ criminal justice journey, and their journey through cope and recovery. The new service brings together five key areas of support for victims and witnesses through a single, integrated service which: provides support to adult (18+) victims of crime; provides specialist support for victims and survivors of domestic abuse (aged 16+); provides access to restorative justice; delivers pre-trial and outreach support (PTOS) for prosecution and defence witnesses; and provides support for people affected by major crime incidents. LVWS integrates several service elements previously delivered by other MOPAC commissioned services to deliver a single offer of support to victims and witnesses in London: • the Multi-Crime Service (including the assessment and referral service and the support service for victims of crime), previously provided by Victim Support; • Restore:London, providing restorative justice services in London; and • the Pan London Domestic Violence Service, previously provided by Victim Support. Key drivers for the implementation of the new service were to improve the experience of victims and witnesses when they are engaged in the criminal justice system, and, ensure the support they are provided enables them to cope and recover from their experience. LVWS works closely with criminal justice partners to increase compliance with the Victim’s Code of Practice to ensure victims and witnesses are informed of their rights and provided with key entitlements. Led by Victim Support, the LVWS comprises a number of specialist partnerships including: • CALM Mediation Service provide Restorative Justice services; • GALOP provide specialist support and expertise to the LVWS around LGBT+ service users; • Sistah Space provide specialist support to victims of African Heritage and expertise to the wider LVWS partnership; • Stay Safe East provide specialist support and expertise to victims of crime who have disabilities; • St Giles Trust bring their extensive experience of support for those who are both victims and perpetrators, particularly those who are gang affiliated or at risk of joining gangs; • Shelter are commissioned to provide specialist advice and support around housing and welfare benefits. The service offers increased accessibility through longer opening hours and is supported by a dedicated in-bound call service and 24/7 support line. There are various gateways for service access, primarily through referral from the police after reporting crime, but also via self and other agency referrals. 6 Specialist support staff - Independent Victim Advocates (IVA’s) and Independent Domestic Violence Advisors (IDVA’s) - provide a needs and risk led support service to victims and witnesses and, at the point of referral into the new service, are the principal point of contact for service users to minimise the amount of contact they have with criminal justice partners, as well as reducing the need for them to re-tell their experience. As a result of partial-devolution of the national Witness Service, LVWS also provides a pre-trial and outreach service to witnesses across London (previously delivered by Citizen’s Advice), with Citizen’s Advice retaining their role of supporting witnesses on the day of their court attendance. This element of LVWS provides support ahead of a trial to witnesses to help them to feel confident in attending court and supported to give their best evidence. Post-trial, LVWS offers support in relation to verdict and sentencing outcomes, information about engagement with other services i.e. Victim Contact Scheme. In some cases, support may be provided by LVWS on the day if a witness is vulnerable, intimidated or has high level support needs. See Appendix A for the LVWS operating model. 7 Methodology The Evidence and Insight (E&I) Unit - MOPAC’s in-house social research and analytical team - were commissioned to undertake an evaluation of the service. The two-year evaluation examines two distinct areas: monitoring of the performance of the service through the routine capture of core project metrics; and, generating in-depth understanding of the processes - from design through implementation – of those working to deliver the service. Year one of the evaluation was designed to explore implementation of the service and early service delivery. The UK Government introduced ‘lockdown’ on Monday 23rd March 2020 in response to the COVID- 19 pandemic. Whilst delivery of LVWS has continued, there has been both an impact on staff delivering frontline services, as well as the evaluation itself – with the amount of fieldwork within the evaluation considerably reduced. In addition, access to detailed service data held by Victim Support has not been available to support the evaluation due to delays in finalising the necessary Information Sharing Agreements. These limitations somewhat limit insight generation and should be noted when reviewing the report. The current evaluation draws upon fieldwork undertaken between January and March 2020, this includes: • Eleven semi-structured interviews with staff delivering LVWS (referred to throughout as participants). • An online survey to stakeholders designed by E&I distributed in January 2020 to staff who deliver LVWS and work for Victim Support. There was a total of 66 respondents3. • An online survey designed by E&I distributed in January 2020 to delivery partners who work as part of the LVWS consortium. There was a total of 6 respondents4. Moving forward into Year two of the evaluation, fieldwork will continue to support the process-based learning around service delivery, as well as analysis of routinely collected performance data to understand how the service is running on a day-to-day basis. Opportunities to explore impact will also be considered in Year two. The ability to successfully complete each element will depend on the quality and quantity of data. One area that requires focus is an improvement in data sharing between Victim Support and MOPAC E&I – this will create opportunities to understand the service user experience more fully; how the service works, for which groups, and in what ways. Without this data the evaluation will be limited. 3 Approximately 145 Victim Support staff deliver LVWS. 4 Approximately 11 staff work within delivery partner organisations to deliver LVWS. 8 Results Performance learning This section reports on Year one service data from 1st April 2019 to 31st March 2020 collected by Victim Support, and data from MOPAC’s User Satisfaction Survey (USS) for quarter one 2020/21. It comprises three sections: the first section presents a brief overview of data in relation to the victim element of the service including an overview of referrals and take-up of service; the second section outlines service activity in relation to the witness element of the service; and, the third section presents findings from MOPAC’s USS. Section 1: Victim Service Service demand Between April 2019, when LVWS launched, and the end of March 2020 there were a total of 167,030 referrals to the service, an average of around 13,900 referrals per month. The majority of referrals come from the police5 (90%, n= 151,008), followed by self-referrals (8%, n= 12,442), and lastly ‘other agencies’ (2%, n=3,580). From October 2019 to March 2020, there was an increase in self-referrals; the number received each month doubled compared to the first six months of the service. This may be as a result of changes to recording practices, further work to understand this increase should be explored. Overall, 23% (n= 38,237) of referrals resulted in take-up of service, which works out as an average of around 3,200 per month. Table one shows the total referrals received by LVWS and the take-up of service rates for each referral source: police; self-referral; and, other agencies. Referrals from the police have the lowest take-up of service rate, though cases referred from the police account for 70% of support provided by the service. Both total referrals and take-up of service rate have fluctuated during the twelve months (see Figure 1). Further work to explore police referral mechanisms and the take-up rate should be considered. Table 1. Referrals and take-up of service by referral source Referral Source Total referrals Take-up of service cases Take-up of service rate Proportion of support provided by LVWS Police 151,008 26,583 18% 70% Self-referral 12,442 9,540 77% 25% Other agencies 3,580 2,114 59% 6% 5 Action Fraud, British Transport Police, City of London Police, Metropolitan Police Service, Other Police Force Areas. 9 Figure One. Total referrals and Take-up of service rate per month Metropolitan Police Service referrals In Year one, the majority (74%, n= 123,735) of referrals came from the Metropolitan Police Service (MPS). Between May 2019 and September 2019, there was a steady decline in the number of referrals made to LVWS; this has since recovered although monthly fluctuations in the number of referrals can be seen in Figure one. In Year One, 19% (n= 22,858) of victims referred by the MPS went on to take- up the service. The take-up of service rate does not appear to relate to the number of referrals from the MPS as you might expect, with lower rates seen in October through January compared to higher referral levels. More detailed data relating to cases referred by the MPS should be explored. Figure Two: MPS referrals and Take-up of service rate per month 10 Non-Domestic Abuse (DA) related referrals During the reporting period, 128,098 referrals related to non-domestic abuse related crime were made to the LVWS, with 16% (n=20,988) of victims taking-up the service offered. The highest number of referrals received related to ‘other theft’ (n= 28,820), followed by: ‘fraud and forgery’ (n=22,809); and, ‘violence with injury’ (n=18,914). See Appendix B, Table 2 for all crime categories. The rate of take-up of service was highest for referrals related to ‘non-crime’ (85%) – further work should be done to understand what this crime category relates to – this was followed by ‘Homicide’ (74%); ‘Rape’ (59%); and, ‘other sexual offences’ (51%). Non-DA related referrals are categorised as Enhanced Priority Referrals (EPR) or non-EPR which is reflective of the level of need or risk associated with the case. The number of EPR and non-EPR referrals into the service was broadly equal; there were 63,454 (49.5%) EPR referrals compared to 64,644 (50.5%) non-EPR referrals during Year one. The take-up of service rate for Year one was higher for EPR referrals than non-EPR referrals; 21% (13,218) compared to 12% (7,770) indicative of greater attrition in relation to non-EPR referrals. Domestic Abuse (DA) referrals During the reporting period, 38,932 referrals were made to the LVWS which related to DA, accounting for 23% of all referrals into the victim element of the service. Of those referrals, 14% (n=5,460) were considered ‘high risk’6, with 86% (n=33,472) considered ‘low/medium’ risk. Using MPS recorded domestic violence offence and incident levels, and referral borough data captured by the LVWS, it is possible to explore whether referral levels to LVWS correspond with MPS recorded DV across London7. According to MPS data, Croydon recorded 6% of DV offences and 6% of incidents; the highest levels across London. LVWS received 1,503 DA related referrals from Croydon, accounting for 4% of the total referrals. A higher proportion of recorded offences compared to the relative proportion of referrals to LVWS was also seen in the boroughs of Greenwich, Barnet and Barking & Dagenham; although the difference is small (see Appendix B, Table 3). Overall, the take-up of service rate for DA referrals was 44% (n=17,249); 41% of service users referred in by the police went on to take-up a service (n=13,495), and 76% of those who self-referred into LVWS went on to take-up a service (n=2,006). This is a higher take-up rate than for non-DA crime. The take- up rate for referrals considered ‘high risk’ was higher than those referrals assessed as ‘low/medium risk’; 58% (n=3,193) compared to 42% (n=14,956). Service user demographics There are data availability issues in relation to user demographics and characteristics. Data relating to religion and sexual orientation have a high level of missing data. Ethnicity data is only available for referrals which resulted in a take-up of service - only 25% of cases have the service user’s ethnicity recorded in these cases. 6 would be referred to an IDVA for support. 7 MPS recorded data is reflective of those offences and incidents which come to the attention of the police. 11 Age data in relation to non-DV related crime (see Appendix B, Table 4) suggests that the take-up of service rate increases with age; 18% (n=2,110) of victims aged 65 and over were provided support, compared to 11% (n=1,930) of 18 to 24 year-olds and 11% (n=3,483) of 25 to 34 year-olds8. Further exploration of age data in relation to crime categories, support provision, case closure and outcomes, would enable a greater understanding of how different groups access and engage with the service. Improved recording of service user demographics would provide valuable insight into which groups are accessing the service and how they engage with the support offered. It is recommended that LVWS review data collection and recording practices in relation to service user information. Support provision Following referral into the service, cases are allocated to either a triage Independent Victim Advocate (IVA), complex IVA or Independent Domestic Violence Advocate (IDVA), to provide ‘immediate’ or ‘ongoing’ support depending on the risk level and needs of the victim9. In Year one, LVWS supported a total of 38,237 victims. Data indicates that in Year one, triage IVAs supported the majority of service users (79%, n= 30,299), providing support to an average of approximately 2,500 cases per month; followed by complex IVAs (12%, n= 4,746) and IDVAs (8%, n= 3,192). Unsurprisingly, DA related service users received ‘ongoing’ support more commonly than non-DA related service users (see figure three). Non-EPR cases most commonly received immediate support (76% of non-EPR cases, n= 5,936) followed by EPR cases (62% of EPR cases, n= 8,139) (see Figure three). Information regarding the nature of immediate and ongoing support, and how cases progress through to closure, would be useful in building understanding of how the service works. Section 2: Witness pre-trial and outreach service In Year one, 3,785 referrals were made to the pre-trial and outreach service, with 56% (n= 2,121) of referrals taking-up the offer of support. The majority (88%, n= 3,325) of referrals came from the MPS Witness Care Units (WCUs) - the expected main source of referrals for this service. The number of referrals from WCUs steadily increased each quarter, as did referrals from other agencies into the service. There may be several factors contributing to this; increased awareness amongst staff working 8 This excludes under 18 victims which data suggests have a 50% take-up of service and those cases in which age is not given (10% of referrals and 27% of take-up of service cases are missing age data). 9 Following referral, service users may be allocated a triage IVA initially, and subsequently reallocated to a complex IVA or IDVA as more information comes to light. Figure Three: Immediate vs. Ongoing support provision by Case type 12 within referral agencies as a result of promotional activity and training, improved referral processes and mechanisms, or improved data recording practices. Victim Support offer witnesses the opportunity of a visit to court ahead of their attendance at the trial (Pre-trial visits); these are intended to help them to feel more confident and comfortable on the day of the trial. A total of 2,109 pre-trial visits were requested and, of these, one third (n= 781) were undertaken during Year One10. In order to accurately reflect the operational activity in this area, changes to the recording of Pre-trial visits - and other aspects of the witness service – are afoot. Section 3: MOPAC’s User Satisfaction Survey (USS) The USS11 now includes a set of questions around provision of support services, specifically LVWS, which can provide useful insight in relation to victims of some crime types12. Data from quarter one 2020/21, indicates that less than half (44%) of victims who were surveyed were offered the services of LVWS and of those, 18% took up the service offered. Of those victims who were supported by LVWS, 83% were satisfied with the service provided13. Amongst those not offered LVWS, 30% felt they may have benefitted from this service, which points towards further training requirements and awareness raising about the service. A broader range of questions around compliance with the Code of Practice for Victims of Crime indicate that: less than one third of respondents were made aware of the Code of Practice for Victims of Crime (31%); half of respondents indicated that they were given the opportunity to provide a victim personal statement; and, less than 1 in 5 (19%) of respondents reported being given information on Restorative Justice. This highlights the important role that LVWS, and other victim support services, have in providing information to victims about their rights under VCOP. Summary of performance learning • LVWS is a high-volume service receiving on average 13,900 referrals per month to the victim service and providing support to on average 3,200 victims each month. • Providing support to DA victims is substantial component of service delivery; although, Domestic- violence related referrals accounted for less than a quarter (23%) of total referrals into the service, 44% of DA victims referred to the service went on to take-up the service. • The take-up rate was higher for referrals deemed higher risk or having a greater level of need (high risk DA and enhanced priority referrals). • In Year one, triage IVAs supported the majority of service users (79%, n= 30,299), providing support to an average of approximately 2,500 cases per month. Further exploration of case 10 Due to court processes there is likely a substantial delay between initial referral and a trial date with many PTVs likely to spread across a number of Quarters. 11 The USS is a telephone survey asking 9,600 victims each year about their experience of reporting a single crime incident to the police. Questions cover initial contact, the response they got and how they were treated by police. Victims of Residential Burglary; Assault; Personal Robbery and Hate Crime are interviewed 6-12 weeks after the report of their incident. Exclusion criteria: Under 18; Domestic Violence; Sexual offences and Police Officers assaulted on duty. 12 burglary, robbery, vehicle, assault and hate crime. 13 Of those contacted by LVWS 13 allocation data and service user journey would enable understanding of service delivery ‘pinch’ points. • In Year one, 3,785 referrals were made to the pre-trial and outreach service, with 2,121 (56%) referrals taking-up the offer of support. In order to accurately reflect the operational activity in this area, changes to the recording of pre-trial visits - and other aspects of the witness service – are afoot. • USS data from quarter one 2020/21, indicates that less than half (44%) of victims who were surveyed were offered the services of LVWS and of those, 18% took up the service offered. Of those victims who were supported by LVWS, 83% were satisfied with the service provided14. Amongst those not offered LVWS, 30% felt they may have benefitted from this service, which points towards further training requirements and awareness raising about the service. • LVWS, and other victim support services have a key role in supporting victims to understand their rights under VCOP. • Data availability and quality issues in relation to service user demographic information should be addressed to enable a fuller understanding of service access and engagement. • To develop our understanding of how the service works, data relating to case closures, service user journey, service user needs and outcomes is required. Process learning This section explores learning from interviews carried out with LVWS staff and surveys with staff delivering LVWS, with the aim of presenting learning related to implementation and early delivery of the service15. Six areas emerged from the data: • Implementing the service • Service demand and workforce capacity • Training, skills and tools to support delivery • Service delivery model • Working with others • Looking ahead Early implementation of the service Timescales associated with commissioning LVWS, meant that the service was required to mobilise more quickly than anticipated. Almost all participants recognised the speed at which the service became operational following award of the contract which, as one participant described, resulted in a large-scale organisational change for Victim Support and delivery partners: ‘If this contract had been secured by another provider, there’d have been a lead in time, we’d have continued to deliver for half a year or however long, but we didn’t have any of that and we went from a full caseload of clients to then a brand new set of work and we inherited a backlog in terms of the pre-trial and outreach service, so I think it has to be acknowledged that we were on the backfoot by us winning the contract’. 14 Of those contacted by LVWS 15 The number of interviews conducted with staff was significantly affected due to COVID-19. This should be borne in mind when reviewing these findings. 14 Whilst some participants felt that they had been well briefed on the transition to the new service, all agreed that in the early stages of delivering the service they were learning ‘on the job’. One participant described how they felt unprepared for the practical aspects of implementing the new service: ‘I got the sense that several positions higher were unprepared for the practical implementation, they might have had knowledge about what it was going to be, but actually how were our day to days going to work, so there was a kind of a fumbling feeling which was quite distressing to work as'. Indeed, ‘communication’ throughout transition and implementation of the service was highlighted by staff as a key challenge. Despite the commendable efforts of LVWS staff to mobilise quickly, accelerated roll out has been linked to early teething problems (e.g., IT, staffing, resources, training, governance) in other programmes and services. Building in adequate time for mobilisation is a central tenet of delivery and implementation – although does not always align with the cultures in which projects roll out16. Key learning can be taken for future MOPAC commissioned services, to ensure sufficient time for planning and preparation prior to service delivery. Service demand and workforce capacity LVWS is a high-volume service receiving on average 14,200 referrals into the service each month17 (see section one). Participants working directly with service users reported heavy caseloads and survey respondents also indicated resource pressures; less than half (n=29/63) agreed that their workload is manageable. Furthermore, only 23/63 respondents agreed that their team has the resources to provide a timely and quality service to victims and witnesses - with a greater proportion of Inbound Support Agents18 disagreeing that their team is adequately resourced. For some participants, their workload was ‘manageable’ but only just, and potentially only temporarily, due to newer roles receiving lower referrals. As one participant described: ‘I think we’re stretched resource wise… people are doing much more to try and maintain the system than was previously. I don’t know how much of a longevity that has for workers, it feels like, you can work people as hard as you can but they buckle and break… there’s always going to be a gap between how much is there to work with and try and make better, and the resource. However, I feel for the sustainability of the people working it’s quite hard at the minute’. Some participants, particularly caseworkers, raised concerns about the impact of increasing demand on service quality, expressing wanting ‘more space and time to offer that quality’ with some fearful that aiming for more referrals and quicker contact times may impact on staff ability to provide quality support. As a result of the demands placed on staff delivering LVWS, some staff highlighted potential effects on staff satisfaction and wellbeing, and in turn staff turnover. Staff support was highlighted by a number of participants as an area requiring greater attention: ‘[staff] can’t do a good job without being looked after ourselves’. Whilst overall, those spoken to felt that their managers were supportive, 16 Dawson, P. and Stanko, B. (2013) ‘Implementation, Implementation, Implementation: Insights from Offender Management Evaluations’ Policing: A Journal of Policy and Practice 7 (3): 289-298. 17 Both victim and witness services. 18 Inbound Support Agents operate from Cardiff and answer inbound calls from self-referrals and other agencies who contact the service directly. 15 it was evident that some participants saw limitations in how much managers are able to do, as one participant described: ‘I think with managers I just personally think naturally they don’t see or feel it because they don’t have caseloads and I know their role is completely different to ours. And to me I’m just getting the impression well there’s nothing that can be done about it so you just have to get on with it in that sense, and it is what it is.’ According to several participants, a significant number of staff resigned during the implementation phase of LVWS which impacted early implementation and delivery of the service, as one participant described: ‘when we first started what we were dealing with was people resigning on mass and then trying to cope with huge staff vacancies’. Subsequently, time was spent recruiting and training new staff to deliver the service. It was evident that staff may still perceive turnover as an issue for some areas of the service: ‘we had a big group of people leaving after the LVWS model came into play but even now when it’s a whole new group of people, we still have a lot of turnover’. In addition, the decision to professionalise the service resulted in a reduction in the number of volunteers, particularly in roles working directly with clients. Several participants reflected on how this change has contributed to workload pressures, one participant summarised: ‘there was some level of control with the load of the work, because you could always go either and recruit new volunteers…at least we had some resources available to us when the caseload got overwhelming, there was some avenues to solve that’. Several participants, particularly those working directly with service users, indicated that staff are required to ‘juggle’ different aspects of work - administrative tasks, providing emotional support and arranging and attending Pre-trial visits (PTVs)19 – which may affect their ability to work efficiently: ‘On a practical level…I found it quite hard in that there is a lot of admin that goes into arranging pre-trial visits, that is fiddly and basically it takes up a lot of time for quite a short thing to achieve, whereas and it’s a very different type of, that admin stuff is very different to the more support role, emotional work you do with clients, the casework’. As such, some participants felt that the witness element of the service could be considered distinct from the case work element of the service with additional staff recruited to perform the role of organising and undertaking pre-trial visits, which may help to alleviate workload pressures. As one participant considered: 'from a work point of view that’s a lot of work from us that we’re doing that is not making the best use of our skills and gets in the way of delivering other cases'. Such variation to service delivery would run counter to the principle of delivering a ‘single point of contact’ model, highlighted as key during service user consultation which informed design of the current model. Training, skills and tools to support delivery Overall, survey respondents reported that they have the training, skills and knowledge to perform their roles. Participants described a comprehensive training package for delivering PTVs and some useful training provided by delivery partners, although this was reportedly somewhat inconsistent. Gaps in training and skills were evident, particularly for some staff groups. A greater proportion of Inbound Support Agents who responded to the survey, did not agree that they had received adequate 19 A visit to the court undertaken by LVWS with a witness ahead of a trial. 16 training and information to carry out their role. Indeed, several participants observed an issue with the quality of information originating from the Inbound Team, pointing towards an area for improvement. Caseworkers reported a lack of training particularly in relation to the practical processes associated with the new structure, and aspects of their new roles, particularly those who were now expected to provide emotional support to service users. As one participant described: ‘I would have liked more training around the emotional support aspects of it just because it was a new territory for us…which meant that we would talk to them once and then go… the waters can get very murky in regards what is emotional support and then what goes into therapy or counselling…so I would have liked more training around that and how we can structure our conversations just so we’re not going into that territory’. Staff respondents reported general satisfaction with the assessment tools available to carry out their role; 48/63 respondents agreed that needs assessment and support planning tools are effective, and 55/63 agreed that risk assessment tools are effective. Participants also considered the tools available to support them in their role were generally effective. However, respondents generally reported lower satisfaction with the systems used for recording their work. Whilst some participants reported that the Case Management System (CMS) usefully holds the necessary information about a case in one place, the majority viewed the system negatively; 'the worst. It’s very slow at times, it’s very long winded just to even put a case on', contributing to inefficiencies across the service. It was evident from staff feedback that the CMS has undergone multiple alterations to reflect the new service delivered which is sometimes challenging for staff to keep updated on. Service delivery model A Pan-London, integrated service LVWS offers service users the choice of where and how they receive support as well as enabling flexibility within the service to cope with fluctuations in demand. The Pan-London aspect of the model was praised by participants – particularly those working with DA cases - as summarised by one participant: 'obviously people don’t exist in a borough way, people live in a Pan-London way and so actually it’s more reflective of people’s lives, and that’s been a reduction of pressure on the other services, they’ve been like oh great if we get someone that isn’t from one of our boroughs you can pick it up'. Although it has enabled the service to be able to cope with increased demand in certain areas, it was evident that more use could be made of this aspect of the model, and the required ‘mindset change’ was still a work in progress. LVWS aims to offer an integrated system of support to service users through integration of the domestic abuse (DA) element and core service - previously two services20 - to form ‘one larger breathing organism’, and through working in partnership with five delivery partners21. Participants spoke positively about the integration between the core and DA teams with much closer working as a result of delays to vetting and honorary contracts. It was reported by one participant that there had been a natural upskilling of staff, with non-DA staff benefiting from having their MARAC referrals quality assured by an Independent Domestic Violence Advocate (IDVA), and DA staff learning about 20 VS Multi-Crime Service and the Pan-London DV Service. 21 Sistah Space; Shelter; Stay Safe East; Galop; and, Calm. 17 some of the practical elements of support, i.e., vouchers, security items, and things that need to be done immediately for a service user. Some participants welcomed further integration of the teams which they felt would help to bring consistency to service delivery. Supporting service users Survey respondents indicated a relatively high level of agreement that: LVWS provides a high-quality service (51/61); delivers improved health and wellbeing outcomes (50/61); provides an improved experience (47/61); and provides improved equality of access (50/61). Overall, staff respondents reported feeling confident in identifying the needs of victims and witnesses and providing the appropriate support, with 51/66 strongly agreeing, and all remaining respondents agreeing. Participants working directly with service users described the ‘fine line’ between providing emotional support vs. counselling and the need to ‘ensure that we’re not going somewhere where we shouldn’t be tapping into’ which requires managing user expectations; setting out the work initially, creating milestones, and utilising an ‘empowerment model’. LVWS aims to minimise the service user touchpoints, making it easier for service users to navigate the system, and reducing the need to retell their experience. In addition, the service aims to allocate any repeat service users to the same staff member. Whilst participants praised the benefits of having one point of contact, it was evident that whilst it was ‘more or less’ happening - 35/63 survey respondents agreed that the single caseworker model works well - it is not always possible due to staff resignations and workloads, and may not be a realistic aim. In addition, some participants felt that the partial devolution22 of the witness service may introduce an unnecessary contact point for the service user. This may point towards the need for further awareness raising amongst staff of the benefits of the new model; although Citizen’s Advice retain their role in providing on-the-day support to witnesses, the new model ensures that service users have someone from LVWS supporting them through every stage of their journey, considered an improved offer to victims in London who also attend court. Survey respondents reported feeling confident in signposting and supporting clients to access additional support outside of LVWS (60/63 agreed), identifying when a client no longer needs support (53/63 agreed), and having conversations about ending support provision (53/63 agreed). A key benefit of the service noted by participants was the absence of a time limit on the support offer, with some service users receiving support for many months particularly in relation to advocacy. Whilst the hope is that all service users will ‘cope and recover’, it was evident from some participants, that there are a group of more complex service users who require greater support than LVWS can provide, particularly those with mental health issues: ‘I think we get a lot of other cases where we are not mental health professionals, so we can’t make any diagnoses over the phone or anything like that, but it seems as the case unfolds it becomes very clear that the problem is actually a mental health issue not the actual crime’. 22 Citizen’s Advice continue to provide support to witnesses on the day of their trial. 18 Working with others LVWS delivery partners Around half of survey respondents reported having regular contact with LVWS delivery partners. Furthermore, most survey respondents reported a good relationship with those partners with whom they have more regular contact, with very few respondents reporting a poor relationship with any partner. Participants were particularly positive about the relationship between VS and Shelter, considered a good example of partnership working and the ambition for all partnerships. The majority of respondents from delivery partner organisations, indicated a good working relationship with Victim Support. Whilst some participants were positive about the contribution of partners to the service and how they increase access points to support for those service users who would not go to the police, it was evident that more work is required to fully implement the partnerships into practice. Overall participants reported minimal contact with partners, some staff had received a training or awareness session, some were aware of partners through networking events or co-locations, pointing towards pockets of partnership working as opposed to a fully integrated partnership. One participant reflected: ‘I think the partners in general, more work needs to be done that to make it feel a little bit more like a partnership’. Introducing advice surgeries, co-location of staff and attendance of partners at area meetings, were reportedly being explored to improve partnership working. IDVA Co-locations Whilst not new, the co-location of IDVAs is a key element of LVWS which has introduced a greater number of co-locations to the model. Overall, co-locations were viewed positively with reportedly increasing referral levels coming from newer co-locations. Co-locations reportedly work well when there is high-level buy-in and the infrastructure to support their implementation. It was clear that each co-location is different and implemented on a case-by-case basis. Some participants described a lack of guidance and materials to support embedding within co-locations, with very little already set- up. It was clear that staff in co-locations must balance the demands of their client facing work, as well as promote and raise awareness of the service through training, ward rounds, meetings and informal routes. Most other challenges associated with co-locations appeared to be practical: having space to sit in predominantly hot desking environments; having somewhere to have confidential conversations with service users; lack of facilities including canteens and Wi-Fi reception; delays to vetting and honorary contracts; and, organising training with busy staff based in co-locations. Looking ahead According to some participants - particularly those in management roles - the service was moving towards ‘business as usual’ with staff levels stabilising following recruitment activity in Year 1. This reportedly allows for more focus on service development and improvement areas for service delivery. Participants identified a number of areas of focus for the service going forwards: • Staff wellbeing and morale: breakout spaces; support for staff; health and wellbeing champions. • Staff turnover and retention: understanding from staff exit interviews why staff leave; considering how volunteers can be utilised to bring additional resources. 19 • Service quality and consistency: bringing increased consistency to the service through further integration of internal teams and processes; enabling more time for staff to deliver quality; addressing resourcing issues. • Development of partnerships with delivery partners –embedding partners fully into the service; learning from the success of the partnership with Shelter. 20 Conclusion The focus of this report has been to highlight findings and issues relating to the performance and processes of the first twelve months of the London Victim and Witness Service implementation and delivery. Data for this report was gathered from several sources, including data from Victim Support’s case management system, interviews with staff, and a survey with staff delivering the service. LVWS has had a busy first year, mobilising quickly to integrate several services into a single support offer for victims and witnesses across London. The service has received 170,815 referrals (167,030 victim referrals and 3,785 witness referrals) and provided support to 40,328 service users (38,237 victims and 2,091 witnesses) in the first 12 months of delivery. Performance will continue to be monitored throughout year two, with particular focus on the service user journey through the service, understanding who is using the service, and outcomes for service users. It was evident from staff that, the principles of a Pan-London, integrated service appear to work well. LVWS aims to make it easier for service users to navigate the CJS and cope and recover following the effects of crime through a single support offer to victims and witnesses across London; staff spoken to consider the LVWS to provide an improved service user experience. Delivery partners appear to bring useful specialisms to the service - supporting victims who may not otherwise access support – though further work is required to fully embed these partnerships. Challenges during early implementation of the service were evident. Staff feedback suggests that mobilisation timescales led to retracted preparation time ahead of delivery, further exacerbated by staff resignations. There is key learning to be taken for future MOPAC commissioned services, to ensure sufficient time for planning and preparation prior to service delivery. Workforce capacity and resourcing issues should be addressed to ensure sustainable ways of working for those delivering the service. Exploring the use of volunteers in the service should be considered, as well as reviewing the roles and tasks performed by LVWS staff, to ensure efficient and effective ways of working. As LVWS progresses into Year two of delivery, opportunities to improve data recording practices and systems should be explored to develop understanding of how the service works. This should include increasing the amount and quality of data captured in relation to take-up of service and case closure. In addition, a key area of focus for Year two is the development and redesign of the service user satisfaction survey. Work is underway to explore viability of an online survey, where appropriate, to yield more robust and reliable results and enhance understanding of service user experience. Evidence and Insight researchers will continue to monitor performance data - seeking opportunities to obtain quality data – and re-visit staff to hear their views on service delivery. In addition, opportunities to understand the impact of the service, as well as gathering the views of service users, will be explored in Year two. The ability to successfully complete each element will depend on the quality and quantity of data, without which the evaluation will be limited. 21 22 Appendix A: LVWS operating model NB. This model has altered during implementation. All high-risk DA that comes via triage does not go to an IDVA as the diagram suggests. IDVAs are now all based in their co-locations and most of their referrals come from these. As a result, some high-risk DA that comes via triage will go to IDVAs, the remaining high-risk DA cases are referred to the appropriate borough service. 23 Appendix B: Performance Data Table 2. Referrals and Take-up of service by Crime category Crime Category Referrals Take-up of service Take-up rate Non-Crime 1187 1011 85% Homicide 284 211 74% Rape 535 314 59% Other Sexual Offences 1355 696 51% Other Crime 15741 4205 27% Violence without Injury 10521 2306 22% Violence with Injury 18914 4077 22% Arson 246 41 17% Fraud & Forgery 22809 2915 13% Burglary - Dwelling 12338 1509 12% Criminal Damage 6904 742 11% Theft from the Person 6394 520 8% Other Theft 28820 2328 8% Burglary - Non-Dwelling 1205 79 7% Bicycle Theft 845 34 4% Total non-DA Crime 128098 20988 16% High risk DA 5460 3193 58% Low/medium risk DA 33472 14056 42% TOTAL DA related Crime 38932 17249 44% TOTAL All Crime 167030 38237 23% 24 Table 3. Domestic Violence related referrals and offence and incident volumes by borough Borough Number of referrals Proportion of total referrals DV offences volume* Proportion of total DV offences in London DV Incidents* volume Proportion of total DV incidents in London Enfield 1627 4% 3869 4% 6207 4% Tower Hamlets 1621 4% 3248 4% 5782 4% Lewisham 1603 4% 3571 4% 5671 4% Hounslow 1551 4% 3238 4% 5403 4% Newham 1549 4% 3559 4% 5876 4% Croydon 1503 4% 4812 6% 7980 6% Greenwich 1489 4% 3940 5% 6122 4% Southwark 1488 4% 3032 3% 4819 3% Ealing 1427 4% 3449 4% 6203 4% Brent 1379 4% 3425 4% 5475 4% Bromley 1373 4% 3042 3% 4783 3% Lambeth 1328 4% 3068 4% 4861 3% Hillingdon 1292 4% 2775 3% 4901 3% Barnet 1276 3% 3123 4% 5034 3% Barking and Dagenham 1223 3% 3190 4% 5396 4% Wandsworth 1163 3% 2262 3% 3938 3% Hackney 1133 3% 2866 3% 4873 3% Waltham Forest 1124 3% 2466 3% 4241 3% Havering 1108 3% 2581 3% 4588 3% Bexley 1073 3% 2573 3% 3892 3% Redbridge 1068 3% 2319 3% 4136 3% Haringey 1042 3% 3028 3% 5030 3% Islington 964 3% 2305 3% 3793 3% Sutton 964 3% 1830 2% 2860 2% Merton 820 2% 1727 2% 2778 2% Harrow 806 2% 2036 2% 3332 2% Hammersmith and Fulham 801 2% 1834 2% 3284 2% Camden 762 2% 2195 3% 3580 2% Westminster 678 2% 2083 2% 3442 2% Kensington and Chelsea 577 2% 1274 1% 2098 1% Kingston upon Thames 574 2% 1373 2% 2196 2% Richmond upon Thames 497 1% 1163 1% 1900 1% * SOURCE: MOPAC crime data dashboard, CRIS data- rolling 12-month data to June 2020 25 Table 4. Referrals and take-up of service rate by Age of service user Age Referrals Take-up of service cases Take-up rate 17 and under 827 411 50% 18 to 24 17414 1930 11% 25 to 34 30424 3483 11% 35 to 44 24511 3152 13% 45 to 54 18539 2628 14% 55 to 64 11622 1623 14% 65 and over 11623 2110 18% Not Given 13138 5651 43% TOTAL 128098 20988 16% Version 1.0 28th July 2017 Page 1 of 3 Tackling disproportionality in the police officer misconduct process Action plan Intervention Description Action to be taken Timescales Lead 1 – Safe space for supervisors Creating a 'safe space' for supervisors to seek support and advice (anonymously if required) prior to considering formal misconduct. A HR advice line is in place, originally set up to advice on Grievance procedures. This is to be utilised for misconduct also. A communication plan to promote this service internally launched in February 2017. Conduct a series of briefings to advice line management staff regarding misconduct. Examine whether any data can be captured to show whether the advice line is being used for 'safe space' advice. Continuous throughout 2017 Complete by end Autumn ‘17 Completed Strategic HR Directorate of Professional Standards (DPS) Strategic HR and Strategy & Governance 2 – Additional training and support for: a) Professional Standards Champions (PSCs) b) Line managers at Sergeant / Inspector level c) Development days for Staff Support Associations (SSA) PSCs, line managers and Staff Support Associations will be offered additional training to ensure they are fully equipped to recognise whether concerns raised to them are actually misconduct or whether the matters should be explored through routine line management. Awareness needs to be raised regarding the disproportionality problem to ensure BME officers’ cases are dealt with in the same way as non-BME officers’ cases. 2a) Develop and deliver a discrimination training package to all existing PSCs as part of continuous professional development. 2b) Deliver a specifically tailored Discrimination Training Package as part of all promotion courses 2c) Deliver regular development days to SSAs to help them support staff in regards to disproportionality Started in July 2017 and will continue quarterly Starting in autumn 2017 Started in July 2017 and will continue regularly DPS Met Training Strategy & Governance 3 – Leading for London training Aims to develop officers and staff by building skills, capability and positive behavioural change. By creating a skilled and inclusive workforce, leaders will lead, manage differences LfL will start with senior leaders, cascading down to all officers and staff to level of Sergeant / Starting in Autumn 2017 Strategic HR Version 1.0 28th July 2017 Page 2 of 3 Intervention Description Action to be taken Timescales Lead programme and diversity better, and have the relevant skills to underpin the new ways of working as the Met transitions. ‘Inclusive Practice’ is consistently woven through the programme objectives and learning outcomes for each part of Leading for London. Band D. In scope on a mandatory basis are 10,300 officers and police staff. Constables and staff below Band D level will receive learning via digital resources. 4 – Changes to the MPS MM1 form (used to record internal conduct concerns) Rollout of a revised MM1 form, aimed to encourage line managers and PSCs to consider informal action where appropriate. Revise MM1 form / guidance and seek feedback from MPS Legal Services. Introduce new MM1, with a communication plan. Ongoing Autumn 2017 DPS & Directorate of Legal Services (DLS) 5 – Evaluating the impact of actions Future analysis and evaluation of the impact of interventions. a) Analyse MPS misconduct data (calendar years 2016 & 2017) as per original research to evidence any changes. b) Conduct quarterly surveys of PSCs and DPS Investigators to capture feedback on training and key challenges. c) Conduct quarterly post- training surveys with new Sergeants and Inspectors to review Met Training. d) Leading for London has a planned evaluation strategy (Kirkpatrick & Project Management methods) forming three core areas of focus: 1. Learner Experience and Performance – Rating and Satisfaction of programme and Learning Content End of January ‘18 Quarterly starting after August ‘17 Quarterly starting once training is delivered Dependent on training delivery MOPAC Evidence & Insight Strategy & Governance Strategy & Governance Strategic HR Version 1.0 28th July 2017 Page 3 of 3 Intervention Description Action to be taken Timescales Lead 2. Programme Performance – Impact, culture, and behavioural change/shift evidenced as a result of delivery 3. Project Performance – Efficiency, effectiveness, cost, impact and quality. e) Quarterly review of performance regarding discrimination in misconduct, grievances & Employment Tribunals as part of the Discrimination Complaints Project Board. Continuous throughout 2017-18 Strategy & Governance 1 Disproportionality in Misconduct Cases in the Metropolitan Police Service December 2016 Daniela Wunsch, Chloe Hughes, Zoe Hobson and Julia Yesberg MOPAC Evidence and Insight 2 Executive Summary .................................................................................................................. 3 1. Introduction ......................................................................................................................... 4 Aim and objectives ................................................................................................................. 4 2. Method and results ............................................................................................................. 5 Review of the existing evidence .............................................................................................. 5 In-depth analysis of five years’ worth of MPS misconduct data .............................................. 6 Interviews with key staff involved in the assessment and investigation of misconduct .......... 10 3. Discussion ........................................................................................................................... 15 Conclusions ........................................................................................................................... 15 Areas for further exploration ................................................................................................. 15 References .................................................................................................................... 17 3 Executive Summary Evidence & Insight, set within the Mayor’s Office for Policing And Crime (MOPAC), were approached by the Metropolitan Police Service (MPS) to undertake research into misconduct cases. The main objectives were to determine the extent to which ethnic disproportionality featured within MPS Officer misconduct data. The research incorporated a review of the key literature, analysis of misconduct allegations made against MPS officers between 2010 and 2015, and 11 in-depth interviews with key staff involved in the assessment and investigation of misconduct in the MPS. Key Findings:  The research found a number of studies that evidenced the existence of ethnic disproportionality in disciplinary proceedings in other police forces - both in the UK and US - as well as other professions.  BAME officers in the MPS are twice as likely as White officers to be subject to misconduct allegations.  Allegations against BAME police officers are more likely to be substantiated, whilst allegations against White officers are more likely to be unsubstantiated.  This disproportionality is not driven by length of service; age of officer; differences in allegation type between BAME and White officers; or differences in on vs off duty behaviour.  Whilst there was no ethnic disproportionality in the write off method for substantiated cases, in unsubstantiated cases BAME officers were more likely than White officers to still receive management action.  There is no disproportionality gap in the number of public complaints made against BAME and White officers.  Those interviewed for the research were confident current processes were fair and robust, putting forward the point that they only dealt with cases they were given and were not focussing on ethnicity – indeed in most cases would not know the ethnicity of the officer the allegation was against.  Interviewees expressed concerns over an overall reluctance to deal informally with problematic staff behaviour or performance issues.  Whilst research to date has been able to evidence the existence of disproportionality, no studies have so far been able to put forward conclusive evidence as to its causes or ‘what works’ in responding to it.  In terms of potential causes, there are three prominent academic theories: 1) Fear of being accused of racism; 2) Conscious / unconscious bias; and 3) Failure to deal with difference. With the research to date and the data available it is impossible to say for certain to what extent - if at all - any of these theories are applicable to the MPS.  It is recommended that the MPS turn their focus to the development of possible interventions - either specific to the misconduct process or to staff more widely. This may include specific training, enhanced information provision, promoting more informal solutions through behavioural ‘nudges’, changes to process, or approaches that encourage and support de-escalation and informal resolution. 4 1. Introduction The overarching aim of this research was to explore the apparent disproportionality in misconduct cases in the Metropolitan Police Service (MPS) that sees Black, Asian and Minority Ethnic (BAME) officers more likely to go through this process. The learning and insights gained from the research will be important in driving understanding into the potential barriers and enablers to fair processes in the MPS for all officers. The overall objectives for the research were to:  Provide a comprehensive review into the disproportionality issue seen within misconduct data available from the MPS, and to ensure clear findings are derived from one agreed data source (provided by the MPS);  Identify whether particular types of misconduct, or other factors, are driving disproportionality;  Understand better current misconduct processes in the MPS and explore whether there are any aspects of them that may contribute to disproportionality;  Further explore possible reasons for the disproportionality and to suggest possible responses / interventions. 5 2. Methodology and Results The research encompassed three elements: A) A review of the existing evidence in this field to garner the learning in the wider literature; B) An in-depth analysis of five years’ worth of MPS misconduct data (2010 – 2015); C) A series of interviews with key staff involved in the assessment and investigation of misconduct in the MPS. A) Review of the existing evidence The literature review explored the extent to which disproportionality is unique to the MPS, reasons for disproportionality and possible interventions. This section presents key findings: Disproportionality in misconduct is not unique to the MPS. A number of studies have evidenced similar patterns in other police forces in the UK (e.g. Greater Manchester Police, West Midlands Police and the British Transport Police; Hagger, Johnson, Smith and Robert, 2013), as well as the US (Lersch and Mieczowksi, 2000; Rojewk and Decker, 2009). Disproportionality is also not unique to policing. Research has found that BAME staff in the NHS are twice as likely to be disciplined compared to white staff (Archibong and Darr, 2010), whilst BAME barristers were not only overrepresented in the number of internal complaints raised against them, but also more likely to have external complaints against them upheld (Bar Standards Board, September 2013). A number of studies place the issue of disproportionality in misconduct within a wider context of the experiences of BAME officers in the police, with some suggesting that despite years of reform, overt racism, as well as covert racism are still common in areas such as recruitment, promotion, and in daily encounters with other officers (EHRC, 2009; House of Commons Report, 2016). Other studies have argued that a resilient dominant white police occupational culture still makes it difficult for BAME officers to fit in (e.g. O’Neill and Holdaway, 2007; House of Commons report, 2016). The existing research suggests that in this respect the experiences of BAME officers in the police are consistent with those of BAME staff in other professions. This includes the prison service (Prison Reform Trust, 2006), Higher Education (Equality Challenge Unit, 2011) and the NHS (NHS Workforce Race Equality Standard, 2016). Academic perceptions of the reasons for disproportionality in internal misconduct proceedings are divided, with three main theories currently being put forward: 1. Some studies suggest that disproportionality results from a fear of being accused of racism, leading to managers and supervisors to shy away from trying to find informal solutions. This in turn either deprives BAME officers of valuable learning and the opportunity to rectify behaviour, or leads to managers taking refuge in more formal procedures (Morris, 2004; Matravers, Motto and Tseloni, 2006; Smith et al, 2012). 6 2. Other studies argue that disproportionality must be understood in terms of implicit bias. However, whilst there are indeed numerous studies evidencing the existence of bias, for example police intervening disproportionately with black people (EHRC, 2016) or an association of black people with crime and violence (e.g. Correll et al, 2007), it must be noted that no quantitative research has been found which examined whether conscious or unconscious bias could explain the disproportionality in misconduct. 3. Lastly, a failure of forces to understand and deal with difference has been identified as an explanation for the disproportionality in misconduct. An independent review concluded that the police ‘appear to have problems managing difference in their workforce, and that the sorts of issues that managers believe can be settled by an informal conversation with white officers often end up in full blown disciplinary proceedings against officers from BAME backgrounds’ (Chapman, 2014, p.28). The literature review was unable to identify any research on ‘what works’ in responding to disproportionality - illustrative of the fact that whilst a number of studies have been able to evidence the existence of disproportionality, none have so far been able to put forward conclusive evidence as to its causes or alter it for the better. Nevertheless, a number of strategies have been suggested to address disproportionality and diversity issues more broadly. This includes increasing BAME staff within the police, particularly in senior ranks, and within professional standards departments (evidence provided by the Black Police Association, cited in House of Commons report, 2016); improving diversity training (O’Neill and Holdwaway, 2007); introducing coaching and mentoring for BAME officers; having more external assessors from BAME background on selection panels; establishing a BAME senior leaders’ forum to provide guidance to BAME officers seeking promotions; and encouraging managers to resolve conduct issues with BAME officers informally. B) In-depth analysis of five years’ worth of MPS misconduct data Analysis of five years’ worth of MPS misconduct data (2010-2015)1 was undertaken. In this time, a total of 4,777 officers received a misconduct allegation, of which 3,710 pertained to White and 1,028 to BAME officers2. This means that whilst BAME officers make up 14% of the MPS workforce, they accounted for 21.5% of those subjected to a misconduct allegation. Over the five years, there was an increase of 48% in the number of officers subject to allegations of gross misconduct3. The number of officers subject to allegations of misconduct4 reduced by 11% during the same five-year period (see Figure 1). 1 Please note that the analysis focused on internal conduct matter pertaining to officers only 2 Please note that the ethnicity of 39 officers on the misconduct database was either unknown or missing. 3 Gross misconduct is defined as a breach of the standards of professional behaviour so serious that dismissal could be justified. 4 Misconduct is a breach of the standards of professional standards within the conduct regulations. 7 Figure 1. Numbers of officers subject to allegations of misconduct each year from 2010 to 2015 The analysis confirmed two key findings around disproportionality: i. BAME5 officers are twice as likely as White officers to be subject to misconduct allegations (see Figure 2). Over the five-year period from 2010 to 2015 there was an average rate of 4.96 per 100 for BAME officers, versus a rate of 2.46 for non-BAME officers. Figure 2. Rate per 100 MPS officers of receiving an allegation of misconduct or gross misconduct ii. Police officers from a BAME background are more likely to have a misconduct allegation substantiated, as compared to officers from a White background (48% vs 39%). White officers are more likely to have an allegation against them unsubstantiated (see Figure 3). This difference is statistically significant. 5 Ethnicity categories within the data set provided referred to White, Black, Asian, Other and Unknown. For the purposes of the analysis, BAME officer was defined as Black, Asian or Other. 8 Figure 3. Outcomes of police officer misconduct allegations in the MPS iii. This disproportionality is not found for Metropolitan Special Constabulory (MSC) officers (see Figure 4) – where the opposite pattern is observed and White officers are more likely to have a misconduct allegation substantiated (43% vs 38%). This difference is not statistically significant. Figure 4. Outcomes of misconduct allegations against officers of the Metropolitan Special Constabulary No causal links could be identified and the data was not able to explain the disproportionality gap. Nonetheless, the analysis was able to reject a number of hypotheses:  A higher rate of misconduct allegations against BAME officers is not associated with length of service or age of officer. BAME officers with less than 10 years’ service are still more likely to be subject to a discipline action than non-BAME officers with the same service length. 9  The disproportionality gap is not driven by differences in allegation types between White and BAME officers. The gap is replicated across all different allegation types for officers (see Figure 5). Per 100 officers, BAME police officers are twice as likely as non-BAME officers to be subject to misconduct allegations – this ratio is similar across allegation type (ranging from 2.2:1 for Failures in Duty to 1.8:1 for Oppressive Behaviour). Figure 5. Average rate per 100 officers of misconduct allegation by allegation type  Disproportionality is not driven by a difference in on vs. off duty behaviour between BAME/White officers. BAME officers are more likely than White officers to be subject to misconduct allegations irrespective of whether the behaviour was on or off duty (see Figure 6). Figure 6. Rate per 100 MPS officers of receiving a misconduct allegation whilst on duty vs off duty  The MPS misconduct data alone is not able to evidence whether BAME officers are more likely to be engaged in behaviour which leads to misconduct allegations. However, there is no disproportionality in the number of public complaints made against BAME vs. 5.5 0.9 3.9 17.0 3.0 1.5 2.1 3.1 0.6 1.9 7.7 1.1 0.8 1.1 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 Oppressive Behaviour Discrimination Malpractice Failures in Duty Inclusivitiy Traffic Irregularity Other Rate per 100 Workforce Police Officers Misconduct by Allegation - Average Rate per 100 Workforce BME White 10 White officers. BAME officers account for just 14% of the workforce and 15% of all public complaints.  The disproportionality in misconduct outcome is not associated with a higher rate of BAME officers resigning/retiring. The disproportionality exists even when resigning/retiring officers are removed from the data (see figure 7). Figure 7. Outcomes of police officer misconduct allegations in the MPS, with those in which officers retire or resign before the process is completed removed from the dataset.  BAME officers do not have a greater chance of receiving a more severe misconduct outcome than White officers. For those misconduct allegations that are substantiated, there is little difference between BAME and White police officers in the write off rmethod (see Figure 8). However, BAME officers are more likely than White officers to receive management action in unsubstantiated cases (see Figure 9). Figure 8. Write off method for police officer misconduct allegations in the MPS that were substantiated Figure 9. Write off method for police officer misconduct allegations in the MPS that were unsubstantiated 11 C) Interviews with key staff involved in the assessment and investigation of misconduct A total of 11 semi-structured one-to-one interviews with key staff were carried out. This included Professional Standards Champions (PSCs), as well as representatives of centrally based Directorate of Professional Standards’ teams6. Interviews focused on four areas:  Understanding better the decision-making that takes place at different stages of the misconduct process in the MPS;  Exploring the processes in place to assess and investigate misconduct in the MPS, and the extent to which they may contribute to disproportionality;  Understanding what happens with alleged misconduct before it enters a formal process; and  Gathering the feedback and opinions of those directly involved in assessing and investigating misconduct on the findings of the data analysis. Analysis of the interviews identified five key themes: Perception of roles Many of those interviewed had not necessarily sought out their roles, especially on borough. Nevertheless, all talked about the importance of representing the MPS and being motivated by a strong belief in upholding standards. ‘We empower our staff. They are heavily supervised but when they walk out of that front door they are on their own and they are representing everybody… I have 50 officers on a response team and they go out every morning and evening and how they behave is very important.’ (Professional Standards Champion) PSCs in particular took a pragmatic approach, seeing their roles as simply coming with the territory. At the same time, they felt - especially when compared with some of the other responsibilities they were holding on their boroughs - that the professional standards aspect was not always seen as a priority by others. ‘No one really gets on my case about complaints. They get on my case about crime…’ (Professional Standards Champion) 6 The Complaints Support Team, Serious Misconduct Investigation Unit and the Hearings Unit 12 Prevention A strong theme emerging from the interviews was a focus on early intervention and prevention and on wanting to be pro-active. ‘There is a real focus on being preventative, and supporting officers to do the right thing and therefore preventing complaints… we always want to take time out to do preventative work.’ (Professional Standards Champion) In this respect – there was almost unanimously positive feedback on the Taylor reforms7, which were perceived as having shifted focus from ‘nailing someone’ to early intervention and identifying learning. ‘early admission and saying sorry probably means we can deal with it at a fairly low level; the ethos around Taylor is not about turning everything into a drama and nailing someone.’ (Professional Standards Champion) However, the ability to be pro-active was often dependent on resourcing and staffing levels, with feedback from the boroughs in particular that current levels make it difficult for them to focus on a pro-active preventative approach. In places where they had been able to, feedback was overwhelmingly positive, with reports of an overall reduction in misconduct cases. Good practice arising from being able to focus on prevention included:  The Complaints Intervention Scheme;  Borough-level professional standards strategies in place;  Analysis of local misconduct data to identify themes and patterns; and  Training and ongoing development - focused in particular on new joiners and newly promoted officers - around standards of behaviour, what constitutes misconduct and types of issues that are current themes in misconduct allegations. A key challenge to prevention and early intervention was seen to come from a culture of reluctance to use informal solutions. Many of those interviewed felt that supervisors often either didn’t have the time, or did not feel comfortable to give informal feedback to their staff. As a result, behaviour was either allowed to escalate, or supervisors defaulted to formal misconduct procedures. PSCs especially talked about conducting initial assessments on a lot of cases they felt could have been dealt with informally (or could have before it escalated). 7 The Taylor reforms refers to legislation passed on 1 December 2008 , which introduced new police misconduct and performance procedures throughout the Police Service in England and Wales. The legislation introduced new professional standards, the reintroduction of Unsatisfactory Performance Procedures (UPP) and new procedures for dealing with misconduct in the police service. It also moved the emphasis of the police discipline framework from punishment to professional development, learning and improvement. 13 ‘…stuff that comes to us shouldn’t come to us,… if supervisors were strong enough, then it would not need to come to us.’ (DPS Officer) Decision-Making Most of those interviewed had not received any formal training for their roles, instead describing it as learning/training on the job. Whilst some expressed the wish for more comprehensive initial, as well as ongoing, training, most felt that the ‘on the job’ approach and support from colleagues had fully equipped them for the role. When assessing potential misconduct they referred to the Assessment of Conduct ‘pyramid’, as well as using common sense. ‘Because some of the judgement is would I have behaved in this way…’ (Professional Standards Champion) Some acknowledged the potential for subjectivity in this, but, nevertheless, all perceived there to be good levels of consensus and consistency in decision-making and the conclusions they came to when assessing a case. Indeed, there were plenty of ‘control points’ along the process, ensuring decision-making did not rest solely with one person. This quality control chain included going back and forth between local professional standards units and central DPS teams. ‘Most of the time it is easy to make an assessment, it only gets complicated when an individual racks up the complaints, but usually the DPS agree with my decision, I only had them disagree once.’ (Professional Standards Champion) MPS Misconduct Processes There was mostly positive feedback on the processes in place for handling misconduct in the MPS – some of those interviewed described them as somewhat clunky and high in paperwork, but overall it was felt that current processes ensured the fairest assessment and robust processes. Amongst those interviewed there was also consensus that these processes were applied consistently and that there were no differences in the way cases of officers and Specials were handled, nor between on and off duty cases. Investigations were approached like any other investigations the police conduct – they were seen as absolutely thorough, although interviewees also stressed the importance of evidence gathering to remain proportionate. When asked about factors impacting on the length of an investigation, two key issues were mentioned: firstly - the work load of the investigating officer and secondly - the complexity of the investigation. In cases where the IPCC became involved 14 this was acknowledged as necessary, but often caused major frustrations because these cases were seen to often suffer from delays and interference. ‘The interference by the IPCC will always be a big bugbear. This is length of time, but also general interference. We can find no case to answer and they can direct a hearing, and I think… can you imagine being an officer accused of gross misconduct and you find out we investigated and found no case to answer, but then… ‘ (DPS Officer) Making sense of disproportionality Interviewees were consistent in their reactions to the findings of the data analysis: whilst they accepted the results, they felt it did not reflect their lived experiences. All insisted that they did not act with any bias, nor thought that the MPS as a whole did. ‘Nope, I really cannot explain that finding. It’s not consistent with my work.’ (DPS Officer) Interviewees pointed out that they were not the ones that initiated the misconduct process and therefore had no influence over which officers received allegations. This suggests that the disproportionally may come from earlier in the chain, although this does not answer the question of why there is a difference in the number of allegations substantiated between BAME and White officers. Furthermore, in assessing misconduct, focus was felt to be on the circumstances of each case, rather than the person. Indeed, both those from central teams and PSCs on borough insisted that they would not necessarily know the officers involved. ‘Cases are dealt with on their merit. I don’t necessarily know the ethnicity of the officer…’ (DPS Officer) In talking about possible causes for disproportionality, interviewees could only speculate. Nevertheless, most commonly talked about was a fear of being accused of racism, which interviewees felt led to supervisors in particular being less comfortable and therefore less likely to resolve issues informally. This increased the likelihood of misconduct allegation in two ways: firstly - a desire to do things ‘by the book’, increasing the likelihood of starting formal procedures; and secondly - behaviour being allowed to escalate until misconduct occurs. ‘People in general are quite worried about being accused of being racist, to the point that they will bend over backwards to avoid being seen as racist. People are frightened of doing the wrong thing so maybe it is people putting things down on paper just to make sure.’ (DPS Officer) 15 4. Discussion In summary, whilst there is strong and consistent evidence for the existence of disproportionality in misconduct, we remain unclear as to its causes and, therefore, possible interventions. In terms of causes, there are three academic theories to be considered: 1. Fear of being accused of racism; 2. Conscious / unconscious bias; and 3. Failure to deal with difference. With the research to date and the data available it is impossible to say for certain to what extent - if at all - any of these theories are applicable to the MPS. However, what is clear is that the findings from the MPS are entirely consistent with those of other police forces and other professions. Considering the wider disproportionality observed in the UK (such as increased representation of BAME populations among those stopped and searched, those in custody, or among the prison population), this suggests that if answers do exist, they lay both within the police service but also beyond it. Amongst those interviewed for the research, concerns over an overall reluctance to deal informally with problematic behaviour or performance issues emerged as a theme. This - in line with the ‘fear of being accused of racism’ theory - may be further exacerbated for BAME officers. Identifying ways to empower investigators and supervisors to explore more informal solutions may lead to a reduction in formalised misconduct cases, including those pertaining to BAME officers. The analysis and research conducted so far is unable to evidence to what extent conscious or unconscious bias might play a role in the disproportionality found within the MPS misconduct data. Those interviewed put forward the point that they only dealt with what they were given - and it is difficult to know, and indeed to research - what may have happened before cases entered a formal process. Some options for exploring this aspect (and wider issues) further may include:  An in-depth series of case-reviews of comparable fact cases;  Further analysis of public complaints data to see whether disproportionality appears later in the process (as has been found in the study pertaining to barristers; Bar Standards Board, 2013);  Further analysis into the process, such as rank of original complainant or workload/ratio of cases for the investigators;  Further analysis of misconduct data with a specific focus on unsubstantiated cases and any ethnic disproportionality in the receipt of subsequent management action. 16  Experimental approaches exploring possible conscious or unconscious bias through a series of scenario-based decision-making exercises. The ‘failure to deal with difference’ theory would suggest that the disproportionality we see in the misconduct data is a symptom of much wider issues – in terms of how White and BAME officers relate to one another - suggesting that we need to look for solutions more widely too. Training will have a crucial role here, and it may be worth examining to which extent current diversity (and other) training addresses issues around ‘difference’, particularly in the context of acknowledging traditional police culture. In addition to any further research that could be conducted to aid our understanding around disproportionality in misconduct, it is recommended that the MPS turn their focus to the development and design of possible interventions to counter the issue. Such interventions may be either specific to the misconduct process or to staff more widely. It is recommended that this includes specific training – focused upon investigators of misconduct, or supervisors around dealing with unsatisfactory behaviour and encouraging early resolution, or indeed more generic training around diversity and dealing with difference. Other options may be found around enhanced information provision, behavioural ‘nudges’, changes to process, or approaches that encourage and support de-escalation and informal resolution. Any intervention trialled, if designed and implemented well can in turn be subject to an evaluation to strengthen the evidence base. 17 References Archibong, U., & Darr, A. (March, 2010). The involvement of Black and minority ethnic staff in NHS disciplinary proceedings. Centre for Inclusion and Diversity, University of Bradford, Yorkshire, UK. Retrieved from http://www.nhsemployers.org/SiteCollectionDocuments/Disciplinary%20Report%20Final%20with%20ISBN.pdf Bar Standards Board (September 2011). Professional Conducts Department Diversity Report 2009-2010. Retrieved from https://www.barstandardsboard.org.uk/media/32396/diversity_report_2009-10.pdf Chapman, C. (October 2014). An Independent Review of the Police Disciplinary System in England and Wales. Retrieved from: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/385911/An_Independent_Rev iew_of_the_Police_Disciplinary_System_-_Report_-_Final....pdf Correll, J., Park, B., Judd, C. M., & Wittenbrink, B. (2007). The influence of stereotypes on decisions to shoot. European Journal of Social Psychology, 37, 1102-1117. doi: 10.1002/ejsp.450. Equality and Human Rights Commission (2016). Stop and think: A critical review of the use of stop and search powers in England and Wales. Retrieved from https://www.equalityhumanrights.com/en/publication- download/stop-and-think-critical-review-use-stop-and-search-powers-england-and-wales Equality and Human Rights Commission (2009). Police and Racism: What Has Been Achieved 10 Years After the Stephen Lawrence Inquiry Report. Equality Challenge Unit (2011). The experience of black and minority ethnic staff in higher education in England. Retrieved from: http://www.ecu.ac.uk/publications/experience-of-BAME-staff-in-he-final-report/ Hagger Johnson, H. H., Smith, G., & Roberts, C. (2013). Disproportionality in internal misconduct and counter- corruption proceedings in three English police services. Policing, 7, 359-369. doi: 10.1093/police/pat030. House of Commons (May 2016). Police diversity. First Report of Session 2016–17. Retrieved from: http://www.publications.parliament.uk/pa/cm201617/cmselect/cmhaff/27/27.pdf Lersch, K. M., & Mieczkowski, T. (2000). An examination of the convergence and divergence of internal and external allegations of misconduct filed against police officers. Policing: An International Journal of Police Strategies & Management, 23, 54-68. doi: 10.1108/13639510010314616. Matravers, A. Motto, M. & Tseloni, A. (2006). Disproportionality in the Metropolitan Police Service (MPS). NHS Workforce Race Equality Standard (May, 2016). 2015 Data Analysis Report for NHS Trusts. Retrieved from: https://www.england.nhs.uk/wp-content/uploads/2014/10/WRES-Data-Analysis-Report.pdf O’Neill, M., & Holdaway, S. (2007). Examining ‘window dressing’: The views of Black Police Associations on recruitment and training. Journal of Ethnic and Migration Studies, 33, 483-500. doi: 10.1080/13691830701234780. 18 Prison Reform Trust (July, 2006). Experiences of minority ethnic employees in prison. Retrieved from http://www.prisonreformtrust.org.uk/Portals/0/Documents/experiences%20of%20minority%20ethnic%20emplo yees%20in%20the%20prison%20service.pdf Rojek, J. & Decker, S. H. (2009). Examining racial disparity in the police discipline process. Police Quarterly, 12, 388-407. doi: 10.117771/1098611109348470 Smith, G., Hagger Johnson, H., & Roberts, C. (2012). Disproportionality in police Professional standards [online]. Greater Manchester Police. Available from: https://www.escholar.manchester.ac.uk/api/datastream?publicationPid=uk-ac-man- scw:170650&datastreamId=FULL-TEXT.PDF YOUTH SURVEY 21-22 REPORT 1 ‘We Are London’ Youth Survey 2021-22 April 2023 MOPAC Evidence and Insight YOUTH SURVEY 21-22 REPORT 2 YOUTH SURVEY 21-22 REPORT 3 YOUTH SURVEY 21-22 REPORT 4 1. Introduction The three years since MOPAC’s last Youth Survey have brought with them unprecedented challenges. Much of this time has been dominated by the COVID-19 pandemic, the impacts of which have been felt globally and across every corner of society. Policing and crime were no exception to this. Political, social and economic shifts over recent years have altered the operational landscape, while several high-profile events placed policing under the public spotlight. It is against this backdrop that the Mayor’s Office for Policing And Crime (MOPAC) launched our latest Youth Survey – ‘We Are London’ – to help understand what matters most to young Londoners, and to give them a say in how their city is policed. Indeed, first release data from the 2021 Census reveals that over a million Londoners are currently aged between 10 and 19 (Greater London Authority, 2022) - reflecting more than ten percent of the capital’s total population and highlighting the importance of undertaking research to understand the views of this next generation. ‘We Are London’ represents the fourth wave in an ongoing series of surveys of young people conducted by MOPAC, with previous iterations taking place in 2018 (‘Youth Voice’), 2015 (‘Youth Matter’) and 2013 (‘Youth Talk’). 1.1. COVID-19, crime and young people Recent years have seen considerable shifts in the picture of crime and disorder across the capital, much of which has been spurred by the COVID-19 pandemic. When ‘lockdown’ took hold and the daily influx of commuters and visitors to the city waned, so too did many types of crime – including certain acquisitive and opportunistic offences such as street theft and robbery (Metropolitan Police Service, 2022). However, as the nation was told to ‘stay at home’, Londoners’ concerns instead shifted to more localised issues and calls to the Metropolitan Police Service about antisocial behaviour contrastingly increased. Such issues were also noticed by young Londoners. Many believed antisocial behaviour and drug dealing had become more visible in their area over ‘lockdown’ (Leaders Unlocked, 2020), while young people’s views of their neighbourhood became increasingly polarised: although some felt residents had ‘pulled together’ in the face of adversity, others believed community tensions had worsened as residents conflicted over Government restrictions (Leaders Unlocked, 2020; Partnership for Young London, 2021). Moreover – despite reductions in recorded knife crime during the COVID-19 pandemic (Metropolitan Police Service, 2022b) – fear of physical and knife attacks nevertheless remained top safety concerns for young people, with many believing levels would rise as the Capital continued to recover from the pandemic (Partnership for Young London, 2021). And to some extent these concerns were realised: levels of violence and knife crime both saw uplifts once Government restrictions were relaxed and ‘lockdowns’ were lifted over the summer months. Even when looking at longer term trends, the picture of youth violence remains mixed. Encouragingly, the number of young victims of non-domestic knife crime (aged under 25) has seen a sustained downwards trend over recent years, while levels of serious youth violence in 2021 still remain far below pre-pandemic levels (MOPAC, 2022c). However, this picture becomes far less positive when focusing in on the most serious violence: in stark juxtaposition to an overall decline in homicides seen across London during 2021 – the number of teenagers losing their lives sadly bucked this trend and reached a record high of 30 (MOPAC, 2022c). For this reason, youth violence remains very much at the forefront of political discourse in London, and a key focus for the Metropolitan Police Service. Indeed, the Mayor of London’s Police and Crime Plan 2021-25 recognises ‘reducing and preventing YOUTH SURVEY 21-22 REPORT 5 violence’ as one of the top priority areas for policing in London, and advocates a public health approach – alongside a dedicated focus on violence affecting young people - to help reduce the harms caused by violent crime across the capital. 1.2. The changing face of risk and vulnerability For many young Londoners, the COVID-19 pandemic also brought important safeguarding implications. With residents increasingly confined to their homes, reports of domestic abuse also increased (Metropolitan Police Service, 2022b), whilst national charities saw record numbers contacting their specialist helplines with child welfare concerns (e.g. NSPCC, 2020). Once again, these issues also resonated with young people themselves: 58% of young Londoners felt their relationships with their family had become more strained during lockdown, while nearly four in five believed their mental health had been negatively impacted by the pandemic (Partnership for Young London, 2021). Further compounding this, around two-thirds of young Londoners said they now spend ‘far more time’ on the internet as a result of the pandemic, bringing with it the potential for increased exposure to a range of online harms including cyber-bullying, inappropriate content or exploitation. However – while such risks increased – opportunities for safeguarding and early interventions decreased. As schooling and education became broadly confined to the virtual classroom, face-to-face contact with peers, teachers and other trusted adults was inevitably affected. For many young people, ‘lockdown’ also meant the loss of valuable wider support services, including youth clubs, recreational activities and healthcare (Partnership for Young London, 2021; Public Health England, 2021). Indeed, the impacts of the pandemic will have been amplified among London’s most vulnerable children, for whom prolonged confinement at home coupled with a reduction in protective services may have intensified pre-existing exposure to adversity, neglect, or even abuse (NSPCC, 2020). 1.3. Policing the pandemic and community relations The COVID-19 pandemic therefore brought considerable shifts in the picture of crime and risk for young Londoners. However – stepping briefly into the realm of operational policing – recent years have also provided a particularly challenging backdrop. In unprecedented times, Government restrictions on freedoms and movements became enshrined in law, and the public looked towards the police to enforce them. However, research reveals mixed views among young people towards the police response to the pandemic; whilst many cited examples of positive engagement between officers and the community over this time, others raised concerns over the way in which restrictions were enforced (Aitkenhead et al., 2022; Leaders Unlocked, 2020). Indeed, although 83% of adult Londoners said they supported the Metropolitan Police Service getting additional powers to enforce Coronavirus measures during the first ‘lockdown’ period, support remained considerably lower for younger age groups and stood at 71% among those aged 16 to 24 (MOPAC, 2020). Compounding this, several high-profile incidents placed policing further into the public spotlight. The death of George Floyd in Minneapolis in May 2020 sparked widespread protests in London as part of the Black Lives Matter movement and became symbolic of structural racism in the police. Among adult Londoners, perceptions that the police ‘treat everyone fairly’ fell by 12 percentage points in the weeks following these events, with such declines once again even more pronounced for younger age groups (MOPAC, 2020). Indeed, for many young people these events appeared pivotal in shaping relationships with police, with the Black Lives Matter movement widely listed as one of the most important issues for their generation (Leaders Unlocked, 2020). YOUTH SURVEY 21-22 REPORT 6 Less than a year later, in March 2021, came the murder of Sarah Everard by a serving Metropolitan Police Officer, which fuelled a national debate on women’s safety and once again placed police standards and integrity under public scrutiny. Tensions between police and communities rose once more – further amplified by widespread criticism of the handling of subsequent vigils – with a national poll showing nearly half of UK women (alongside 40% of men) said their trust in the police had ‘declined since the details of the Sarah Everard case emerged’ (End Violence Against Women Coalition & YouGov, 2021). Specific to London, public trust in the Metropolitan Police Service saw a notable downturn in the weeks following the murder of Sarah Everard: an effect that was once again more pronounced amongst females (MOPAC, 2021). Although this gender disparity was relatively brief – with females’ opinions quickly returning to levels more in line with males – these events also spurred more sustained reductions in public trust that continued throughout 2021 and into 2022 (MOPAC, 2021; MOPAC, 2022a). More than a year on, the murder of Sarah Everard and allegations of misogyny remained very much at the forefront of Londoners’ minds in shaping their trust in police, highlighting the influence of these issues on relationships between police and communities (MOPAC, 2022b). Once again, young people were no exception to this: many cited an explicit mistrust in the police following the Sarah Everard case and called for reform to help tackle perceived misogyny and the ‘male culture’ of policing (Leaders Unlocked, 2021). Indeed, the declines in public confidence and trust in the Metropolitan Police Service over recent years have been particularly pronounced amongst London’s younger residents (those aged 16 to 24 (MOPAC 2022d)). Moreover, evidence has revealed emerging impacts on how this generation engage with the police, with disproportionate declines also seen in measures of compliance – from supporting operational tactics and reporting crime, through to obeying the law and following police orders (MOPAC, 2022a). Importantly then, falls in public trust and confidence bring potentially tangible impacts on how communities engage with police, with support amongst London’s youngest generation seeming critical to maintaining future police legitimacy. 1.4. Building the wider context of youth Reflecting back then, the last three years have presented a volatile environment for crime and policing in London, the impacts of which have been felt even among London’s youngest generation. However, the events of these years have not happened in isolation. Indeed, the fall in public trust and confidence in the Metropolitan Police Service can be traced back to 2017 (MOPAC 2022e). Looking more specifically at those aged under 16, previous waves of MOPAC’s Youth Surveys show that the views of this group consistently track below levels seen for adult residents (see for example Charleton, Harrison & Dawson, 2015; Ramshaw, Charleton & Dawson, 2018). And all of this also sits atop an existing body of research that has long-evidenced the unique position of children and young people within the criminal justice system. Estimates from the Crime Survey for England and Wales (CSEW) Youth Module consistently show relatively high levels of victimisation among those aged 11 to 16 – including for low-level violent offences and theft of personal property (Office for National Statistics, 2020). This was echoed in the last wave of MOPAC’s Youth Survey, whereby over one in ten young Londoners said they had been a victim of crime during the last year (Ramshaw, Charleton & Dawson, 2018). Furthermore, several authors have revealed an overlap between victimisation and offending among young cohorts (see for example Beckley et al., 2018), with analysis also revealing a complex interplay of victimisation, safety issues and involvement in violent crime among those growing up in the capital (Ramshaw, Charleton & Dawson, 2018). YOUTH SURVEY 21-22 REPORT 7 Looking at interactions with police, young people are far more likely to be the subject of police Stop and Search (Metropolitan Police Service, 2022c) and for many these encounters form ‘prototypical’ interactions with officers (Skarlatidou et al. 2021). Indeed, Stop and Search remains a particular concern among young people, with experiences and perceptions of perceived disproportionality in the use of such powers influential in driving wider mistrust in the police (Jackson & Smith, 2013; Murray et al., 2020). Specific to London, MPS Safer Schools Officers also sit in a unique position to engage with young people in the school environment, despite previous waves of MOPAC’s Youth Survey highlighting mixed awareness and engagement with these officers (Ramshaw, Charleton & Dawson, 2018). Moreover, by the very virtue of their age - and magnified by a range of potential personal circumstances - young people also face a range of specific vulnerabilities, including to exploitation or abuse (Children’s Commissioner, 2017): as such, the police possess a unique responsibility to help safeguard and protect the next generation. 1.5. ‘We Are London’ – hearing the voice of young Londoners It is against this backdrop that the Mayor of London has published his Police and Crime Plan 2021-25: the document that sets the strategic direction for crime and policing over the next four years. Accordingly, the Plan possesses a strong emphasis on youth and vulnerability, and keeping young Londoners safe from harm. The Plan commits to build trust and confidence in the police across London’s communities, and to ensure that the voices of young people are heard in the decisions that affect them. With all this in mind, the Mayor’s Office for Policing And Crime (MOPAC) launched the ‘We Are London’ Youth Survey 2021-22, which aimed to gather the views of young Londoners aged 11 to 16 across a range of crime and policing issues that affect them. Importantly, this group represent London’s future – the next generation to transition through adolescence into adulthood – bringing with them their views and experiences of the criminal justice system. Against this backdrop of change and uncertainty, the survey seeks to understand what matters most to those growing up in the capital, and to give young Londoners the opportunity to have a say in how their city is policed. YOUTH SURVEY 21-22 REPORT 8 2. About the Survey 2.1. Methodology and Fieldwork The ‘We Are London’ Youth Survey 2021-22 was conducted by the Mayor’s Office for Policing And Crime (MOPAC) and was hosted online using a secure portal provided by Opinion Research Services (ORS). Fieldwork took place between 29th November 2021 and 21st January 2022. Note that this timeframe followed the return to face-to-face schooling for most pupils in March 2021 as part of the gradual easing of Government COVID-19 restrictions in the United Kingdom. The survey was distributed to schools offering secondary provision (academic years 7 to 11) in London via a central email distribution list1. This included state-funded schools, private schools, and specialist educational establishments such as Pupil Referral Units. A total of 1,007 schools were invited to participate2, with two reminders sent during the fieldwork period. The survey covered the following topic areas: • Safety • Victimisation • Perceptions of the police • Safer Schools Officers • Stop and search • Serious youth violence • Online safety • Sexual harassment and healthy relationships. 2.2. Sample A total of 11,874 responses were received from young people living or going to school in London. Full sample demographics are provided in Appendix 1. Respondents were aged between 11 and 16 years old, representing secondary school years 7 to 11. Those in lower school years were more likely to take part than those in upper years, likely reflecting curriculum demands. 2% of the sample (188 of 8,080) said they had attended a Pupil Referral Unit (PRU). 53% of young people identified themselves as female (6,042 of 11,489), while 45% identified themselves as male (5,153 of 11,489), revealing a slight skew towards female respondents. A further 3% identified themselves as a non-binary gender (294 of 11,489). Young people were from a broad range of ethnic groups, with the achieved spread broadly reflecting the secondary school population as captured by the Department for Education’s School Census3 . 1 Please note that this approach differs to that used in previous waves of the Youth Survey (2015/2018), where the survey was distributed via the Metropolitan Police Service’s Safer Schools Officer network. 2 1,013 schools appeared on the central distribution list, but six emails bounced back with a failure to deliver error. 3 Data for 2019 shows 38% of London Secondary School pupils were from a White Background, 22% from an Asian Background, 21% from a Black Background, 10% from a Mixed Background and 6% from an Other Background. https://data.london.gov.uk/dataset/percentage-pupils-ethnic-group-borough YOUTH SURVEY 21-22 REPORT 9 2.3. Ethical Considerations Distributing the survey via schools provided an ethical safeguard for young people by ensuring that consent from appropriate adults (teachers or parents) was obtained before participation. To retain confidentiality and comply with the General Data Protection Regulation (GDPR), the survey did not gather any personally identifiable information relating to individuals or to schools. Before taking part, young people were provided with information about the research, and informed that participation was voluntary. Participants were able to skip any questions they did not want to answer, while questions relating to more sensitive topics also included explicit ‘prefer not to say’ options. Topic areas relating to online risks and sexual harassment were only asked to those in schools years 10 and 11 to ensure content was age-appropriate. Throughout the survey, young people were provided with links to relevant support organisations, with a detailed list also provided upon completion. 2.4. Survey Scope and Report Structure The aims of the ‘We Are London’ Youth Survey 2021-22 were two-fold. Firstly, the survey sought to understand how the views of young Londoners may have changed over time in light of the current climate of crime and policing. To this end, many questions therefore built upon those asked in previous waves to enable effective benchmarking. Secondly, the survey also sought to generate novel insight into young people’s views across a range of topical issues, and to this end several questions were newly introduced to this latest wave. The current report is divided into four sections, each of which explores a different aspect of young people’s views of crime and policing4. The first section, ‘Young People and Crime’ examines young people’s feelings of safety in London and provides an updated snapshot of criminal victimisation and experiences of reporting to the police. The second section, ‘Young People and the Police’ places a particular focus on trust and confidence in light of recent events and generates new insight into young people’s interactions with police in London. The third section, ‘Young People and Serious Violence’ explores the prevalence of exposure to weapon-possession and gang issues among young Londoners – including vulnerability to forms of criminal exploitation. Finally, the fourth section, ‘Young People and Safeguarding’ examines exposure to online risks among young people, and revisits early attitudes and opinions towards sexual harassment and healthy relationships. The report also concludes with a reflection across the topic areas discussed – ‘Building the Picture of Risk’ - with a particular focus on examining the potentially cumulative impact of exposure to harm amongst the small cohort of young Londoners exposed to multiple different forms of violence and exploitation. Please note that all differences quoted in this report are statistically significant at the p<0.05 level. 4 These sections broadly align with the structure of the survey questionnaire design. YOUTH SURVEY 21-22 REPORT 10 3. Young People and Crime 3.1. Feelings of safety at home and at school In line with MOPAC’s last Youth Survey in 2018, the majority of young Londoners continued to feel safe both at school and in their local area (see Figure 3.1)5. However, around one in ten said they felt unsafe at school (9%, 966 of 10,481), while one in five felt unsafe in their local area (20%, 2,102 of 10,704)6. Moreover, there was an overlap between these two locations, such that 5% of the total sample – or 493 respondents – said they felt unsafe both at home and at school. Importantly, this reveals a small but potentially vulnerable group of young Londoners growing up in the capital without a sense of safety across either of these locations. In turn, young people felt a range of crime issues were problematic both at school and in their local area (see Figure 3.2). Hate crime was most often perceived to be a problem, with 28% of young Londoners (2,968 of 10,620) feeling this was a ‘big problem’ in the area where they live, and 22% (2,243 of 10,420) feeling it was a ‘big problem’ at school. Beyond this, young people also appeared concerned about violence in their local area, with around a quarter feeling that people carrying knives (26%, 2,803 of 10,609), people joining gangs (25%, 2,612 of 10,602) and people being violent (25%, 2,660 of 10,569) were a ‘big problem’. This is in line with previous research that suggests fear of physical attack remained salient 5 88% feel safe at school (9,225 of 10,481); 76% feel safe at home (8,172 of 10,704). 6 Remaining respondents said they ‘did not know’. Regression modelling of safety (blue box) predicted whether or not a young person felt safe in their local area by each of the individual crime problems (big problem vs. bit/not a problem) whilst controlling for individual demographics (age, gender, ethnicity). Second stage also included interaction terms. YOUTH SURVEY 21-22 REPORT 11 for young Londoners despite COVID-19 (Partnership for Young London, 2021). All of the crime issues listed were perceived to be more of a problem locally than in school – supporting the lower levels of safety seen here. Certain groups of young people were less likely to say they felt safe at school and in their local area, including those from older age groups and those from Black or Mixed ethnic backgrounds. Importantly, young people who had attended a Pupil Referral Unit (PRU) were also significantly less likely to feel safe in their school (74%, 139 of 188) than those who had not attended a PRU (90%, 5,235 of 5,811). Results from MOPAC’s ‘We Are London’ 2021- 22 Survey also bring unique and valuable insight into smaller and potentially ‘hard to reach’ groups of young Londoners for whom safety concerns may be particularly prominent. Illustrating this, young people identifying as a non-binary gender were significantly less likely to feel safe either at school (70%, 181 of 257) or in their local area (59%, 154 of 263) when compared with males or females, and in turn were more likely to feel that hate crime was a ‘big problem’ (39% at school (100 of 255); 45% in the local area (118 of 261)). This highlights a minority group of young Londoners for whom safety concerns may be more salient - particularly with regards to hate and discrimination. Beyond this, results also revealed lower levels of safety among the small number of young people exposed to gangs and knife crime - including those who knew others involved, or who had felt pressured or been personally involved (see Sections 5 and 7 for more details). Once again, this highlights a minority of young Londoners – those potentially most vulnerable to violence - for whom safety concerns may be particularly salient. 3.2 London as a safe place for different groups Beyond their own personal feelings of safety, young Londoners were also asked whether they felt their local area was safe for different groups of people. While nearly eight in ten felt their area was a safe place for boys and young men, less than two-thirds felt it was a safe place for women and girls or for children and young people7 (see Figure 3.3). Responses for these questions are broadly in line with those seen for adult Londoners as measured by MOPAC’s Public Attitude Survey. However, it is interesting that the proportion of young Londoners feeling their local area is a safe place for children and young people tracks below the proportion saying they personally felt safe here. 7 Safe for women and girls, 6,632 of 10,639; safe for men and boys, 8,313 of 10,593; safe for children and young people, 6,834 of 10,625). Understanding the Link between Crime Concerns and Safety Regression modelling6 suggests that young people’s concerns about a range of local crime issues shape whether they feel unsafe in the area where they live – but that concerns about violence and drugs are most influential here. Holding other crime concerns and demographics constant, those who felt knife crime was a ‘big problem’ showed 1.9 times increased odds of saying they felt unsafe in their local area. This was followed by people using or dealing drugs (1.7 times) and people being violent (1.6 times). The only crime issue not to be independently associated with increased odds of feeling unsafe in the local area was hate crime. However, results also suggested that hate crime may be more influential in shaping feelings of safety amongst those identifying as a non-binary gender. YOUTH SURVEY 21-22 REPORT 12 When looking at young people’s views of gendered safety, further analysis reveals that perceptions of the local area as a safe place ‘for women and girls’ and ‘for men and boys’ are both grounded in wider concerns about crime8 – in particular knife crime. However, whereas a young person’s own gender9 and concerns about hate crime appear strongly influential in shaping their views of their area as a safe place for women and girls, this is not the case for judgements about men and boys. Instead, concerns about gangs and ethnicity are more important here, with young Black Londoners also showing independently increased odds of feeling their local area was NOT safe for men and boys. Together, these findings suggest that young people’s views of their local area as a safe place ‘for women and girls’ and ‘for men and boys’ are closely interlinked but are also grounded in slightly different issues. In particular, the link between gangs, ethnicity, and male safety appears prominent for young Londoners: this is to some extent evidenced in current crime trends, with violent crime disproportionately affecting young Black males in the capital, and half of teenage homicides showing links to gangs (MOPAC, 2022c). However, such factors appear far less influential in shaping perceptions of female safety, with a young person’s own gender identity and wider perceptions of discrimination and hate in their local area instead having a more important role here. 3.3 Young Londoners’ experiences of crime When asked about their own personal experiences of crime, one in ten young Londoners said they had been the victim of a crime in the last year (10%, 987 of 9,837), consistent with levels seen last time MOPAC’s Youth Survey was run in 2018 (12%). This result remains higher than that recorded by the Crime Survey for England and Wales (CSEW) Youth Module for children aged 10 to 15 (7%)10, with the CSEW having recorded a decrease compared with the previous year (ONS, 2020). However, direct comparisons are limited as the CSEW data does not span the COVID-19 pandemic period. In turn, certain groups of young Londoners were more likely to have experienced a crime. To illustrate, those identifying as a non-binary gender were around three times more likely to say they had been the victim of a crime (28%, 61 of 217), while the likelihood of victimisation also increased with age - from 5% of those aged 11 (82 of 1,738) to 18% of those aged 16 (76 of 427). Moreover, early victimisation was also associated with a range of wider risk-factors. Young Londoners exposed to elements of gang and knife crime were disproportionately more likely to also have been the victim of a crime during the last year (see Section 5 and 7 for further discussion), establishing an 8 Results are from two binary logistic regression models predicting feeling the local area is NOT safe ‘for women and girls’ and ‘for men and boys’ from crime problems in the local area (big problem vs. bit/not a problem) controlling for demographics (age, gender. Ethnicity). 9 Holding other variables constant, young females showed c. 3x increased odds and non-binary c. 5x increased odds of feeling their area was NOT safe for women and girls compared with young males. 10 CSEW result is the ‘Preferred measure’ that takes into account factors identified as important in determining the severity of an incidence. The ‘Broad measure’, which counts all incidents which would be legally defined as crimes and therefore may include low-level incidents between children, indicates a prevalence rate of 11% in R12 to March 2020. YOUTH SURVEY 21-22 REPORT 13 important overlap between these experiences. In turn, victimisation also appeared to have profound impacts on young people’s feelings of safety, with only just over half of young victims saying they felt safe in their local area (57%, 557 of 985). Young victims of crime were also far less likely to hold positive views of the police: just 28% said they had a good overall opinion of the police (275 of 986) compared with 48% of non-victims (3,624 of 7,599). Importantly, this highlights a range of wider issues affecting this potentially vulnerable group. When asked what type of crime they had experienced, young victims were most likely to classify the incident as a hate crime (30%, 272 of 894) followed by a theft or burglary (22%, 200 of 894) (see Figure 3.4). It is notable here that the proportion saying they experienced a hate crime has more than doubled compared with the last time MOPAC’s Youth Survey was conducted in 2018. This may in part reflect greater awareness of behaviours that constitute hate over this time, and is in line with previous findings that revealed hate crime as the issue young Londoners were most likely to feel was a ‘big problem’ - both at school and in their local area (see Figure 3.2). Certain groups of young victims were more likely to say they had experienced a hate crime, including those from Minority Ethnic Backgrounds (35%, 180 of 532 across all groups combined, compared with 24%, 87 of 362, from a White Background). Large differences were also seen here by gender: while 21% of young male victims classified their crime as a hate crime (86 of 413) this increased to 35% of young females (145 of 409) and stood at 56% amongst the small number identifying as a non-binary gender (31 of 55). 3.4. Reporting crime to the police Despite the prevalence of early victimisation and its potential overlap with wider vulnerabilities, only around a third of young victims said they reported the crime they experienced to the police (34%, 317 of 920). This is below results seen during the previous wave of MOPAC’s Youth Survey conducted in 2018 (44%). Furthermore, levels of reporting fell to just one in six among those who had experienced a hate crime (17%, 43 of 259); the lowest proportion for any of the crime types. Indeed, the type of crime seemed particularly influential in shaping whether or not a young person sought such support, with regression modelling showing that – even when controlling for aforementioned demographic differences and wider relationships with police – crime type remained a strong driver of whether or not young victims told the police about the crime they experienced.11 11 Binary logistic regression predicting if a young victim told the police about their crime by individual demographics (gender/age/ethnicity), perceptions (overall opinion of the police, confidence, trust), having an SSO at school, crime type. YOUTH SURVEY 21-22 REPORT 14 Amongst the young victims who DID report their crime to the police, just one in three said they were happy with the way the police dealt with the crime (31%, 97 of 317), while approaching half said they were unhappy (44%, 139 of 317).12 These results are broadly consistent with the last time MOPAC’s Youth Survey was run in 2018. Results suggest that negative experiences of reporting a crime may detrimentally impact on young people’s wider views of the police, with just one in five of those unhappy with the way the police dealt with their crime saying they had a good overall opinion of the police (20%, 27 of 138). Importantly though, opinions of the police were similarly low amongst the cohort of young victims who did NOT report their crime (23%, 129 of 562, compared with 36%, 115 of 316, of those who DID report), perhaps suggesting that poor opinions of the police may also serve as a barrier to seeking help in the first place. Reflecting back then, these findings highlight the prevalence of crime victimisation amongst young Londoners, coupled with relatively low levels of reporting and mixed experiences of the police response to such reports. Indeed, this picture of early victimisation becomes particularly concerning in light of the identified overlap with wider risk factors, including reduced feelings of safety and increased exposure to elements of gang and weapon-related violence, and this is discussed further in Sections 5 and 7. 12 The remaining young victims said they were neither happy nor unhappy (22%, 71 of 317) or did not know (3%, 10 of 317). YOUTH SURVEY 21-22 REPORT 15 4. Young People and the Police 4.1 Young Londoners’ overall opinions of the police Moving beyond crime and safety, MOPAC’s ‘We Are London’ Youth Survey 2021-22 sought to understand more about young Londoners’ perceptions and interactions with police in the capital. Overall, young people held mixed opinions of the police. 42% (4,921 of 11,855) said that they had a good opinion of the police, while one in six (15%, 1,720 of 11,855) said they had a bad opinion (see Figure 4.1). Results here are below levels seen last time MOPAC’s Youth Survey was run in 2018, where 50% held positive views. Remaining respondents either held no opinion (29%, 3,404 of 11,855) or said they did not know (15%, 1,810 of 11,855), highlighting a considerable proportion of young Londoners who may not yet have formed strong views of the police. Consistent with MOPAC’s last Youth Survey in 2018, certain groups of young Londoners were less likely to hold good opinions of the police. This includes those from minority ethnic backgrounds, with lowest results seen for young Black Londoners at 22% (369 of 1,712). Opinions also became more negative with age: whereas 58% of those aged 11 (1,177 of 2,024) said they had a good overall opinion of the police, this fell to just 25% among those aged 16 (141 of 557). Results also reveal an emerging gender gap, predominantly driven by worsening opinions among young females. Whilst 48% of young males (2,489 of 5,144) said they held good overall opinions of the police, this figure stood at just 38% for young females (2,304 of 6,034), a statistically significant gap of ten percentage points. The gap has widened considerably since the last time MOPAC’s Youth Survey was run 2018, when 53% of males and 51% of females held good opinions. These findings are in line with recent trends seen among young adults in MOPAC’s Public Attitude Survey, whereby perceptions of the police have disproportionately worsened among young females (aged 16 to 25) during the most recent financial year (MOPAC, 2022d). To offer further insight into young Londoners’ impressions of the Metropolitan Police Service, respondents were also asked to write the first three words they think of when they think about police in London (see Figure 4.2). Several positive words were mentioned, such as ‘brave’, ‘helpful’, ‘friendly’, ‘protect / protection / protective’, ‘safe / safety’, ‘strong’ and ‘trustworthy’. However, a range of more negative words also emerged - most prominently ‘racism / racist’ and ‘scary’ - but also including ‘bad’, ‘biased’, ‘corrupt’, ‘intimidating’, ‘rude’, ‘unhelpful’, ‘unreliable’, ‘untrustworthy’, ‘useless’ and ‘white’. Again, these findings highlight mixed opinions of the police among young people growing up in London. YOUTH SURVEY 21-22 REPORT 16 4.2. Wider perceptions and views of the police When looking at young people’s wider perceptions of the police, a mixed picture once again emerged (see Figure 4.3). Young Londoners appeared most positive about the police’s ability to keep them safe from crime, but were less positive about how police engage with young people. Illustrating this, while just under half believed the police would protect them from crime (49%, 5,812 of 11,829) or would be there when needed (43%, 5,083 of 11,840), only around a third felt the police listen to the concerns of young people (28%, 3,297 of 11,808) or deal with the issues that matter to young people (31%, 3,631 of 11,806). However, young Londoners were least confident in aspects of police fairness. Only a quarter believed the police treat everyone fairly, whatever their skin colour or religion (24%, 2,883 of 11,812), while just 17% (1,974 of 11,796) felt the police treat young people the same as they treat adults. Over one in three young people actively disagreed with these statements. Perceptions of fairness were again particularly low among certain groups: for example, while 40% of those aged 11 believed the police YOUTH SURVEY 21-22 REPORT 17 treat everyone fairly (803 of 2,013), this declined to just one in ten by age 16 (11%, 59 of 555), with similarly low levels seen for young Black Londoners at 12% (197 of 1,695). Moreover, young Londoners exposed to other forms of risk were also less likely to respond positively across a range of these perception measures. As discussed earlier in Section 3.3, those who had been the victim of a crime during the last year were less likely to hold positive views of the police, and it is perhaps particularly concernning that only one in three here believed the police could protect them from crime (33%, 324 of 980, compared with 56% of non-victims, 4,214 of 7,581). Beyond this, it is notable that those who felt unsafe where they live were also more likely to respond negatively – including for feeling the police do a good job in their local area (12%, 261 of 2,095, compared with 37% of those feeling safe, 3,031 of 8,154) - while those exposed to elements of violence or exploitation were also less likely to hold positive views across the suite of perception measures (see Section 5 and 7 for further discussion). These findings highlight some important challenges to relationships with police amongst groups of young Londoners who may be at increased risk of wider harms. In line with previous waves of MOPAC’s Youth Survey, results for many of the perception measures for young Londoners also continue to track below levels seen for adult counterparts as measured by the Public Attitude Survey (PAS). However, although recent years have seen a sustained decline in perceptions of the police among adult residents, comparable measures13 for young Londoners appear to have remained more stable and are broadly in line with those seen in MOPAC’s last Youth Voice Survey 2018. An important exception to this is feeling the police treat everyone fairly, which has seen a reduction of 13 percentage points over this time from 37% in 2018 to 24% in the current wave. Once again, this highlights perceptions of police fairness as a particular issue for young people growing up in the capital. 13 Several measures included in MOPAC’s ‘We Are London’ Youth Survey 21-22 are similar to those asked in the PAS, including feeling police can be relied on to be there when needed; deal with local issues; listen to local concerns; treat people fairly; and do a good job in the local area. Blue box: Trust was predicted by individual demographics (gender/age/ethnicity) and mean scores across perception questions on three factors identified by factor analysis: reliability (police can protect me from crime/will be there when I need them), fair treatment (police treat everyone fairly/treat young people the same as adults) and engagement (police are helpful and friendly/deal with issues that matter to young people/listen to concerns of young people). Exploring Young People’s Trust in the Metropolitan Police Service MOPAC’s Youth Survey 2021-22 newly captured trust in the police among young Londoners. Here, just 43% agreed the Metropolitan Police Service was an organisation they could trust – far below levels seen for adult Londoners in MOPAC’s Public Attitude Survey at 73% in 2021. Trust was even lower for certain groups of young people, at just a quarter among those aged 16 (26%, 146 of 557) and those from Black Backgrounds (24%, 416 of 1,700). A considerable gender gap once again emerged, with 38% of females trusting the police (2,283 of 6,019), compared with 51% of males (2,626 of 5,126). Regression modelling13 shows that trust in the Metropolitan Police Service is closely interlinked with young people’s wider views of the police. However - even controlling for these wider perceptions – the demographic differences above continued to persist. Furthermore, analysis reveals that perceptions of police reliability (i.e., feeling the police can protect me from crime/will be there when needed) are particularly influential in shaping wider trust; even more so than aspects of police engagement or treatment. This suggests that – for the youngest generation – a belief in the police’s ability to keep them safe seems a core element of trust. YOUTH SURVEY 21-22 REPORT 18 4.3. Young people’s priorities for policing in London When asked to select their top three priorities for policing in the capital, young people placed a clear focus on tackling violence and higher-harm issues. Those surveyed were most likely to feel the police should prioritise gun and knife crime (70% listed this in their top three), closely followed by domestic abuse or sexual violence (54% listed this within their top three). Many young people also placed a focus on gangs and gang-related crime and keeping children and young people safe (see Figure 4.414). These top priority areas consistently emerged across a range of demographic groups, suggesting a consensus among young people. However, results also revealed some more nuanced differences. For example, females and non-binary young people were far more likely to feel domestic abuse and sexual violence should be the first priority for policing in London than males (females: 27%, 1,494 of 5,461; non-binary: 38%, 101 of 264; males: 10%, 477 of 4,683). Similarly, the proportion listing building better relationships between the police and the public as their first priority increased with age, while young Londoners with ‘bad’ overall opinions of the police were also more likely to place this within their top three choices. Taken together, these findings demonstrate that young Londoners overwhelmingly agree that violence and safeguarding issues should form key priorities for policing in the capital, but also highlight some specific groups of young people for whom certain issues may be attenuated. 4.4. Young people’s interactions with the police MOPAC’s Youth Survey 2021-22 sought to bring new insight into young people’s personal interactions with police officers in the capital. Here, a quarter of those surveyed (25%, 1,406 of 5,534) said that they had spoken to or had contact with a police officer in London during the last year15. However, young people’s experiences of this contact were divided: half felt the interaction had been positive overall (45%, 615 of 1,356), but a quarter felt it had been negative (25%, 344 of 1,356) and 22% described it as mixed (300 of 1,356). 14 Total calculated by adding the proportions ranking as first, second or third priority. Please note that numbers therefore do not add up to 100%. Total N for first priority = 10,752; total N for second priority = 10,459; total N for third priority = 10,199. 15 This question was randomised and asked to half of respondents. YOUTH SURVEY 21-22 REPORT 19 Young people were asked to outline their contact with police in their own words. School-based contact with officers (such as in assemblies or lessons, or at careers events) or informal contact (such as saying hello, smiling or asking questions) were the most common forms - and were also generally seen as positive by young people. However, results also revealed an array of wider safeguarding, enforcement and victimisation issues that bring young people into contact with police in London (see Figure 4.5). Analyses once again suggest that early interactions with officers may be critical in shaping young people’s wider views of the police – in particular their trust in the Metropolitan Police Service. Illustrating this, while 46% (1,658 of 3,591) of those who had NOT spoken to an officer in the last year agreed the Metropolitan Police Service was an organisation they could trust, this stood at just 13% (43 of 344) among those who HAD spoken to an officer but felt the interaction had been negative. Crucially though, among those feeling their interaction had been positive, trust in the Metropolitan Police Service was conversely higher than those having no contact at all, at 58% (352 of 612). Importantly, this highlights the potential for recent interactions to both positively and negatively impact views: although negative encounters were associated with worse perceptions of the police, positive encounters may hold the potential to improve early perceptions – albeit to a lesser extent16. This is in line with a body of research that has also established an asymmetrical effect of citizen contact on perceptions of the police among adult populations (see for example Skogan, 2007). 16 These relationships remained true even when controlling for other variables, such as individual demographics (e.g. age, gender and ethnicity) and young people’s wider experiences of crime and policing (e.g. whether they had been stopped and searched, whether they had been the victim of a crime, whether they knew their school had an SSO). YOUTH SURVEY 21-22 REPORT 20 YOUTH SURVEY 21-22 REPORT - DRAFT 21 4.5. Partnership working between police and schools Despite the generally mixed perceptions of police outlined earlier, it is encouraging that nearly nine in ten young people surveyed believed it was important for the police to work together with schools (88%, 8,811 of 10,041). This highlights a clear appetite for engagement between officers and young people within the educational setting. However, when given a list of specific activities that police may undertake in school, young people’s views were more divided (see Figure 4.617). Encouragingly, around two-thirds said they would feel ‘more safe’ if Safer Schools Officers worked with schools to respond to crimes and support pupils - including by dealing with people who commit crime in school, being available for pupils to report crime or get advice, and helping find support for those experiencing difficulties. However, young people were less likely to feel that engagement or educational activities would make them feel ‘more safe’, including having Safer Schools Officers help to organise school activities or present in lessons or assemblies. Despite this, the number of young people believing such activities would make them feel ‘less safe’ remained small – while earlier findings also highlight the potential for early encounters with the police in an educational setting to hold wider benefits to building 17 Approximately 9,850 young people answered these questions. Support for school searches among those affected by serious youth violence Perhaps counterintuitively, young people who said they had been hurt or threatened with a knife, or those who had been pressured to carry a knife or drugs for others, were more likely to say school searches would make them feel ‘less safe’ at school: arguably two groups that such searches would most aim to protect. Importantly though, multivariate analysis shows this relationship weakened once controlling for a young person’s own direct involvement in gangs and/or weapon possession – suggesting the overlap between victimisation and offending is key to understanding this. YOUTH SURVEY 21-22 REPORT - DRAFT 22 positive relationships between young people and police (see Section 4.4). Importantly, young people’s views were particularly divided on having Safer Schools Officers help to search pupils if they thought they were carrying things they should not be. While half of young people believed this would make them feel ‘more safe’ at school, 17% (1,639 of 9,891) said this would make them feel ‘less safe’ – levels over three times higher than those seen for other activities. In general, certain groups of young people were less likely to say that activities carried out by Safer Schools Officers would make them feel ‘more safe’ at school – including older age groups, young people from Black Ethnic Backgrounds, and those who had attended a Pupil Referral Unit. Importantly though, gaps tended to be smaller when asked about Safer Schools Officers organising activities for young people; in fact, young Black Londoners were here most likely to say this would make them feel ‘more safe’ at school (56%, 737 of 1,318, compared with 46%, 1,741 of 3,748, of White respondents). This could highlight the potential for Safer Schools Officers assisting with school trips and clubs to help reach young people from a range of backgrounds. 4.6. Engagement and Interactions with Safer Schools Officers Just over a third of young people surveyed said they were aware that their school had a dedicated Safer Schools Officer (37%, 3,628 of 9,908). This is lower than levels seen the last time the Youth Survey was run in 2018 (57%), likely reflecting changes to the survey methodology which had previously been distributed in partnership with the Metropolitan Police Service Safer Schools network (see Figure 4.7). A large proportion of young people said they did not know whether their school had a Safer Schools Officer (47%, 4,662 of 9,908), while a further 16% (1,618 of 9,908) believed their school did not have an officer. Taken together, these findings highlight the potential to improve young people’s awareness of dedicated police officers working with schools and educational establishments across London. Among those who were aware that their school had a Safer Schools Officer, 41% (1,441 of 3,532) believed the presence of this officer made them feel more safe at school, while only a small minority said they felt less safe (3%, 99 of 3,532). These results are broadly in line with those seen in 2018. However, less than half of young people (47%, 1,658 of 3,527) said they would be likely to speak to their Safer Schools Officer if a crime was to happen to them or if they were worried about something. In line with the last time the Youth Survey was run in 2018, results were particularly low for those feeling unsafe at school (20%, 62 of 311), highlighting an important barrier to communication with schools officers among this group of potentially vulnerable pupils. Similarly, willingness to speak to a Safer Schools Officer was also lower amongst the cohorts of young people who had personally been the victim of crime during the last year (26%, 105 of 404) or who had been threatened or hurt with a knife (29%, 56 of 196), once again highlighting potential barriers to help-seeking amongst young Londoners potentially most at risk of harm from crime. YOUTH SURVEY 21-22 REPORT - DRAFT 23 Just over one in ten young people saying their school had a Safer Schools Officer said they had personally spoken to or had contact with this officer (13%, 445 of 3,529). This group18 most often said it had been on an informal basis, such as for a chat, or in a lesson, assembly or club (37%, 166 of 445). However, approaching a quarter said they had spoken to their schools officer as they had been the victim or a witness of a crime (23%, 101 of 445); 15% (68 of 445) wanted help or advice, and 12% (47 of 445) wanted to give information. Some young people said they had spoken to their Safer Schools Officer as they had been accused of doing something wrong or committing a crime (15%, 67 of 445), while 7% (29 of 445) said they had been searched by their schools officer. Overall few demographic differences were seen in the proportion of young people saying they had contact with their Safer Schools Officer or the type of contact experienced – including by age, gender, or ethnicity. Nevertheless, these findings highlight a range of scenarios that bring young people into contact with their Safer Schools Officers. Accordingly, young people’s experiences of such contact were also divided: although half felt their interactions had generally been positive (54%, 235 of 438), one in ten felt they had been negative (10%, 44 of 438) and a further 22% (97 of 438) felt they had been mixed. This mirrors earlier findings into young people’s interactions with police officers more widely (see Section 4.4), and once again different types of contact were viewed differently. Encouragingly, three- quarters of young people saying they spoke to their Safer Schools Officer on an informal basis or in a lesson/assembly/club felt this contact had been positive (73%, 121 of 166), again highlighting such routes as key to building relationships between Safer Schools Officers and pupils. 4.7. The prevalence of Stop and Search experiences To further explore young people’s early contact with the police, MOPAC’s ‘We Are London’ Youth Survey 2021-22 also sought to understand their experiences of Stop and Search. Here, 6% of young people said they had personally ever been Stopped and Searched by the police in London (577 of 9,543), while 25% said they knew someone else who had been Stopped and Searched in London (2,328 of 9,494). These results have both decreased compared with the last time the survey was run in 2018 (from 10% and 38% respectively). However, new questions also revealed that around half of those Stopped and Searched said this had happened more than once (47%, 262 of 563) - with a fifth saying they had been stopped more than three times (21%, 118 of 563). This highlights a considerable number of young people who may have had contact with police in this way on multiple occasions. As noted in the previous wave of this survey, the proportion of our young cohort saying they have been Stopped and Searched may be higher than expected among those aged 11 to 16.19 This may in part reflect young people’s own understanding of what constitutes ‘Stop and Search’, or experiences of related procedures (such as ‘Stop and Account’). Nevertheless, it is important to recognise that this proportion of young people feel or believe that they have been Stopped and Searched in London, and results in this section therefore emphasise the value of each interaction with the police; regardless of how such encounters would be officially recorded. The prevalence of self-reported Stop and Search experiences increased with age - from 3% of those aged 11 (53 of 1,675) to 12% of those aged 16 (50 of 414) - while young males were more than twice as likely to say they had been Stopped and Searched as young females (9%, 362 of 4,202, compared 18 Young people could select all types of contact that apply, and ‘other – please specify’ comments were back-coded where possible. 19 Stop and Search data recorded by the MPS shows 12.9 per 1,000 population searches involving a subject aged 10 to 14 (volume: 5,911, population: 456,696) (MPS, 2022c) YOUTH SURVEY 21-22 REPORT - DRAFT 24 with (4%, 183 of 4,815). Young Londoners from Black Backgrounds were most likely to say they had been Stopped and Searched (9%, 116 of 1,263); while those from Asian Backgrounds were least likely (4%, 91 of 2,479). Once again, young Londoners exposed to a range of wider vulnerabilities were more likely to have said they had been Stopped and Searched by the police. In particular, such experiences were disproportionately concentrated amongst those exposed to gangs or weapon-related violence: nearly a quarter of young people who knew someone in a gang said they had personally been Stopped and Searched in London (22%, 194 of 863), and this increased further to around half of those who had been asked or pressured by someone else to carry or hide a knife (50%, 71 of 143), or who had personally belonged to a gang (57%, 49 of 86). Such findings inevitably reflect the intelligence-led nature of Stop and Search employed as a tactic to specifically target those involved in criminality. However, it is worth noting that - even when controlling for exposure to such elements of gang and weapon-related violence – regression modelling continued to reveal increased odds of experiencing Stop and Search amongst the demographic groups outlined above. Furthermore – and in line with the overlap in such risk-factors identified earlier (see for example Section 3.1) – the prevalence of self-reported Stop and Search was inevitably also increased amongst those with wider vulnerabilities, including those who had been the victim of a crime during the last year (20%, 177 of 896) and those who felt unsafe in their local area (11%, 213 of 1,884) or at school (17%, 143 of 857). In turn, these findings once again serve to reinforce the particularly complex picture of wider harm and vulnerability affecting young Londoners coming into contact with the police through Stop and Search, and emphasise the importance of recognising such issues during these interactions. 4.8. Young people’s experiences of Stop and Search interactions Young people who said they had been Stopped and Searched by the police in London were also asked to reflect on their experiences of these interactions. Here, less than half of those stopped said the police were polite (40%, 220 of 544), treated them with respect (39%, 212 of 542) or explained why they carried out the Stop and Search (42%, 230 of 542) during their most recent interaction. This picture remains broadly similar compared with the previous wave of the Youth Survey conducted in 2018, but results are considerably more negative when compared with experiences of adult Londoners as captured by MOPAC’s Public Attitude Survey.20 Furthermore, young people who said that they had been subject to a Stop and Search more than once were in turn less likely to have responded positively about these three aspects of the interaction compared with those who said that they had been Stopped and Searched on just one occasion.21 In line with the last wave of MOPAC’s Youth Survey, results once again highlight the influence of early Stop and Search interactions in shaping young people’s wider views of the police. In general, young people who said they had personally been Stopped and Searched were also far more likely to have a bad overall opinion of the police (40%, 232 of 576) compared with those who had not been stopped (12%, 1,011 of 8,503). However, further analysis reveals that the nature of the Stop and Search 20 PAS result for the police were polite is 68%, treated you with respect is 68% and explained why they stopped and searched you is 82% for FY 2021/22, excluding those who said ‘I don’t know’. 21 Among young people who said that they have been subject to a Stop and Search more than once, 31% (79 of 254) said the police were polite, 29% (73 of 249) said the police treated them with respect and 38% (95 of 251) said the police explained the reason for the Stop and Search, compared with 53% (109 of 205), 52% (107 of 207) and 50% (104 of 208) respectively among those who said that they have been subject to a Stop and Search only once. YOUTH SURVEY 21-22 REPORT - DRAFT 25 interaction remains key here. Illustrating this, those who had been stopped - but who felt police were polite, respectful and explained the process - were no more likely to say they had a bad overall opinion of the police (15%, 18 of 119) than those not stopped. In contrast, a particularly large disparity emerged among young people who felt police had not done these three things, with nearly half in this group now saying they had a bad overall opinion of the police (48%, 202 of 421). Crucially, these findings reveal the importance of how – and not simply whether – young people experience Stop and Search, and emphasise the value of each individual encounter in shaping early views of the police. This is in line with a large body of research that has shown the link between fair and procedurally just interactions and wider public confidence and legitimacy amongst adults (e.g. Sunshine and Tyler, 2003; Tyler, 2003), with the potential for negative interactions to be particularly detrimental to public opinions (Bradford, Jackson and Stanko, 2009; Stanko et al., 2012). Looking beyond young people’s own experiences of Stop and Search, results from the ‘We Are London’ Youth Survey 2021-22 also brought new insight into what young people hear about Stop and Search from others. Indeed, among those saying they knew someone else who had been subject to a Stop and Search in London, more than half said they had heard negative things about this person’s experience (55%, 1,212 of 2,206), while only 5% (120 of 2,206) had heard something positive.22 4.9. Young people’s support for Stop and Search MOPAC’s ‘We Are London’ Youth Survey 2021-22 also sought to understand young people’s wider attitudes and support for police use of Stop and Search. Opinions were divided here: only around a third agreed that the police should carry out Stop and Search (35%, 3,298 of 9,521), with relatively large proportions saying they were neutral or that they did not know (see Figure 4.8). Although more young Londoners believed police use their Stop and Search powers fairly, confidence in this was still low, with just 51% agreeing here (4,795 of 9,489). When compared with previous waves of MOPAC’s Youth Surveys, the proportion of young people believing police should conduct Stop and Search is lower here (44% in 2018, 56% in 2013), perhaps revealing declining support for the tactic amongst young people. This result also stands below levels 22 The remaining respondents said they heard something mixed (19%, 420 of 2,206), hadn’t heard anything (11%, 239 of 2,206) or did not know (10%, 215 of 2,206). YOUTH SURVEY 21-22 REPORT - DRAFT 26 seen amongst adult Londoners as recorded in MOPAC’s Public Attitude Survey (PAS).23 Furthermore, whereas results from the Youth Survey show that young Londoners are more likely to agree that Stop and Search powers are used fairly than to agree that Stop and Search should be carried out, the reverse is seen among adult Londoners.24 In line with the mixed overall support for police use of Stop and Search, young people were also divided about the impacts of the tactic. To illustrate, half of young people believed that Stop and Search will help to stop people from carrying knives (53%, 5,005 of 9,478), while even fewer agreed that they felt safer knowing the police can use Stop and Search (38%, 3,641 of 9,474). Once again, attitudes towards Stop and Search varied across demographic groups. Support for all four of the measures outlined in Figure 4.8 declined by age, while young males were more likely to respond positively towards the tactic than either females or those identifying as a non-binary gender. To illustrate, while nearly half of young males agreed that they feel safer knowing police can use Stop and Search (45%, 1,892 of 4,163), this fell to just a third of young females (34%, 1,647 of 4,785) and to just 15% of non- binary young people (34 of 228). Those from White or Asian Ethnic Backgrounds were also more likely to respond positively across many measures than those from Black, Mixed, or Other Ethnic Backgrounds. However, few differences by ethnicity were seen in young people’s agreement that police use Stop and Search fairly; this is despite far larger inequalities seen amongst adult Londoners in the PAS. In turn, it is also worth noting that attitudes towards Stop and Search tended to be more negative amongst young people who said they had personally been Stopped and Searched, with only a quarter of this group saying they feel safer knowing police can use the tactic (25%, 143 of 564). 23 PAS result for agree that the police should conduct Stop and Search is 74% for FY 2021/22, excluding those who said ‘I don’t know’. 24 PAS result for confident that the police in the local area use their Stop and Search powers fairly is 63% for FY 2021/22, excluding those who said ‘I don’t know’. Blue box shows results from a regression model predicting agreement that police should carry out Stop and Search by individual demographics (gender/age/ethnicity), perceptions of the police, crime concerns in the local area, feeling safe in the local area, ever being subject to a Stop and Search, knowing someone ever subject to a Stop and Search. Exploring drivers of support for Stop and Search among young people, and the role of gender Young female Londoners tended to be less supportive of Stop and Search and have less positive perceptions of the tactic compared with young male Londoners, despite also being less likely to have said that they have been personally subject to Stop and Search. Regression modelling shows that the strongest driver of support for Stop and Search is having a good overall opinion of the police.24 However - even when controlling for these opinions and wider factors (such as personal experiences of Stop and Search and concerns about crime in the local area) - females still remain less likely to support the tactic. YOUTH SURVEY 21-22 REPORT - DRAFT 27 5. Young People and Serious Violence 5.1 Young people’s experiences of gangs and knife crime Young people taking part in MOPAC’s ‘We Are London’ Youth Survey 2021-22 were asked a bank of questions that aimed to capture exposure to and involvement in violence (see Figure 5.1). Several questions mirrored those asked in the last wave of the Youth Survey in 2018 – allowing for comparisons over time – whilst other measures were newly introduced to explore exposure to wider elements of criminal exploitation. Here, one in ten young Londoners said they know someone who is in a gang (10%, 874 of 8,869) – a decrease from around a quarter (23%) when the survey was previously run in 2018. Similarly, the proportion of young people saying they know someone who has carried a knife also declined this wave to one in seven (14%, 1,204 of 8,699) – from around a quarter (26%) in 2018. Looking at personal involvement in these behaviours, fewer young Londoners said that they themselves had belonged to a gang or had carried a knife or another form of weapon. However, these proportions remained broadly stable with those seen in the last wave of MOPAC’s Youth Survey in 2018.25 These findings are also in line with those from the UK Millennium Cohort Study where 3.7% of young people aged 14 reported carrying or using a weapon (Villadsen and Fitzsimons, 2021). Beyond this, a minority of young people also said they had been exposed to behaviours that may indicate elements of criminal exploitation. For example, 3% of our young cohort said they had been asked or pressured by someone to carry or sell drugs (231 of 9,022), while 2% said they had been asked or pressured by someone to carry or hide a knife (145 of 9,095). Together these findings highlight a range of violence-related risks experienced by young people growing up in the capital. Exposure to such risks became particularly attenuated amongst the small group of young Londoners who said they had attended a Pupil Referral Unit (PRU). Here, approaching a quarter said that they 25 2018 Youth Survey result for belonged to gang is 3%, carried a knife is 3% and carried another weapon is 5%. YOUTH SURVEY 21-22 REPORT - DRAFT 28 had been threatened or hurt with a knife (22%, 38 of 170), while around one in five said they had been asked or pressured to carry or sell drugs for someone (18%, 30 of 163). In turn, over one in ten young Londoners who had attended a PRU said they had personally belonged to a gang (11%, 18 of 170) or had personally carried a knife (13%, 22 of 172). This aligns with a body of previous research that identifies children outside of mainstream schooling to be at increased risk of gang involvement and youth violence (Dempsey, 2021). Certain demographic groups were also more likely to have been exposed to elements of gang and weapon-related violence. In particular, prevalence increased with age; such that those aged 16 were around five times more likely to say they had been threatened or hurt by someone with a knife (11%, 41 of 389) when compared with those aged 11 years old (2%, 38 of 1,603). Across most measures, young females were less likely than males to say they had been exposed to elements of gang or weapon-related violence. However, in some cases results were significantly higher amongst those identifying as a non-binary gender: to illustrate, 13% of this group said they had been threatened or hurt by someone with a knife (27 of 211) and one in ten said they had carried a weapon or something they intended to use as a weapon (9%, 19 of 206). In general, differences seen by Ethnicity tended to be comparatively small, with young people from Asian backgrounds significantly less likely to have been exposed to many of the measures of gang and weapon-related violence than those from other backgrounds. It is worth noting here that considerable overlap was seen between many of the eight measures of exposure to gang and weapon-related violence. For example, those who knew others in a gang were significantly more likely to say they themselves had been part of a gang (9%, 73 of 831) or that they themselves had carried a knife (9%, 77 of 832). In turn, it is perhaps particularly concerning that young people experiencing elements of exploitation – including being asked or pressured to carry drugs or knives – were also far more likely to have engaged in such behaviours. Moreover, we see that around half of young people who said they had been asked or pressured to sell drugs (43%, 95 of 222) or to carry a knife for someone else (55%, 74 of 134) said that they had been threatened or hurt with a knife – compared with just 4% of young people not experiencing such forms of pressure26. Importantly, young people exposed to elements of gang and knife crime are often overlapping populations, and the potentially cumulative impacts of exposure to multiple different forms of such risks are discussed further in Section 7 of this report. Building on this, results from MOPAC’s ‘We Are London’ Youth Survey 2021-22 also reinforce the overlap between involvement in violence and wider criminal victimisation. As mentioned earlier in this report (see Section 3.3) young people exposed to or directly involved in a range of gang and knife- related activities were far more likely to say they had been the victim of a crime. Although caution should be taken in interpreting results due to low base sizes, victimisation was particularly high among those saying they had personally belonged to a gang (54%, 43 of 79) or had been asked or pressured to carry or hide a knife (47%, 61 of 129). Once again, these findings highlight the complex relationship between victimisation and offending, and the potential for increased vulnerability to harm among young people exposed to or involved in serious youth violence. This reflects a body of existing research 26 N=302/8,404 for those who had been asked or pressured to carry or sell drugs, and n=332/8,558 for those who had been asked or pressured to carry or hide a knife. YOUTH SURVEY 21-22 REPORT - DRAFT 29 that finds victims and perpetrators of violence among young people are not distinct, but rather largely overlapping groups exposed to multiple, chronic sources of violence in their lives (Kincaid et al., 2021). In turn, this vulnerability was once again further reflected across broader measures throughout the survey – including through lower personal feelings of safety (see Section 3.1) and through poorer perceptions of the police (see Section 4.1). Indeed, building on previous findings, it is notable that across each of the eight measures of exposure to elements of gang and weapon-related violence, those saying ‘yes’ were around 20 percentage points less likely to say they have a good opinion of the police or that they trust the Metropolitan Police Service27. As an example, just one in five young Londoners who have been asked or pressured by someone else to carry or sell drugs for them believed the Metropolitan Police Service was an organisation they could trust (20%, 46 of 230, compared with 47%, 3,964 of 8,470), bringing potential implications for how police can best engage to safeguard this potentially vulnerable group. Moreover, this also sits atop previous findings that revealed young people exposed to elements of serious violence were also more likely to have come into direct contact with the police, including through Stop and Search (see Section 4.7). This is perhaps unsurprising given the intelligence-led nature of such tactics, but nevertheless again builds an increasingly complex picture of early contact and relationships with police amongst the cohort of young Londoners potentially most vulnerable to harm, exploitation and victimisation. These issues are also explored further in Section 7. Crucially, results once again highlight the value of recognising opportunities for safeguarding interventions through these frequent police interactions – with previous research emphasising the value of adopting trauma-informed approaches with perpetrators or victims of violence to help build trust in the face of detachment among those who need help (Kincaid et al., 2021). 5.2. Young people’s views on what drives violence Building on this, findings from MOPAC’s ‘We Are London’ Youth Survey 2021-22 also reveal that young people themselves broadly recognise the interplay of wider risks and vulnerabilities that can underlie violence. Illustrating this, young people cited a range of drivers when they were asked what they thought the main reasons were that some young people get involved in violence (see Figure 5.2). Responses most commonly described issues related to young people’s social environments (such as peer influence, interpersonal disputes and acting to be popular or appear strong), home environments (such as adverse family situations, domestic abuse, and poverty or the need to make money), fear, vulnerability and harm (such as being forced or pressured into violence, concerns for personal safety, and gang-related activity) and psychological wellbeing (such as mental health, emotional distress, trauma, and lack of support). Once again, many of these themes reinforce findings already highlighted in this report – including lower levels of safety among young people exposed to gangs and knife crime (see Section 3.1). Taken together, young people’s responses clearly highlight the broad and intersecting range of vulnerabilities and risk factors that may lead young people to involvement in violence, closely reflecting a body of research that has already identified factors in a young person’s background and upbringing as linked to an increased likelihood of violent offending (e.g. Dempsey, 2021; Early Intervention Foundation, 2015; Kincaid et al., 2021; Villadsen and Fitzsimons, 2021). 27 Compared with those saying ‘no’ or that they ‘do not know’ in response to survey measures relating to exposure to or involvement in violence. YOUTH SURVEY 21-22 REPORT - DRAFT 30 Figure 5.2. Exploring drivers of young people’s involvement in violence YOUTH SURVEY 21-22 REPORT - DRAFT 31 Figure 5.2 (continued) YOUTH SURVEY 21-22 REPORT - DRAFT 32 6. Young People and Safeguarding 6.1. Young people’s experiences and safety online MOPAC’s ‘We Are London’ Youth Survey 2021-22 also sought to understand young people’s exposure to wider safeguarding issues on the internet and social media. Positively, four in five young Londoners said that they feel safe when they are online (81%, 6,931 of 8,582, see Figure 6.1). However, despite this, around one in five said that something had happened to them online during the last year that had made them feel worried (19%, 1,542 of 8,021)28. Furthermore, a quarter of young Londoners said that someone had said mean things to them, bullied them, or upset them online during the last year (25%, 1,992 of 7,991) – with 40% of this group believing they had been specifically targeted because of a protected characteristic (40%, 769 of 1,933). Together, these findings suggest that many young Londoners may have been exposed to potentially harmful experiences online. However, it is particularly concerning that the small cohort of young Londoners identifying themselves as a non-binary gender were around twice as likely to say that someone had said mean things to them, bullied them or upset them online (51%, 93 of 182), and were in turn also more likely to believe they had been targeted because of a protected characteristic (78%, 72 of 92). These findings link back to earlier results that revealed the prevalence of hate crime amongst young victims (See Section 3.3), and importantly highlight the extent to which such issues may also pervade into the online sphere. Moreover, results indicate that experiencing something worrying online may have a particularly negative impact on shaping young people’s feelings of safety online: controlling for individual demographics, this emerged as an even stronger predictor of whether or not someone felt safe than having been bullied or upset online29. In turn – despite previously seeing that non-binary young people were more likely to have been exposed to harmful online experiences - this group still continued to feel less safe online even when controlling for such experiences, suggesting other factors may also be influential in shaping lower security amongst this potentially vulnerable group. Beyond this, results from MOPAC’s ‘We Are London’ Youth Survey 2021-22 also sought to explore the prevalence and sources of potentially harmful or inappropriate content – including violent content and content that encourages hate or discrimination – on social media platforms (see Figure 6.2).30 Here, young Londoners were most likely to say they had seen such content on TikTok, with over a quarter saying they had seen violence or something that encouraged violence on the platform (28%, 989 of 3,473) and almost two-fifths said they had seen something encouraging hate or discrimination 28 Apart from overall online safety question, all other measures in this section exclude young people who say they ‘don’t use the internet or social media’. 29 Binary logistic regression predicting safe online by individual demographics (gender/age/ethnicity), having had someone bully/upset/say mean things online and having had something worrying happen online 30 Respondents were randomly allocated to answer either ‘In the last year, have you seen violence or anything that encouraged violence in any of the following places?’ or ‘In the last year, have you seen anything that encouraged hate or discrimination (such as racism or sexism) in any of the following places?’. YOUTH SURVEY 21-22 REPORT - DRAFT 33 (38%, 1,329 of 3,531). However, it should be noted that responses may reflect whether - and how often - young people access each of these different platforms; alongside the prevalence of such content. Nevertheless, less than half of respondents believed the police understand online issues that affect young people (45%, 3,585 of 8,030), with agreement declining by age (from 60% of those aged 11, 853 of 1,416, to 27% of those aged 16, 95 of 355). Furthermore – despite potentially increased exposure to online targeting– only one in five non-binary young people believed the police understand such issues (19%, 36 of 186). 6.2. Acceptability of sexual harassment, controlling and coercive behaviours Building on this picture of online safety, MOPAC’s ‘We Are London’ Youth Survey 2021-22 also sought to explore early attitudes and experiences of wider safeguarding risks including healthy relationships and sexual harassment. Please note that these questions were only asked to young people in school years 10 and 11 (aged 14 to 16) due to the nature of these topics. Young people were given a range of scenarios that could be considered to be coercive control, sexual harassment or domestic abuse, and were asked whether they felt it was ‘always ok’, ‘sometimes ok’ or ‘never ok’ for people to display these behaviours (see Figure 6.331). Responses for these questions were mixed. Many young people felt it was ok to display elements of coercive or controlling behaviour in a relationship, although fewer felt sexual harassment was acceptable – such as staring and wolf-whistling or making sexual comments or jokes. In turn, young people were least likely to feel it was acceptable to physically hit or push a partner or to share intimate pictures without their consent. In line with results seen in 2018, young females were more likely to feel many behaviours were ‘never’ ok, while young males were more likely to say they ‘did not know’. This could highlight greater uncertainty among young boys growing up in the capital around the acceptability of behaviours that constitute harassment or abuse. Illustrating this, young males were over four times more likely than young females to say they ‘did not know’ whether it was ok to share intimate pictures of a partner without their consent (9%, 85 of 932, compared with 2%, 24 of 1,094). 31 N for these questions = 2,154 to 2,170 YOUTH SURVEY 21-22 REPORT - DRAFT 34 Importantly, young people exposed to a range of wider violence-related risks were also more likely to feel such behaviours were acceptable, perhaps also indicating an underlying trend towards the normalisation of potentially harmful behaviours among this vulnerable cohort. For example, those who knew other people in gangs were over twice as likely to feel it was ok to hit or push a partner during an argument (8%, 26 of 308) than those not knowing anyone in a gang (3%, 48 of 1441). Moreover, this effect was even more pronounced among the small group of young people personally involved in violence, with around one in five of those who had previously carried a weapon here believing it was acceptable to hit or push a partner (19%, 15 of 80). This builds upon the already complex interplay of risk factors seen amongst young Londoners exposed to gang and weapon-related violence (see Section 5). Indeed, it is perhaps particularly concerning that young people exposed to multiple forms of violence or exploitation were in turn even more likely to normalise abusive behaviours, and this is discussed further in Section 7. Encouragingly, results reveal some positive shifts in young Londoners’ attitudes towards sexual harassment over recent years. When compared with the last time the survey was run in 2018, young people were now less likely to feel it was acceptable to stare or wolf-whistle at people, to make sexual comments or jokes, or to force someone to dance at a party32. However, results for other measures have remained more stable over this time (including feeling it is acceptable to hit or push a partner or to check their phone or social media), while the proportion feeling it is ok to tell a partner not to hang out with friends increased from 31% to 39%. Importantly, this suggests that improvements identified earlier were largely confined to behaviours constituting sexual harassment; in contrast, less progress appears to have been seen over this time in changing young Londoners’ attitudes towards healthy relationships or domestic abuse. 32 In 2018, 31% of young people felt it was acceptable to start or wolf-whistle at people, 23% felt it was acceptable to make sexual comments or jokes, and 23% felt it was acceptable to try to dance with someone at a party even if they don’t want to dance with you. YOUTH SURVEY 21-22 REPORT - DRAFT 35 6.3. Young people’s experiences of sexual harassment Young Londoners in school years 10 and 11 were also asked about their own experiences of sexual harassment and inappropriate sexual content. Here, half of our young cohort said someone had stared at them in a way that had made them uncomfortable during the last year (51%, 1,063 of 2,104), while just under a third said that someone had wolf-whistled them (29%, 602 of 2,110), made sexual jokes or taunts (29%, 619 of 2,102), or made sexual comments about them (29%, 609 of 2,099). These results represent a slight increase compared with the last time the survey was run in 2018 – in particular for unwanted sexual jokes and unwanted sexual comments, which are both seven percentage points higher than levels seen in the last wave. This is despite the current survey wave being conducted shortly after the COVID-19 pandemic, which inevitably limited opportunities for social contact and the use of public space for young people growing up in the capital. Gender differences were also stark (see Figure 6.4). In line with results from 2018, young females remained far more likely to say they had experienced forms of sexual harassment than young males. Illustrating this, females were over three times more likely to say they had experienced unwanted sexual comments (41%, 431 of 1,056, compared with 13% of males, 118 of 925), and were over six times more likely to have experienced wolf-whistling (46%, 486 of 1,055, compared with 7% of males, 63 of 934). Although base numbers are low for non-binary young people, exposure was also particularly high for this group. For example, 67% of non-binary young people said they had experienced unwanted sexual comments during the last year (36 of 54), while 60% had experienced sexual jokes or taunts (32 of 53). YOUTH SURVEY 21-22 REPORT - DRAFT 36 6.4. Young people’s experiences of sexual content MOPAC’s ‘We Are London’ Youth Survey 2021-22 also brought new insight into young Londoners’ exposure to other forms of sexual content, including messages and images (see Figure 6.4). Here, a quarter of young people in school years 10 and 11 said that they had received unwanted sexual messages or nude pictures (25%, 536 of 2,106), while a similar proportion said someone had asked them to send nude pictures (24%, 510 of 2,105). Large gender differences again emerged, with young females around three times more likely to say they had received or been asked to send nude pictures compared with young males. A smaller group of young people in school years 10 and 11 – 6% – said that someone had shared their nude pictures or selfies in a way that they did not want (125 of 2,119). Despite low base numbers, prevalence was particularly high here among young people identifying as a non-binary gender (20%, 11 of 56). 6.5. Understanding cumulative exposure to sexual harassment and content Using the seven measures of sexual harassment and inappropriate content measured in MOPAC’s ‘We Are London’ Youth Survey 2021-22, it was possible to calculate cumulative exposure to inappropriate sexual behaviours – i.e., how many young people had experienced more than one type of risk during the last year (see Figure 6.5).33 From this, 42% of young people (851 of 2,024) in school years 10 and 11 said that they had experienced none of the behaviours listed – leaving 58% who had experienced at least one (1,173 of 2,024). Worryingly, a quarter of young people had experienced four or more of the seven listed forms of inappropriate sexual behaviour – with a small minority (3%, 69 of 2,024) experiencing all seven (Figure 6.7). Again, the effect of gender was particularly pronounced here: females were nearly five times more likely to have experienced four or more forms of inappropriate sexual behaviour (39%, 388 of 1,005) than males (8%, 71 of 909) – and this increased even further to 62% of non-binary young people (31 of 50). In turn, results once again also revealed an overlap with a range of wider vulnerabilities, including victimisation, exploitation and violence. Indeed, such issues were particularly attenuated among the cohort of young people experiencing four or more different forms of sexual harassment or inappropriate sexual content: here, one in six said they had been threatened or hurt by someone with a knife (15%, 74 of 489) and one in eight had been pressured by someone else to carry or sell drugs (12%, 55 of 477) – levels over five times higher than those seen for young people not experiencing any of the listed safeguarding risks (at 3%, 22 of 839, and 2%, 13 of 840 respectively). In turn, only around 33 Totals calculated for young people giving valid responses (i.e. yes/no/don’t know’) across all seven measures in Figure 6.4. YOUTH SURVEY 21-22 REPORT - DRAFT 37 half of those experiencing four or more forms of sexual harassment or inappropriate content said they felt safe in the area where they live (56%, 283 of 505): significantly below all other groups. Again, this adds to the increasingly complex picture of risk and vulnerability amongst young Londoners, with those most exposed to multiple different forms of sexual harassment and inappropriate sexual content at an early age seemingly also experiencing a range of wider adversities; this is explored further in Section 7. 6.6. Seeking help for experiences of sexual harassment and inappropriate content Overall, less than a quarter of young people who had experienced any of the seven listed sexual harassment and/or inappropriate content risks during the last year said that they had spoken to anyone or sought help for their experience(s) (23%, 270 of 1,181). This is in line with the last time the Youth Survey was run in 2018.34 Although young people experiencing multiple forms of sexual harassment and inappropriate content were more likely to say they had spoken to someone or sought help, levels nevertheless remained low: 28% of those experiencing four or more forms said they had spoken to someone about their experiences (134 of 484), compared with 17% of those experiencing just one (48 of 290). Furthermore, young males were significantly less likely to say they had spoken to anyone (15%, 50 of 344) than females (26%, 196 of 759), perhaps highlighting an increased hesitancy to seek support here. Young people who had sought support were most likely to say they had spoken to friends or family (79%, 212 of 270), with one in five saying they had told someone at school or a teacher (19%, 51 of 270).35 However, far fewer said they had spoken to the police (9%, 23 of 270) or to a dedicated organisation or charity (4%, 12 of 270). Together this highlights considerable under-reporting of sexual harassment and inappropriate content among young Londoners, with those affected far more likely to turn to informal support networks such as peers and relatives than more formal channels. Consistent with the last time the survey was run, results once again revealed a range of barriers that prevented young people from seeking support (see Figure 6.6). Those who did not speak to anyone about their experiences most often said this was because they didn’t feel like they needed to (59%, 373 of 804)36. However, results also highlight a degree of normalisation of behaviours among young Londoners: many felt their experience wasn’t important or serious enough (40%, 325 of 804), happens too often (30%, 239 of 804) or that it was just normal or ‘banter’ (16%, 128 of 804). These themes also emerged in young people’s own comments. Some young people also expressed concerns about the consequences of seeking help, or felt unwilling to speak to anyone as they felt ashamed or frightened. 34 Please note that the previous Youth Survey in 2018 asked only about sexual harassment experiences, so figures are not directly comparable. 35 Respondents could select all that apply, so percentages do not add up to 100%. 36 Respondents could select all that apply, so percentages do not add up to 100%. YOUTH SURVEY 21-22 REPORT - DRAFT 38 YOUTH SURVEY 21-22 REPORT - DRAFT 39 7. Building the Picture of Risk – a focus on young people ‘most at risk’ from violence and exploitation 7.1. Defining cumulative risk Reflecting back across the topics covered by MOPAC���s Youth Survey 2021-22, it is clear that findings paint a complex picture of risk and vulnerability that pervades across many of the areas discussed. The final section of this report seeks to explore this interplay further with a specific focus on understanding the cohort of young Londoners exposed to multiple different forms of violence and exploitation. To achieve this, a composite measure was calculated to reflect the total number of risks experienced across the eight measures of exposure to serious violence or exploitation included in the survey – including vicarious and personal exposure to gangs and knife crime37. Young people were then grouped into three categories as outlined in Figure 7.1. 7.2. The prevalence of cumulative risk Positively, the majority of young Londoners said that they had not experienced ANY of the eight identified violence or exploitation risks (81%, 6,606 of 8,151) – forming our ‘low risk’ group. However, 17% of young people fell into the ‘medium risk’ category (1,412 of 8,151), while results revealed a small minority of young people – 2% - who had experienced four or more of the eight identified risks and formed our ‘high risk’ cohort (see Figure 7.1). This equated to 133 young people in total – with 32 of these saying they had experienced all eight of the identified violence and exploitation related risks. Importantly, when looking across key demographics, certain groups of young people were disproportionately more likely to fall into this ‘high-risk’ cohort (see Figure 7.2). Perhaps unsurprisingly, exposure to cumulative risk increased with age. In addition, young people identifying as a non-binary gender were around three times more likely to fall into this ‘high risk’ category (6%, 11 of 179). Particularly prominently – and building on findings in Section 5.1 – results once again also revealed an exceptionally high prevalence of exposure to multiple forms of violence and exploitation risks amongst those who had attended a Pupil Referral Unit (PRU). Indeed, within this small group over one in six young people said they had experienced four or more forms of violence or exploitation related risk 37 Calculated for young people giving valid responses (i.e. yes/no/don’t know) across the eight measures listed in Figure 5.1. YOUTH SURVEY 21-22 REPORT - DRAFT 40 (16%, 23 of 142). This is far higher than all other demographic groups, and again highlights a disproportionate concentration of risk within such specialist educational settings. 7.3. The complex picture of wider vulnerabilities In turn, further analysis revealed a range of attenuated vulnerabilities amongst young Londoners more exposed to violence and exploitation in the capital. Crucially though – and building on previous findings throughout this report – this section reveals a cumulative increase in risk; such that particularly poor outcomes are seen amongst the small cohort of young Londoners most exposed to these issues. To illustrate, when looking at safety, just 43% of this ‘high risk’ group said that they felt safe in their local area (57 of 133): significantly below levels seen either for our ‘medium’ risk (67%, 942 of 1,409) or ‘low risk’ group (82%, 5,394 of 6,600). A similar pattern was seen at school, with half of the ‘high risk’ group feeling safe here (53%, 70 of 133); compared with nearly all of those exposed to none of the violence or exploitation risks (92%, 6,069 of 6,586). Accordingly, young people in the ‘high risk’ group were around three times more likely to feel that issues such as ‘people using or dealing drugs’, ‘people carrying knives’, or ‘people joining gangs’ were a big problem where they live – and these gaps became even more pronounced in the school environment (where those in the ‘high risk’ group were around five times more likely to feel these issues were a ‘big problem’ than those in the ‘low risk’ group). In turn, it is perhaps even more concerning that over half of the young Londoners in our ‘high risk’ group for violence and exploitation said that they had personally also been the victim of a crime during the last year (55%, 66 of 119): levels twice those seen for the ‘medium risk’ group (22%, 296 of 1312) and over ten times higher than the low risk group (5%, 330 of 6392). Building on earlier findings, this once again highlights the overlap between victimisation and exposure to wider aspects of criminality – with a seemingly disproportionate impact on the small group of young Londoners potentially most exposed to multiple forms of violence and exploitation. Furthermore, this effect also pervaded across wider safeguarding issues – including into the online sphere, where our ‘high risk’ group were over twice as likely to say someone had bullied or upset them online during the last year (53%, 60 of 114) than the ‘low risk’ group (18%, 1,047 of 5,764). YOUTH SURVEY 21-22 REPORT - DRAFT 41 Although base numbers were low38, young people in our ‘high risk’ group also saw significantly increased exposure to inappropriate sexual behaviours – including jokes or taunts of a sexual nature and unwanted sexual comments. In many cases, exposure to such behaviours were also elevated amongst the ‘medium risk’ cohort; but findings were particularly stark in relation to inappropriate sexual content. Here, nearly two-thirds of our ‘high risk’ cohort said someone had sent them unwanted sexual messages or images during the last year (62%, 34 of 55) – significantly higher than levels seen both for our ‘medium risk’ (44%, 197 of 446) or ‘low risk’ (15%, 206 of 1,343) groups – while around one in three said their own nude pictures had been used or shared in a way they did not want (29%, 16 of 55). This is striking when compared with just 3% of our ‘low risk’ cohort (38 of 1,354) – and is once again also significantly above the ‘medium risk’ group (9%, 40 of 446). With this in mind, it is perhaps particularly concerning that the small group of young Londoners most exposed to multiple forms of violence and exploitation were in turn also disproportionately more likely to believe such inappropriate behaviours were acceptable. To illustrate – not only were our ‘high risk’ cohort more likely to have had their own nude images used in a way they did not want – but nearly a quarter of this group believed it was ‘always or sometimes OK’ for people to share intimate pictures of a partner without their consent (23%, 13 of 56). Again this is stark when compared with just 1% of young people in our ‘low risk’ group (14 of 1,375) and 4% of those in the ‘medium risk’ cohort (18 of 457). Furthermore, half of our ‘high risk’ group believed it was OK to insult a partner during an argument (50%, 28 of 56), while a quarter believed it was OK to physically hit or push a partner (25%, 14 of 56): demonstrating an acceptance of domestic abuse behaviours significantly higher than those seen amongst either the ‘low risk’ or ‘medium risk’ cohorts. Importantly, these findings paint a particularly complex picture of wider risks and vulnerabilities amongst young Londoners potentially most exposed to multiple forms of violence and exploitation- related risks – with potentially cumulative impacts that bring implications for effective safeguarding. 7.4. Challenges to relationships with police Building on this, results also reveal unique challenges to wider engagement with the police amongst this cohort of young Londoners. In light of the picture already established, it is perhaps unsurprising that those most exposed to multiple forms of violence and exploitation were more likely to have come into contact with the police. To illustrate, nearly all of the ‘high risk’ cohort said they had spoken to a police officer in London (83%, 45 of 5439). In turn, around half said they had specifically been Stopped and Searched by police (53%, 69 of 130) – again significantly above both our ‘low risk’ (3%, 174 of 6,591) and ‘medium risk’ (11%, 156 of 1,405) cohorts. This is in line with earlier findings in Section 4.7, and likely reflects the targeted nature of the tactic. However, in turn, it is worth noting that young people forming the ‘high risk’ cohort were also disproportionately more likely to have been stopped on multiple occasions, with 41% of this group saying they had been Stopped and Searched more than three times (28 of 69). This was significantly higher than our ‘medium risk’ (19%, 30 of 155) and ‘low risk’ (11%, 77 of 173) groups, and again illustrates the need for effective interventions to help break this cycle. However – and reinforcing the inherent challenges in this – it is perhaps particularly concerning that over half of this group of young Londoners exposed to four or more forms of violence and exploitation risk said that they had a bad overall opinion of the police (54%, 72 of 133). This pattern was reflected across wider perception measures, with young people in our ‘high risk’ cohort disproportionately 38 These questions were only asked to young people in school years 10 and 11 – representing around 60 young people in our ‘high risk’ cohort. 39 Please note low base as question was randomised to half of the survey sample. YOUTH SURVEY 21-22 REPORT - DRAFT 42 more likely to hold negative views of police when compared with both the ‘low risk’ and ‘medium risk’ groups. Taken together then, these findings also reveal a unique and particularly challenging picture of relationships between police and this small cohort of young Londoners potentially most exposed to multiple forms of violence and exploitation. Indeed, given the intelligence-led nature of police operations alongside the aforementioned array of overlapping risk-factors, it is perhaps unsurprising that this group are disproportionately more likely to come into contact with police. Inevitably, such contact also brings valuable safeguarding opportunities; but results also highlight specific challenges in trust and confidence amongst this group that may serve as a barrier to effective engagement. With all this in mind, it therefore appears particularly important to recognise that many young people coming into contact with the police in London may have experienced a range of wider adversities, and to ensure that this complex interplay of vulnerabilities is considered across these interactions. YOUTH SURVEY 21-22 REPORT - DRAFT 43 8. Conclusion MOPAC’s ‘We Are London’ Youth Survey 2021-22 brings rich insight into a range of crime and safety issues affecting those growing up in the capital. Reaching nearly 12,000 young people aged 11 to 16, the survey offers an unrivalled capture of what matters most to young Londoners, and gives the next generation a say in how their city is policed. This latest wave of MOPAC’s Youth Survey took place at a unique point in time – closely following the COVID-19 pandemic and several high-profile events affecting policing in London. Overwhelmingly though, results reveal a broadly stable picture across many topic areas when compared with the previous wave in 2018. Indeed, for many young people the picture of crime and safety seems to have remained largely static: the same issues appear at the forefront of young people’s minds, even in light of this potentially changed context. Illustrating this – and in line with the last survey wave – it is positive that most young Londoners continue to feel safe in the capital. Nevertheless, results reveal a considerable minority who feel unsafe, and suggest that young people remain affected by a range of crime issues – with many still concerned about violence, and around one in ten directly experiencing criminal victimisation themselves. Consistent with 2018, results highlight a relatively stable minority of young people personally involved in gangs and weapon possession, although it is encouraging that fewer now say they know others involved in such behaviours. More widely, results again reveal potentially concerning normalisation of inappropriate sexual behaviours among young people, while experiences of sexual harassment remain prevalent among our young cohort, even in spite of recent COVID-19 ‘lockdowns’. Importantly though, MOPAC’s ‘We Are London’ Youth Survey 2021-22 also reveals some emerging areas of concern. Compared with the last wave, issues relating to hate crime now appear more prominent on young people’s minds. Indeed, this now forms the top ‘crime problem’ for young people both at school and locally, while the proportion of young victims saying they have personally experienced a hate crime has more than doubled over this timeframe. Although this may in part reflect improvements in young people’s understanding of what constitutes hate, findings nevertheless highlight this as an emerging issue for young Londoners. In addition, new insight from the survey also reveals the prevalence of discriminatory content online, in terms of online bullying as well as exposure to content promoting hate across social media platforms. Furthermore, results from this wave of MOPAC’s Youth Survey also highlight emerging risks to young people’s wider views and relationships with the police. In line with the 2018 wave, early opinions of the police remain mixed, while trust and confidence stand below levels seen for adult Londoners. Importantly though, young Londoners’ opinions of police have also become more negative since 2018, in particular among the capital’s young females. Compounding this, significant declines have also been seen in the proportion believing police ‘treat everyone fairly’, with just one in ten young Black Londoners now agreeing with this statement. Moreover this picture is not unique to youth: MOPAC’s Public Attitude Survey also reveals a downwards trend in perceptions over this timeframe among adult Londoners, once again more pronounced for younger females (MOPAC, 2022d). Despite this, a key limitation of the current survey is that it remains a ‘snapshot’ capture of young Londoners’ views at a specific point in time. It is possible that recent events – including the Black Lives Matter protests and the murder of Sarah Everard by a serving police officer – may have been particularly influential in driving these declines, and the lasting impact of such events on young YOUTH SURVEY 21-22 REPORT - DRAFT 44 Londoners remains unknown. To this end, future waves of MOPAC’s Youth Survey will be invaluable in continuing to monitor these trajectories on an ongoing basis. Nevertheless, our findings here highlight young people’s perceptions of the police as a key area of emerging risk, and emphasise the importance of work to build positive relationships between the police and young people. Related to this, MOPAC’s ‘We Are London’ Youth Survey 2021-22 also brings new and innovative insight across a range of novel topics. One such area includes identifying the diverse range of situations that bring young people into contact with police officers – as victims, witnesses, vulnerable people, and through school engagement – with results here highlighting the particular value of informal interactions in helping to build positive relationships between young people and the police. This stands atop findings that emphasise the importance of procedural justice during police encounters – treating young people politely and with respect – in helping to maintain good opinions of the police. Encouragingly, we see that the majority of young people also support partnership working between police and schools, again highlighting opportunities for positive engagement between young Londoners and the Metropolitan Police Service. Finally – and perhaps most importantly – MOPAC’s ‘We Are London’ Youth Survey 2021-22 also brings new insight into the complex interplay of risk and vulnerability affecting many young people growing up in the capital. Adding to the existing evidence base, we identify a considerable minority of those aged 14 to 16 who have had their images shared with others in a way they did not want, and reveal extensive cumulative exposure to sexual harassment and inappropriate content among the young cohort. Looking specifically at youth violence, results newly identify a small but potentially vulnerable group of young Londoners exposed to elements of criminal exploitation – including being asked to carry knives or sell drugs on behalf of others. Once again, the survey reveals an overlap between exposure to such elements of youth violence and an array of wider harms, including increased criminal victimisation, lower feelings of safety, poorer perceptions of the police, and normalisation of potentially harmful sexual behaviours. In turn, results also show that this potentially vulnerable cohort are also disproportionately more likely to come into contact with the police – including through Stop and Search – once again reinforcing the value and opportunity for early interactions with officers to help safeguard and support young Londoners. On top of this, new analyses reveal a small cohort of young Londoners exposed to multiple different forms of violence and exploitation – for whom many of the risks outlined above are disproportionately further attenuated – and emphasise some unique challenges to effectively engaging with this particularly vulnerable group. To conclude, MOPAC’s ‘We Are London’ Youth Survey 2021-22 provides a robust evidence base across a range of crime and policing issues affecting young Londoners at a unique – and potentially pivotal – point in time. Results from this survey will continue to be disseminated across key partners and stakeholders to provide a catalyst for ongoing improvements to service delivery, and to ensure the views of young Londoners are heard in the decisions that affect them. YOUTH SURVEY 21-22 REPORT - DRAFT 45 References Aitkenhead, E., Clements, J., Lumley, J., Muir, R., Redgrave, H. & Skidmore, M. (2022). Policing the pandemic. The Police Foundation. https://www.police-foundation.org.uk/2017/wp- content/uploads/2010/10/policing_the_pandemic_final.pdf Beckley, A., Caspi, A., Arsenaeault, L., Barnes, J., Fisher, H., Harrington, H., Houts, R., Morgan, N., Odgers, C., Wertz, J. & Moffitt, T. (2018). The developmental nature of the victim-offender overlap. Journal of Developmental and Life-Course Criminology, 4, 24-49. Bradford, B., Jackson, J., & Stanko, E. (2009). Contact and confidence: revisiting the impact of public encounters with the police. Policing and society, 19(1), 20-46. Charleton, Harrison & Dawson (2015). YOUth Matter! Listening to the voice of young London. https://www.london.gov.uk/sites/default/files/youth_matter_report_final_version.pdf Children’s Commissioner (2017). Defining child vulnerability: definitions, frameworks and groups. Technical paper 2 in Children’s Commissioner project on vulnerable children. Cordis Bright. https://www.childrenscommissioner.gov.uk/wp-content/uploads/2017/07/CCO-TP2-Defining- Vulnerability-Cordis-Bright-2.pdf Dempsey, M. (2021). Still not safe: The public health response to youth violence. London: Children’s Commissioner. https://www.childrenscommissioner.gov.uk/wp- content/uploads/2021/02/cco-still-not-safe.pdf Early Intervention Fund (2015). Preventing gang and youth violence: A review of risk and protective factors. Cordis Bright/Home Office. https://www.eif.org.uk/report/preventing-gang-and- youth-violence-a-review-of-risk-and-protective-factors End Violence Against Women Coalition & YouGov (2021). YouGov/EVAW Survey Results [data tables]. https://www.endviolenceagainstwomen.org.uk/wp- content/uploads/2021/11/EVAW_SexualViolence_211028-1.xls-Compatibility-Mode.pdf Gordon, F. (2021). Online harms experienced by children and young people: acceptable use and regulation. https://static1.squarespace.com/static/5d7a0e7cb86e30669b46b052/t/618b7cd8b5872f4721 c9d59a/1636531420725/Online+Harms+Research+November+2021+-+Full+Report.pdf Greater London Authority (2022, 28 June). 2021 Census: First release. https://airdrive-secure.s3-eu- west-1.amazonaws.com/london/dataset/2021-census-first-release/2022-06- 29T12%3A38%3A53/2021%20census%20first%20release.pdf?X-Amz-Algorithm=AWS4-HMAC- SHA256&X-Amz-Credential=AKIAJJDIMAIVZJDICKHA%2F20221003%2Feu-west- 1%2Fs3%2Faws4_request&X-Amz-Date=20221003T144346Z&X-Amz-Expires=300&X-Amz- Signature=e4f35915ca94cc52de2f5a2d1122f02f8fa18e35267e37ba8c3773c9aed09317&X- Amz-SignedHeaders=host Jackson, L. & Smith, L. (2013). Research into young Londoners’ experiences and perceptions of Stop and Search. https://www.london.gov.uk/sites/default/files/gla_migrate_files_destination/14- 02-06-OPM%20-%20Young%20people%27s%20views%20stop%20and%20search%20- %20FINAL.pdf YOUTH SURVEY 21-22 REPORT - DRAFT 46 Kincaid, S., Lumley, J., and Corlett, M. (2021). Violence and Vulnerability. Crest. https://64e09bbc- abdd-42c6-90a8- 8992ce46e59.usrfiles.com/ugd/64e09b_6a5b47a71f744e3e92e2d82164dfeea6.pdf Leaders Unlocked (2020). Policing the Pandemic: Exploring young people’s experiences and recommendations. http://leaders-unlocked.org/luwp/wp- content/uploads/2020/07/PolicingPandemic_v1.pdf Leaders Unlocked (2021). IOPC Youth Panel: Findings and recommendations. https://www.policeconduct.gov.uk/sites/default/files/Documents/research- learning/Youth_Panel_report_2021.pdf Metropolitan Police Service (2022). Crime Data Dashboard [online dashboard]. https://www.met.police.uk/sd/stats-and-data/met/crime-data-dashboard/ Metropolitan Police Service (2022b). Hate Crime or Special Crime Dashboard [online dashboard]. https://www.met.police.uk/sd/stats-and-data/met/hate-crime-dashboard/ Metropolitan Police Service (2022c). Stop and Search Dashboard [online dashboard]. https://www.met.police.uk/sd/stats-and-data/met/stop-and-search-dashboard/ MOPAC (2020). Public Voice Reporting Pack Q1 20-21. Internally published document. For further information please email evidenceandinsight@mopac.london.gov.uk. MOPAC (2021, August 21). MOPAC Q1 Report 2021-22: MOPAC Q1 Public Voice. https://www.london.gov.uk/mopac-publications-0/mopac-q1-report-2021-22. MOPAC (2022a, 23 May). MOPAC Q4 Reports 2021-22: MOPAC Q4 Public Voice. https://www.london.gov.uk/mopac-publications-0/mopac-q4-reports-2021-22. MOPAC (2022b, 23 May). Understanding Londoners’ Trust in the Police – Qualitative Analysis. https://www.london.gov.uk/sites/default/files/understanding_londoners_trust_in_the_police _-_qualitative_analysis_v2.pdf. MOPAC (2022c). A problem profile of violence, gangs and young people. https://www.london.gov.uk/sites/default/files/serious_youth_violence_problem_profile.pdf MOPAC (2022d, 21 February). MOPAC Q3 Reports 2021-22: MOPAC Q3 Public Voice. https://www.london.gov.uk/mopac-publications-0/mopac-q3-report-2021-22. MOPAC (2022e, 02 December). MOPAC Trust and Confidence Dashboard. https://data.london.gov.uk/mopac-pcp-dashboard/increase-trust-and-confidence-dashboard/ Murray, K., McVie, S., Farren, D., Herlitz, L., Hough, M. & Norris, P. (2020). Procedural justice, compliance with the law and police stop-and-search: A study of young people in England and Scotland. Policing and Society, 31(3), 263-282. National Youth Agency (2021). Between the lines. https://static.nya.org.uk/static/f3fcc0c77f1f2d3b579af6274648540b/Between-the-lines-final- version.pdf YOUTH SURVEY 21-22 REPORT - DRAFT 47 NSPCC (2020). Social isolation and the risk of child maltreatment, in lockdown and beyond. https://learning.nspcc.org.uk/media/2246/isolated-and-struggling-social-isolation-risk-child- maltreatment-lockdown-and-beyond.pdf NSPCC (2020, October 7). Record numbers contact the NSPCC with concerns during lockdown. https://www.nspcc.org.uk/about-us/news-opinion/2020/2020-10-07--helpline-recovery-plan/ Office for National Statistics (2020). Crime in England and Wales, year ending March 2020 – Appendix tables. Retrieved from: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/datasets/crimeine nglandandwalesappendixtables Partnership for Young London (2021). Mapping young London: a view into young Londoners after a year of lockdown. https://www.partnershipforyounglondon.org.uk/_files/ugd/3532bf_24ea9bda6d6f45c498b09 8d5b3a33a64.pdf Public Health England (2021). The impact of COVID-19 on London’s children and young people. https://www.bacaph.org.uk/images/PDFs/The%20impact%20of%20COVID- 19%20on%20London's%20CYP.pdf Ramshaw, Charleton & Dawson (2018). Youth Voice Survey 2018. MOPAC Evidence & Insight. https://www.london.gov.uk/sites/default/files/youth_voice_survey_report_2018_final.pdf Skarlatidou, A., Ludwig, L., Solymosi, R. & Bradford, B. (2021). Understanding knife crime and trust in police with young people in East London. Crime & Delinquency [online] https://doi.org/10.1177%2F00111287211029873 Skogan, W. (2007). Asymmetry in the impact of encounters with police. Policing and Society, 16, 99- 126. Stanko, E., Jackson, J., Bradford, B., and Hohl, K. (2012) A golden thread, a presence amongst uniforms, and a good deal of data: studying public confidence in the London Metropolitan Police. Policing and Society: An International Journal of Research and Policy, 22(3), 317-331. Sunshine, J. & Tyler, T. (2003). The Role of Procedural Justice and Legitimacy in shaping public support for Policing. Law & Society Review, 37(3), 513-548. Tyler, T. (2003). Procedural justice, legitimacy, and the effective rule of law. Crime and Justice, 30, 283- 357. Villadsen, A. and Fitzsimons, E. (2021). Carrying or using a weapon at age 17: Evidence from the UK Millennium Cohort Study. London: Centre for Longitudinal Studies. YOUTH SURVEY 21-22 REPORT - DRAFT 48 Appendix 1: Survey sample demographics. YOUTH SURVEY 21-22 REPORT - DRAFT 49 GPS Tagging Interim Evaluation Report September 2017 MOPAC Evidence and Insight 1. Introduction Background to the Pilot In early 2016 the Ministry of Justice (MOJ) announced its intention to pilot the use of GPS tags. The MOJ agreed to work with the Mayor’s Office for Policing And Crime (MOPAC) to test the use of mandatory GPS tagging for prolific offenders through the Gripping the Offender (GtO) pilot. Although GPS tags have been used on a voluntary basis in several small scale pilots across London, this is the first time mandatory use has been piloted. An electronic monitoring requirement has been available in Thames Magistrates Court and Highbury Corner Magistrates Court and the local Crown Courts since 20 March 2017. Two dedicated pre-sentence report writers from the National Probation Service based at Thames and Highbury have the option to recommend GPS tags as a requirement of a Community Order or Suspended Sentence Order. It is envisaged that the electronic monitoring requirement is aligned with other rehabilitative measures (e.g., a Rehabilitation Activity Requirement). The main aims of the MOPAC GPS tagging pilot are to: 1. Increase compliance with the requirements of a Community Order or Suspended Sentence Order 2. To offer sentencers an alternative to custody 3. To support crime detection and/or the apprehension or prosecution of offenders by providing information on an offender’s location at a specific date and time Up to 100 prolific offenders in the 8 GtO pilot boroughs1 will be tagged during the 12 month pilot. The Evidence and Insight team at MOPAC have been asked to evaluate the proof of concept pilot, capturing learning around the implementation and design of the pilot, gaining insight into the experiences of practitioners and offenders, and measuring success. GPS Tagging Evaluation The evaluation of the GPS tagging pilot forms part of the wider GtO evaluation. The overarching aims of the evaluation are to explore: Performance Monitoring the key performance indicators, including numbers receiving the tag, types of orders, compliance rates, violations, breaches, order completions. Process Understanding the experiences of practitioners and offenders involved in the pilot through surveys, interviews and focus groups. 1 Camden, Enfield, Hackney, Haringey, Islington, Newham, Tower Hamlets, Waltham Forest Indicative Impact A robust impact evaluation is not possible due to the projected sample size. The evaluation will capture indicators of success (e.g., compliance rates) and will use qualitative methods to gain learning about the success of the pilot. GPS tagging will be included in the wider impact evaluation of GtO. Interim Report This report shares findings and learning from the first six months of the GPS tagging pilot. It consists of performance and process aspects, including feedback from offender managers, magistrates and police officers, and initial feedback from tag wearers. 2. Results Performance GPS tags imposed • Between 20 March and 31 August 2017 a total of 25 GPS tags were imposed by the courts (18 at Highbury Corner and 7 at Thames): an average of 1 tag per week (see Figure 1). Figure 1. GPS tags imposed per week of the pilot • A total of 36 GPS tags were recommended as sentencing options by the dedicated report writers based at these courts, which equates to a success rate of 69% (25 tags imposed of the 36 recommended). • Of the GPS tags imposed, 11 were imposed as a requirement of a Community Order and 14 as part of a Suspended Sentence Order. Alongside the electronic monitoring requirement, all orders also included a Rehabilitation Activity Requirement, and a 0 1 2 3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Number of tags Week of Pilot handful also included an Alcohol Treatment Requirement or a Drug Rehabilitation Requirement. • The majority of tags were 24 hour whereabouts (92%; n=23), with two including a restrictive requirement (i.e., an Exclusion Zone). • The average length of the tag order was just over 3 months (range 1- 6 months). • Tags were imposed for a range of offences, including theft (29%; n=7), violence (17%; n=4), public order (13%; n=3), burglary (8%; n=2), criminal damage (8%; n=2), breach (4%; n=1), driving related (4%; n=1) and drug possession/supply (4%; n=1). • The vast majority of cases were tagged at court on the day of sentencing (92%; n=23). Compliance • Of the 25 GPS tags imposed, only 4 Points of Interest have been set up by offender managers. No Inclusion Zones have been set up. • In total, there have been 349 violations of the GPS tag, the majority for ‘No Communication’, ‘No Location’ and ‘Low Battery’ (see Table 1). In addition, there were 13 violations of an Exclusion Zone. The number of violations has steadily increased over time, reflecting the increase in the number of people subject to a tag. • A total of 389 confirmed alerts have been escalated to offender managers. Table 1. Types of GPS tag violations Type of Violation April ‘17 May ‘17 June ‘17 July ‘17 Aug ‘17 Total Low Battery 6 9 20 41 33 109 No Communication 10 15 22 40 35 122 No Location 10 18 21 30 26 105 Exclusion Zone 0 10 0 1 2 13 Total 26 52 63 112 96 349 • The violations reported to the offender manager resulted in six breaches of the tag order, the majority for not charging the tag. • As of 31 August 2017, five offenders had successfully completed their tag order and seven orders had been revoked. Orders were revoked for a combination of non- compliance with the GPS tag or other order requirements and for committing new offences. Process Findings Practitioner feedback Feedback was captured from a range of practitioners, including offender managers, police and magistrates. Initial views of GPS tagging Prior to the pilot going live, a number of training sessions were held with practitioners. Overall, feedback collected from these sessions revealed initial positive views about the pilot.2 The 2 Based on feedback from 16 probation officers, 32 magistrates and 17 police officers (total n=65). majority of people who attended the trainings agreed: they would use GPS tagging (92%; n=60); that they see GPS tagging as a useful way to tackle prolific offending (87%; n=57); that they understand the aims and objectives of GPS tagging (83%; n=54); and that they understand how to use GPS tagging in their role (86%; n=56). Magistrates who attended the training sessions agreed they would feel confident imposing a GPS tag (83%; n=27) and police officers agreed it would be a useful additional tool in their role (88%; n=15). Uses of GPS tagging A survey was sent to practitioners in August 2017, five months after the pilot went live. A total of 20 practitioners responded to the survey: 11 probation staff and 9 practitioners from local IOM teams (i.e., police officers, local authority staff). Furthermore, a questionnaire was administered to offender managers in late – early September 2017 (n=7). Respondents were asked to rank what they believed to be the most important uses of GPS tagging. The highest ranked use of the tags was “as an alternative to custody” followed by “to monitor compliance with exclusion zone or other restrictive requirement”. Other highly ranked uses included “to deter service users from reoffending” and “to use location data to match with potential new offences”. Offender managers were asked how they are using the tags as part of their work with offenders. Of the 16 offender managers surveyed, 12 had experience managing an offender on a GPS tag. Most offender managers had some awareness of the tag, but only as a requirement to be enforced rather than something that could be used proactively to manage cases. To illustrate, only one offender manager reported having logged on to Buddi’s Eagle platform and none reported having set up an Inclusion Zone or Point of Interest. When asked why they had not set up an Inclusion Zone or Point of Interest, the most common reasons given were: they were not aware that they could, they did not know how to, or that IT problems were causing issues. Positive learning and challenges Positive themes that emerged from offender manager feedback were: the good communication with the GPS tagging provider Buddi; and the easy access to Buddi’s Eagle system (although only one offender manager reported having logged on to the Buddi system). In addition, a number of positive case examples were given. For example, one offender manager reported an example of a tag wearer being more conscious of his alcohol problem and the need to avoid certain pubs due to the GPS tag. Another example was given of the tag providing the offender manager information about where the tag wearer spends his time in order to protect a vulnerable victim. Despite these positive case examples, offender managers mentioned a number of areas for improvement, including: charging issues undermining several cases; a lack of knowledge about how to make use of GPS tags for Inclusion Zones or Points of Interest; a lack of knowledge about GPS when being allocated cases; cases being given a tag who were not suitable (i.e., chaotic cases with severe learning disabilities and mental health issues); the need for the whole GtO system to support the tag (i.e., Police and partner agencies) not just the responsible officer; and the clash between needing the offender to be on the tag for a long time to bring about behaviour change and the unwillingness of the Court to impose long tags because it was felt to be too punitive. To address the above issues, additional training was delivered to offender managers in August – September 2017. The training was designed to give offender managers information on how to make full use of the GPS tag. Following the training, offender managers reported a much clearer understanding of how they can use GPS tags to manage their cases, including how to use Inclusion Zones and Points of Interest and how and when they can share GPS tag data with partner agencies.3 Ongoing training and support will be offered to offender managers to enable them to integrate the management and information received from the tag into their existing offender management and risk reduction plans for each case. Offenders A questionnaire is administered to offenders at the time of fitting and removing the GPS tag. The survey is voluntary and some choose not to respond. In total, 9 fitting surveys and 1 removal survey were completed at the time of writing this report. Only findings from the fitting survey are reported here. At the time of fitting the tag, offenders were asked about their thoughts on being given a GPS tag and their expectations of the order. Findings showed that: - All offenders understood why they had received the GPS tag - All understood how the GPS equipment works - All understood what they must do to comply with the tag order - All understood how the tag fits in with their other order requirements - Most (6/9) stated they were worried about what their friends and family would think of the tag - All were fairly or very confident they would successfully complete the tag order Offenders were also asked about their expectations of how the tag will affect various aspects of their life: - All thought it would make their life in general better - All thought it would lead to them committing less offences - 5 out of 9 tag wearers thought the tag would improve relationships with their family - 4 out of 9 tag wearers thought the tag would improve their ability to go out and socialise 3 Based on feedback collected from 8 offender managers who received training. “My personal experience is that it was issued to an offender where it was not really required or necessary” “Good communication from tagging team and support staff in the field” - The majority of tag wearers thought the tag would have no impact on: their relationship with friends; their financial situation; their physical health and mental wellbeing; they housing situation; and their employment situation. 3. Discussion Since the GPS tagging pilot went life on 20 March 2017, a total of 36 electronic monitoring requirements have been proposed by the dedicated pre-sentence report writers based at Thames and Highbury Corner Magistrates Courts’ and 25 tags have been imposed: a success rate of 69%. All electronic monitoring requirements have been imposed alongside a rehabilitative requirement, consistent with the intended use of the tags as a package of support, monitoring and punishment. Overall, there were positive initial views on the GPS tagging pilot from offender managers, police officers and magistrates and the most popular use of the tag was that it provided an alternative to custody. Feedback collected from offender managers suggested more information was needed around how GPS tags could be used to manage their cases. Very few offender managers had set up an Inclusion Zone or Point of Interest on Buddi’s Eagle system due to a lack of awareness about the full use of the tags. Recent training to address these issues has led to an increased awareness from offender managers. Offenders themselves had positive expectations of the tag. They were confident they would successfully complete the tag order and knew what they needed to do to comply. Furthermore, all tag wearers thought the tag would have a positive impact on their life and lead to them to committing less crime. Overall, findings from the first six months of the pilot show that GPS tagging has been well received by practitioners and offenders’ expectations of being subject to a tag order are mostly positive. Steps are currently being taken by the implementation team to address the issues identified in this report, including more training and awareness raising for offender managers and ensuring the suitability of cases given a tag. The next interim evaluation report is scheduled for March 2018. “Wearing this tag is going to help me stop driving illegally” “My drug and alcohol addiction is getting worse so anything to help stop me offending will be welcomed. I hope to have my own family in the future so am positive with help I can get back on track” 1 GPS Tagging First Year Interim Report July 2018 MOPAC Evidence and Insight 2 Contents Executive Summary ......................................................................................................... 3 1. Introduction ......................................................................................................................... 5 Background to the Pilot .......................................................................................................... 5 Location Data Available on Pilot .............................................................................................. 6 GPS Tagging Evaluation .......................................................................................................... 7 2. Methodology ........................................................................................................................ 8 3. Results .................................................................................................................................. 9 Performance ............................................................................................................................ 9 Practitioner Feedback ............................................................................................................ 14 Tag Wearer Feedback ............................................................................................................ 18 4. Discussion ........................................................................................................................... 23 3 Executive Summary In 2016 the Ministry of Justice (MOJ) agreed to work with the Mayor’s Office for Policing And Crime (MOPAC) to test the use of mandatory GPS tagging for prolific offenders through the Persistent Offender Programme (POP). The main aims of the MOPAC pilot are to test whether GPS tagging: increases compliance with the requirements of a Community Order (CO) or Suspended Sentence Order (SSO), offers sentencers an alternative to custody, and supports crime detection and/or the apprehension or prosecution of offenders by providing information on an offender’s location at a specific date and time. Consequently, an electronic monitoring requirement has been available in Thames Magistrates’ Court and Highbury Corner Magistrates’ Court and the local Crown Courts in London since 20 March 2017. Two dedicated pre-sentence report writers from the National Probation Service based at Thames and Highbury have the option to recommend GPS tags as a requirement of a CO or SSO. Under the POP pilot, allocated probation Offender Managers (OMs) can access and monitor location data for their cases. Where the tag wearer is the subject of an SSO the police can also undertake crime mapping, matching location data with crime data where it is relevant and justified for the detection of crime. Between 20 March 2017 and 12th March 2018, a total of 73 GPS tags were imposed by the judiciary (54 at Highbury Corner Magistrates’, 18 at Thames Magistrates’, 1 at Snaresbrook Crown Court): an average of 1.4 tags imposed per week. This figure is just below the minimum target of 75 tagged offenders in the first year of the pilot. The success or concordance rate (i.e. the % of GPS tags recommended by the PSR writers which were subsequently imposed by the judiciary) over the first year was 70%, one percentage point higher than the figure reported in the period of the first interim report (the first five months of the pilot). Of the GPS tags imposed, 39 were imposed as a requirement of a CO and 34 as part of an SSO. Nearly all the tags were 24-hour whereabouts (95%; n= 69), with four including a restrictive requirement. Tags were imposed for a range of offences, including: theft (37%), burglary (14%), violence (12%), and public order (10%). The majority of tag wearers were Male (88%) and aged under 40 years (78%). Just over half of the sample was White (53%), followed by Black (21%) and Asian (15%). Of the 73, 41 cases had completed successfully, while 33 had been unsuccessful, breached and revoked for either a failure to comply and/or imprisonment following further offending (a compliance rate of 56%). Lack of attendance and failure to charge the tag were the two most common reasons for non- compliance. There were few measurable differences between the compliant and non-compliant tag wearers; however, non-compliant wearers were slightly older than compliant wearers and had significantly lower levels of motivation. For both groups, the most prevalent risk factor displayed was drug use, followed by mental health. Findings from surveys undertaken with Offender Managers and tag wearers, and interviews with those involved in the production of Pre-Sentence Reports (PSRs) during the first 12 months of the pilot suggest GPS tagging has been well received by practitioners and tag wearers alike. Feedback from the PSR staff suggested that the judiciary were generally receptive to recommendations made for tagging requirements, and were happy to adopt these recommendations, although there was a perception that judges’ understanding of GPS tagging was better than magistrates’. There was a general feeling that there was scope for greater use of GPS tagging, and for its use with higher risk offenders, but that current staffing levels – at the time of interview there was only one PSR writer and one administrative support to service both courts – acted as a barrier to this being realised. 4 In comparison to E&I’s earlier evaluation report (produced after 6 months), it appears there is greater awareness of the GPS tag on the part of Offender Managers (at least, those that completed the survey), and willingness to use interest and exclusion zones as part of supervision. As with the first interim report, offenders had positive expectations of the tag. They were confident that they would successfully complete the tag order and knew what they needed to do to comply. Furthermore, tag wearers thought the tag would have a positive impact on their life and lead to them committing less crime. At tag removal, the majority of tag wearers surveyed (admittedly a small number, n=22) thought being on the tag had improved their offending behaviour and had made their life better in general. As in the earlier report, concerns were expressed (by tag wearers and practitioners) about the difficulties caused for this cohort by the requirements to keep the tag charged. Early non-compliance by tag wearers, and the level of lost equipment, suggest there are issues around the identification of suitable individuals to be tagged that need to be addressed. Other challenges that have emerged are the repercussions arising from the changes to PSR writing policing around making recommendations for SSOs in April 2018, which resulted in a reduction of SSOs with tagging requirements, and a reduction in eligible cases for the crime mapping strand of the pilot (which can only be undertaken on SSO tag wearers). Question marks about the effectiveness of the crime mapping strand in its current form are also raised, particularly given the low number of ‘hits’ between tag wearers’ location data and police crime data. Nevertheless, the overall compliance rate for the scheme, and the positive views expressed by practitioners and tag wearers about their experience with the GPS tag, provides grounds for optimism. The next evaluation report is scheduled for Summer 2019 following the completion of the pilot’s extension. This report will provide updated results for the final GPS cohort, looking at issues around compliance, as well as reporting back on the experiences of tag wearers, probation and court staff. 5 1. Introduction Background to the Pilot In early 2016 the Ministry of Justice (MOJ) announced its intention to pilot the use of GPS tags. The MOJ agreed to work with the Mayor’s Office for Policing And Crime (MOPAC) to test the use of mandatory GPS tagging for prolific offenders through the Persistent Offender Programme (POP) pilot, starting in March 2017. Although GPS tags had been used on a voluntary basis in several small-scale pilots across London, this was the first-time mandatory use had been piloted with such a cohort. Initially running for 12 months, MOPAC has successfully applied to the MOJ to have the statutory instrument for this pilot extended for an additional 12 months, until March 2019. Under the POP pilot, offender whereabouts and location data can be accessed by the allocated probation Offender Manager (OM) as part of an Electronic Monitoring Requirement imposed under a Community Order (CO) or Suspended Sentence Order (SSO). The routine use of location data as part of offender management practice is a new development and MOPAC have invested time, training and innovation to build momentum around this in local practitioner working cultures. New tools available to probation OMs include bespoke email notifications highlighting key locations visited by the offender in real time, heat maps and “Top 5 locations of the week” reports, so that patterns of behaviour associated with risk can be uncovered, explored and challenged. Where the offender is the subject of an SSO the police can also undertake crime mapping, matching location data with crime data where it is relevant and justified for the detection of crime. There is also the capacity for the police to make external requests for location data from the GPS Tag operator (Buddi) in relation to specific crime incidents. In the first year of operation there were 13 requests made, 6 of which were granted. Police have been active supporters of using GPS with cases to manage their risk and integrate location data into existing containment strategies. An electronic monitoring requirement has been available in Thames Magistrates’ Court and Highbury Corner Magistrates’ Court and the local Crown Courts since 20 March 2017. Two dedicated pre-sentence report (PSR) writers from the National Probation Service (NPS) based at Thames and Highbury have the option to recommend GPS tags as a requirement of a CO or SSO.1 It is envisaged that the electronic monitoring requirement is used in conjunction with other rehabilitative measures (e.g., a Rehabilitation Activity Requirement). The main aims of the MOPAC pilot are to test whether GPS tagging: 1. Increases compliance with the requirements of a Community Order or Suspended Sentence Order; 2. Offers sentencers an alternative to custody; and 1 Since April 2018, PSR writers have been no longer able to recommend SSOs. 6 3. Supports crime detection and/or the apprehension or prosecution of offenders by providing information on an offender’s location at a specific date and time. Figure 1. Outline of GPS tagging process Location Data Available on Pilot The type of location data available to the various agencies involved in the pilot differs depending on the type of order made. Community Orders with Exclusion Zones or residence requirements • Probation only get alerted if the offender breaches the exclusion zone or residence requirement. • The police have no access to location data (apart from External Agency Requests (EARs)). Community Orders with 24/7 whereabouts monitoring • All the location data captured can be shared with the probation OMs. Alerts can be set against inclusion zones, points of interest (specific addresses) or interest zones. Requests for other data can be made via a secure email sent directly to Buddi - heat maps, top location reports etc. Sentencing Tagging Induction Case Management Completion - Whenever possible the subject is tagged at court. Data sharing and consent forms processed - Specialist author recommends GPS requirement in pre- sentence report - Judiciary decide whether to impose GPS under electronic monitoring requirement as part of Community or Suspended Sentence Order – 24/7 or restrictive - Subject meets with their OM in NPS or CRC, GPS tag rules are reinforced - OM creates GPS zones and monitoring set up - OM uses location data to monitor and challenge offender & manage risk - Enforcement action is taken when required - Tag is removed - OM and subject review time on GPS requirement 7 • Other agencies are involved in helping to suggest monitoring locations and support compliance i.e. drug workers, police Intel Analyst, CBO exclusion zones. • The police have no access to location data (apart from EARs). Additional functions available for Suspended Sentence Orders • The police will carry out automatic crime mapping and may be alerted to location data that matches a crime, so they can investigate further. They can also make EARs. GPS Tagging Evaluation The Evidence and Insight team at MOPAC have been asked to evaluate the GPS pilot, capturing learning around the implementation and design of the pilot, gaining insight into the experiences of practitioners and offenders, and measuring success. The evaluation of the GPS tagging pilot forms part of the wider POP evaluation. The overarching aims of the evaluation are to examine: • Performance - monitor the key performance indicators, including numbers receiving the tag, types of orders, compliance rates, violations, breaches, order completions. • Process - understand the experiences of practitioners and offenders involved in the pilot through surveys, interviews and focus groups. • Indicative Impact - a robust impact evaluation is not possible at this point due to the small sample size. The evaluation will capture indicators of success (e.g., compliance rates) and will use qualitative methods to gain learning about the success of the pilot. Interim Report This report shares findings and learning from the first 12 months of the GPS tagging pilot. It consists of performance and process aspects, including feedback from offender managers, pre- sentence report staff, and tag wearers. 8 2. Methodology Performance Performance data was gathered from a range of sources, including the PSR writers, the tagging provider Buddi, the MPS analyst responsible for crime mapping, and from MOPAC delivery staff. This data included information about the tag wearers and the types of orders they were on, compliance with the tag, and data on the crime mapping function. Interviews with Pre-Sentence Report Staff Interviews were conducted with two members of staff involved in the production of pre-sentence reports (PSRs) at Highbury Corner and Thames Magistrates’ Courts in May 2018. The interviews were designed to capture their experiences of the GPS tagging pilot, including how the pilot has been received by the courts and any challenges they have faced. Face-to-face interviews were conducted lasting approximately one hour. Interviews were recorded and later transcribed. Interviewees were asked to sign a consent form outlining confidentiality and anonymity. Offender Manager Surveys A survey was sent to offender managers (OMs) who were responsible for managing GPS cases in late February 2018. This survey was a follow-up to a similar survey sent to OMs in August 2017, the results from which were contained in the first interim report. The survey asked OMs about their experiences working with GPS cases, their perceptions of the tag, and how they have used the tag to manage their cases. At the time of writing, responses had been received from nine OMs. Tag Wearer Surveys Tag wearers were asked to complete a short survey at the point the tag was fitted, and again when it was removed. The surveys aim to collect their anticipated and actual perception of the impact of wearing the tag on their offending behaviour, their relations with friends and family, and their lifestyle more generally. At the time this report was written data were available from 22 individuals where a tag had been fitted, and 12 cases where the tag had been removed. It should be noted that this is a small data set and is unlikely to be representative of the GPS tagging cohort overall because it comprises compliant offenders (both at the point of fitting and removal). Surveys were administered by the tag fitter (a staff member from Buddi). 9 3. Results Performance GPS tags imposed Between 20 March 2017 and 12th March 2018, a total of 73 GPS tags were imposed by the judiciary (54 at Highbury Corner Magistrates’, 18 at Thames Magistrates’, 1 at Snaresbrook Crown Court): an average of 1.4 tags imposed per week. The success rate (i.e. the % of GPS tags recommended by the PSR writers which were subsequently imposed by the judiciary) over this whole period was 70%, one percentage point higher than the figure reported in the period of the first interim report. Of the 73 GPS tags imposed, 39 were imposed as a requirement of a Community Order and 34 as part of a Suspended Sentence Order. Alongside the electronic monitoring requirement, most orders imposed also included a Rehabilitation Activity Requirement, and 16 also included an Alcohol Treatment Requirement or a Drug Rehabilitation Requirement. The average length of the tag order was just over three months (range 1-6 months). The majority of GPS tags imposed were 24-hour whereabouts (95%, n=69), which means the tag wearer’s location data from any time or location can be accessed by the Offender Manager where this is relevant to the management of the case. Four tags were imposed with a restrictive requirement (e.g. data only available in relation to a requirement not to go to a particular place or location). Tags were imposed for a range of offences, including theft (37%, n=27, 3 of which were motor vehicle related), burglary (14%, n=10), violence (12%, n=9), public order (10%, n=7, 2 of which were racially aggravated), drug possession/supply (8%, n=6), criminal damage (7%, n=5), breach (7%, n=5), possession of an offensive weapon (3%, n=2), harassment (1%, n=1) and Other offence (dangerous dog) (1%, n=1). See Figure 2 for breakdown. Figure 2. Breakdown of offence types Theft 37% Burglary 14% Violence 12% Public Order 10% Drugs 8% Criminal Damage 7% Breach 7% Weapon 3% Harassment 1% 10 Characteristics of tag wearers The majority of tag wearers were Male (88%, n=64) and the average age was 32 years (SD=9.4). More than three-quarters of tag wearers were aged under 40 years: 27% (n=20) were 18-24 years and 51% (n=37) were 25-39 years. Twenty-one percent (n=15) of tag wearers were aged 40-54 years, and only 1 was aged 55+. In terms of the ethnicity of the 73 tag wearers, 53% (n=39) were White, 21% (n=15) Black, 15% (n=11) Asian, 10% (n=7) Mixed, and 1% (n=1) ‘Other ethnic group’. In this respect, the characteristics of those receiving tags reflects the characteristics of the POP cohort overall.2 During the first year of GPS tagging, there were 191 individuals who received the PSR writing service whose ethnicity was known: 45% (n=86) were White and 55% (n=105) Black, Asian or Minority Ethnic (BAME). Of these, 23% (n=20) of White service users and 25% (n=26) of BAME service users were recommended GPS tagging, suggesting the report writers proposed GPS tags on similar proportions of White and BAME individuals. Compliance with tag Of the 73 tags imposed during the first year of the pilot (i.e. up until 20 March 2018), 41 were completed successfully (‘compliant’ tag wearers), while 32 were unsuccessful (‘non-compliant’ tag wearers), either revoked for a failure to comply with the order and/or imprisonment following the commission of further offending, or were in the process of being revoked: a compliance rate of 56%.3 For the 73 GPS tags imposed, 33 Interest or Exclusion Zones had been set up by OMs. OMs receive an alert any time the tag wearer enters these zones. In the first year of the pilot, there were 44 Interest or Exclusion Zone alerts sent to OMs. Compliant versus non-compliant tag wearers Characteristics of compliant and non-compliant tag wearers are presented in Table 1. In terms of the demographic characteristics, both groups consisted of predominantly males. The average age of the compliant group was 30.6 years, slightly younger than the non-compliant group (average 34.6 years, although this difference was not statistically significant). For both the compliant and non-compliant groups, the proportion of White and BAME service users were similar (White individuals made up just over 50% of the groups). In terms of disposal, the non-compliant group had a higher proportion of tag wearers on a CO; however, again this difference was not statistically significant. Information was collected on the risk factors the two groups displayed (see Table 1). Specifically, individuals were measured on drug use, alcohol use, housing, physical health, mental health, and 2 POP refresh data from January-February 2018 consisted of 91% Males and an average age of 30 years. A pan- London dataset compiled from CRC data from January 2018 and NPS data from October 2017 showed the following ethnic composition; Asian 14%, Black 26%, Mixed 8%, Other ethnic 2% and white 50%. 3 This compares to a compliance rate of 51% (combining technical and non-technical parole violations) in a study testing the effect of GPS tagging with a sample of high-risk gang offenders in California (Gies et al., 2013). 11 motivation levels. On average, the non-compliant group displayed two risk factors, out of the potential maximum of six. The compliant group had a slightly lower average number of risk factors (1.3), but this difference was not statistically significant. For both groups, the most prevalent risk factor displayed was drug use, followed by mental health. Comparing individual risk factors across the two groups, there was only one significant difference: the non-compliant group displayed significantly lower motivation than the compliant group.4 The number tag wearers assessed as having housing issues was low for both group, but most likely reflect the fact that (relatively) stable accommodation is a prerequisite for the recommendation/ adoption of a GPS tag. Table 1. Characteristics of compliant and non-compliant tag wearers Non-compliant (n=32) Compliant (n=41) Gender Male 28 36 Female 4 5 Age (mean) 34.6 30.6 Ethnicity White 18 21 BAME 14 20 Order type CO 20 19 SSO 12 22 Risk factors Drug use 21 22 Alcohol use 7 10 Housing 5 4 Physical health 3 5 Mental health 15 17 Low motivation 15 7 There were several reasons specified for non-compliance.5 Lack of attendance was the most common reason, followed by failure to charge the tag. Other reasons were ‘total non-compliance’, non-engagement and the tag having never been fitted. In 12 cases (out of 15 cases where data was available), the tagging equipment was recorded to have been lost or ‘almost certainly’ lost. Details of the number of appointments attended with probation was available for 30 of the 32 individuals in the non-compliant group. On average, these tag wearers attended 3.9 appointments overall (ranging from 0 to 24 appointments). Interestingly, the pattern of attendance for those who did, and who did not, attend their first appointment with probation was markedly different. Individuals who attended their first appointment (n=18) attended an average of 5.5 appointments overall. Conversely, of the 12 individuals who failed to attend their first appointment, the average was 1.4 appointments overall. What this pattern appears to suggest is that there is a group of tag wearers who are largely non-compliant right from the 4 Fisher exact test statistic value is 0.0096, significant at p < .05 5 Reason for non-compliance was known for 23 out of the 32 non-compliance cases. 12 outset of the order, compared to other tag wearers who, while ultimately being breached, engage with probation over a longer period of time. Police Crime Mapping Under the crime mapping intervention, which began in September 2017, a dedicated police analyst maps the movements of tag wearers on SSOs against police crime data. Where the analyst believes there is a sufficiently close match, details are forwarded to the relevant police IOM team for further investigation. In deciding whether there is a potential match, the analyst considers the time of the offence (and the precision with which the time of the offence had been recorded), the location of the tag wearer, the movement of the tag wearer (i.e. their speed and direction of travel) and other factors, such as suspect descriptions. To assist in making a referral decision, the analyst devised a weighting system based on the above factors, automatically discounting scores below a certain level.6 Initially crime data was uploaded only for the eight boroughs covered by the POP pilot, but since late October 2017 crime data has been uploaded for all London boroughs. Table 2 shows the number of crimes uploaded by month. The table also shows the number of matches or ‘hits’ that were recorded by the analyst each month: a total of 865 for the 5-month period where data is available (0.3% of all crimes uploaded). However, not all the ‘hits’ were referred to IOM teams (see Table 2), as a certain proportion were discounted based on the weighting system developed by the analyst. In the period from September 2017 to January 2018, 46 referrals were made to the various IOM teams: only 5% of all apparent ‘hits’ generated. The vast majority of matches (95%) were discounted by the analyst. Table 2. Crime mapping data per month Sep-17 Oct-17 Nov-17 Dec-17 Jan-18 Total Crimes uploaded 12,483 30,673 76,424 70,448 74,499 264,527 Number of ‘hits’ 162 190 108 215 190 865 ‘Hits’ referred to IOM teams 13 20 3 4 6 46 To assess the outcome of referrals, the police analyst kept a database of what happened with respect to the matches that he provided to the police IOM teams. This covered issues such as the response from the IOM team when details of the match were emailed to them, subsequent actions (if recorded on CRIS), whether the individual was subsequently arrested, and local use of the GPS match (again, if the details were recorded on CRIS). 6 In practice the mapping process proved to be more time-consuming than anticipated due to differences in the way data was geo-coded on the Met’s and Buddi’s systems, meaning the analyst had to manually confirm the apparent proximity of the crime location and offender in some cases. 13 In 13 cases, receipt of the referral was acknowledged but no further action recorded. In a similar number of cases there was nothing more than a read receipt. In seven cases there was no response at all. In nine cases the decision was taken by the IOM team not to refer the case to the officer in charge due to what were often described as ‘evidential difficulties’ (e.g. the location was not specific enough, the time-frame provided by the victim was too wide, and there was no CCTV or forensic evidence). In 13 cases, there was an acknowledgment the referral had been passed on to either the officer in charge, the Crime Reduction Unit, or a sergeant. In two further cases the decision was made to look at the incident further; one case had been allocated for secondary investigation, and in the other, an EAR request had been sent to Buddi. To make a judgement about whether the GPS data was used in the investigation of the crime, the analyst looked at information subsequently put onto the CRIS system. In the majority of cases (88%, n=57)7 information from CRIS suggested the GPS data was either not used in the investigation of the crime, or there was no information available. In the eight cases where there were references on CRIS to the use of GPS, in no cases was the matter proceeded with (i.e. it was deemed the GPS data was not sufficient to proceed with an arrest, or there was a lack of additional corroborative evidence). At first sight the data above may appear disappointing, bearing in mind the large number of crimes up-loaded onto the system. However, it is important to remember the GPS-tagged population able to generate ‘hits’ was relatively small. Only 73 individuals had been tagged in total, and of these, only 34 were on SSOs and eligible for crime mapping. Of these, 12 were non- compliant at the outset (i.e. removed or did not charge their tags after fitting). Thus, only 22 individuals in total were eligible for crime mapping. Another complicating factor was the type of offences for which the matches were generated (largely theft and burglary). These offences, by their nature, are likely to provide the greatest latitude in terms of the reported time that they took place (and thus, the greatest difficulty in placing the offender at ene of the crime). This is reflected in the responses from the IOM officers about the limited utility of the data (and in contrast to a contact offence like robbery, for example). In these circumstances, it is perhaps unsurprising that the level of hits was so small. It is difficult to know, as a result, whether the low number of yields from the crime mapping reflects failings in the implementation of the scheme, or in the theory behind the concept of crime mapping, particularly bearing in mind the type of offences which many of the POP cohort are committing (thefts and burglaries). The box below provides two case examples of how GPS location data has been used to link tag wearers to other crimes, following earlier police investigations. Both are example where the IOM teams made EARs to Buddi about the specific movements of individuals. 7 Note this data covers the period September 2017 to April 2018. A total of 65 ‘hits’ occurred during this time 14 Practitioner Feedback Practitioner feedback was collected from nine offender managers (OMs) and two pre-sentence report (PSR) staff.8 The main themes are presented below. Offender Manager Feedback Use of GPS location data Eight of the nine OMs surveyed stated they were confident using GPS tagging in their role. Five had previously logged on to Buddi’s Eagle platform and seven had accessed location data. Of those who had accessed location data, six respondents had used it to set up Interest Zones, four to request heat and other maps, three to request information on a specific location, two to set up Points of Interest, and two to request ‘top location’ information. OMs were also asked how they used the location data to manage their case(s). Five had used it to inform discussions with the tag wearer, three to monitor attendance at a place linked to risk of harm, three to monitor attendance at substance misuse treatment, three to find out new information about the tag wearer’s lifestyle, three to try and locate a tag wearer who had gone missing, two to monitor attendance at a place linked to risk of reoffending, and two to challenge the tag wearer about his or her behaviour in supervision. Other uses mentioned by OMs were: to praise the tag wearer for compliance and positive behaviour; to protect known potential victims; and to monitor attendance at the tag wearer’s workplace. Respondents were also asked how easy it was to access location data. Of the seven for whom the question was relevant, six felt it was ‘easy’ or ‘very easy’, while one respondent felt it was ‘difficult’. For the two respondents who had not accessed location data, one stated this was because they did not think an Interest Zone or Point of Interest would be useful, the other because the case had only been allocated to them at the end of the requirement. Respondents 8 Given the size of the research cohort, caution should be used when considering the results. Case Example 1 The first case involved an Enfield offender who the police IOM team had linked to 19 burglaries in the Chingford, Walthamstow and Leytonstone areas via his MO. The individual was described by the police as a prolific offender who they were keen to have remanded because they were concerned that if granted bail he would reoffend and not attend court. Data obtained via an EAR request to Buddi allowed the police to place the offender at the burglaries. Case Example 2 The second example was where the offender had been linked to a theft from a motor vehicle via his bus pass which had been left in the vehicle. The offence had been reported to the police at 21:50 on the 11th February 2018. Again, via an EAR request, Buddi were able to provide the police with data placing the offender at the crime scene between 21:32 and 21:40 on the night of the offence. The offender was sentenced to 8 weeks imprisonment (consecutive to other matters). 15 were complimentary about the quality of the communication with Buddi, the GPS tagging provider, with eight describing it as ‘excellent’ and one as ‘good’. The responses above show some interesting developments in terms of OM perceptions of the GPS tag, compared to their views back in August 2017, some five months into the pilot. At this point, although most OMs had some awareness of the tag, this tended to be as a requirement to be enforced rather than something that could be used proactively to manage cases. At this point also, only one OM (of seven) reported having logged on to Buddi’s Eagle platform and none had set up an Inclusion Zone or Point of Interest. When asked the reasons for this the replies were that they were not aware that they could, they did not know how to, or that IT problems were causing issues. In the second survey, OMs were also asked to select their top three from six possible uses of GPS tagging. The uses that was selected most frequently were ‘to deter service users from reoffending’ (selected by 8 respondents) and ‘to monitor compliance with an exclusion zone or other restrictive requirement’ (6 respondents). Other commonly selected uses were ‘to monitor attendance with RAR/DRR/ATR appointments’ (5 respondents), ‘as an alternative to custodial sentences (4 respondents), and ‘to use location data to match with potential new offences (4 respondents). In the most recent survey, all nine respondents agreed or strongly agreed they understood when they could access and use location data, and seven out of nine indicated they knew how to access and use location data. Similarly, eight of nine respondents agreed they understood the restrictions and processes that applied to sharing location data with other agencies, and seven out of nine agreed they understood the different ways GPS tagging worked for COs and SSOs. When asked if they felt that GPS tagging was a useful way to tackle prolific offending, eight of the nine respondents either agreed or strongly agreed with the statement (the remaining respondent being neutral). However, when asked whether GPS location data had improved their ability to manage the case and engage with the offender, OMs’ feelings were slightly more equivocal, although most (n=6) agreed (two respondents disagreed, and one was neutral). Three of the offender managers provided comments on their experience of GPS tagging. One stated that he found the Buddi system difficult to access and use. The comments from the other two are presented below. “I have found GPS tagging useful in many ways. Firstly in regards to monitoring service users’ whereabouts and their attendance to certain places…More importantly in my opinion I think it does act as a deterrent to offending. One of my service users regularly comments on the fact that having the tag on has deterred him from offending. It makes them think twice before doing something stupid…It is also good in a way for eliminating crimes that the IOM Police may think they have done” “I had two cases, one for drug requirements and one who was a DV perpetrator. I got a mile exclusion set up on one, asked for a snap shot of where one was when he said he could not leave the house due to illness etc. Buddi were brilliant, both when they came to fit the equipment and any questions that I have had in relation to service users, very efficient.” 16 Pre-Sentence Report Staff Feedback Types of offenders being tagged When the PSR staff were asked about the use of GPS tags within the POP pilot, there was a recognition that often GPS tags tended to get used on minor offenders, precisely because these individuals were in stable accommodation, which was a prerequisite for a successful tag (although, conversely, it was also identified that there could be problems getting a tag imposed on an offender for a minor offence precisely because it was viewed by the courts as a disproportionate intervention). However, there was also the perception that tagging could be used on higher risk offenders, such as domestic violence or sex offenders, or those with mental health issues, precisely because the tag provided the ability for ‘additional oversight and better understanding’ (although one of the respondents was keen to stress that one of the factors he felt precluded any consideration for a tag was paranoia on the part of the offender). It was also suggested that GPS tags had a role to play in reinforcing curfew orders, or to provide effective enforcement of exclusion zones, which were areas that had previously proved problematic for probation. It was also felt that tagging could have a useful role to play with individuals involved with, or on the brink of involvement with, gangs. Notwithstanding the range of uses outlined above, both respondents felt that there was scope for much greater use of tagging, but what prevented this at present were issues related to staffing levels. “What can sometimes happen is that some of those people [shop-lifters, acquisitive criminals with drug issues] end up in relatively stable accommodation whether it’s through St Mungo’s or one of the other accommodation providers and actually at that point we’re able to get them on a tag and because there has been so much chaos around them it is a really good idea to do it at that point” “I think might be able to have more if we had more staff. It’s just hard. When I first [started] I was like a bloodhound, going to all the different courts. Then burnt myself out, couldn’t get everything…could probably have more [tags] if we had more resources”. “I know MOPAC are looking at knife crime and the people who have been involved in gangs and guns at a more lower level, from some of the conversations that I have had with offenders who are clearly on the periphery of guns and gangs …it can be a really easy way to make them undesirable to bigger gangs, because they do not want someone with GPS tagging in their gang”. 17 Wider understanding of GPS pilot When asked for their views about the level of understanding of, and support for, GPS by other probation staff, the response was mixed. One respondent suggested the dedicated CRC IOM officers were particularly impressive and had a clear idea of how the GPS tag could be used. One respondent suggested the presence of a dedicated report writer meant that other report writers did not take responsibility for proposing GPS tags. Furthermore, as above, both respondents felt the current level of staffing was too low and was hindering their effectiveness. One respondent suggested the ideal model for the team would be a PSR writer at each court, together with a support officer at each court providing administrative assistance. Asked for their perceptions about judges’ understanding of GPS tags the respondents were generally positive. However, their perception of magistrates’ understanding was more equivocal, with both respondents feeling that not all magistrates understood GPS tags, and that some were unclear about the distinction between curfew and GPS tags. Both respondents stressed the importance of their presence in court to be able to inform the court about the tags and to clear up any misunderstandings. However, one respondent was keen to put this in context, explaining that there were a large number of magistrates at the court, who might only serve one day a month, so seldom came across GPS tagging cases. Both respondents stated that where recommendations for GPS tags were made by the PSR writer they tended to be adopted (as is suggested by the data elsewhere in this report), to the extent that one respondent described himself as being ‘aggrieved’ if a tagging recommendation was not adopted. Both respondents identified the difficulties that had arisen since April 2018 when PSR writers had been instructed not to make recommendations for SSOs in reports. However, the respondents both stressed that while they could not be seen to be making a recommendation, they endeavoured to keep the courts aware that the SSO (and GPS tagging) was available to them as an option. Staff perceptions of offenders’ views about the tag The research was not able to talk directly to offenders on the tag, however staff spoken to felt that at the point of their court appearance, offenders would say anything to stay out of prison and were happy to consider a tag as an alternative to custody. One of the respondents suggested that where an offender was strongly opposed to a proposal for a tag it was an indication that “Half the time they [magistrates] don’t know what it is, they don’t know what we’re doing. When I’m in court and the probation officer says ‘we can put an electronic tag on him’ they just get confused and call it a curfew. They don’t really know what it is.” “We are quite fortunate that we have 4 resident DJs [district judges]…[who] are very aware of the GPS tagging and when I recommend it give it genuine thought.” 18 they did not want probation to know about their whereabouts and might be ‘involved in something really dodgy’. The same respondent also suggested that some offenders might be actively seeking to avoid the imposition of a GPS tag (e.g. by saying they do not have an address). Equally there was a recognition that the type of offender being considered for a tag under the POP pilot was likely to struggle with the organisational/logistical requirements of the tag, particularly requirements to keep it charged. There was also the recognition that the tag was something of a ‘double-edged sword’ for the wearer; its ability to provide additional oversight meant that it was likely to appeal to the court as an alternative to custody, but there was a greater likelihood for the tag-wearer to be breached because of wearing the tag. However, when asked directly if they were concerned about setting offenders up to fail by recommending them for a tag, this did not appear to be the case. As one respondent said: Views on Buddi The two respondents’ views on Buddi and their staff were positive. One respondent stated that he found the staff helpful, and with a good understanding of the issues that he faced. However, one respondent mentioned a problem he had had in the past getting Buddi to respond to multiple tagging requests on the same day. Tag Wearer Feedback All 22 tag wearers who were surveyed stated they understood why they had received the GPS tag, how the equipment worked, what they had to do to comply with the order, and how the tag fitted with their other requirements. When the tag was fitted, respondents were asked the extent to which they agreed or disagreed with a number of statements about the anticipated impact of the GPS tag. When asked if they were worried about what their friends and family would think about the tag, most respondents disagreed (n=14). Respondents were also asked to indicate how confident they were that they would complete the tag order successfully. All tag wearers described themselves as being either ‘fairly confident’ (n=9) or ‘very confident’ (n=13). Respondents were asked what impact they thought wearing the GPS tag would have on specific aspects of their life (‘better’, ‘worse’, or ‘no impact’, see Table 3). Generally, respondents felt the tag would either have a beneficial effect, or no impact. Respondents were most positive about “Everyone that we’ve talked to about it are happy, ‘Don’t want to go to prison, give me a GPS tag’ The ones in the cells will always say that ‘cos they’re about to come in front of the judiciary and they want to be the friendliest person around.” ‘I am very clear with them that if this tag is fitted then we are going to know what you are up to and if you choose to offend that is going to happen’. 19 the impact of the tag on their relationships with family (10 tag wearers agreed the tag will make their relationships with family better). Respondents appeared to be more pragmatic in terms of their expectations of wearing the tag on their relationship with friends, their financial situation, their physical health, their mental well-being, their housing situation, employment status, ability to go out or socialise, and their education situation. In every case the most popular response was that the tag would have no impact. When asked what the anticipated impact of wearing the tag would be on their life overall, 18 respondents felt that it would get better, while 3 felt the tag would have no impact; nobody thought their life would get worse. In terms of the impact of the tag on their offending behaviour, 20 respondents thought it would lead to an improvement. Table 3. Tag wearers’ perception of anticipated impact of wearing GPS tag Anticipated impact of wearing tag on… (n=21 unless otherwise specified) Better No impact Worse N/A Relationship with family 10 9 1 1 Relationship with friends 7 13 1 0 Money/current financial situation (n=20) 6 13 0 1 Physical health 7 13 0 1 Mental well-being 8 11 1 1 Housing situation 7 14 0 0 Employment situation 2 15 1 3 Ability to go out/socialise 7 12 2 0 Education situation 2 15 1 3 Respondents were also asked to provide general comments about their expectations on having the tag fitted (16 responses were received). The most popular assertion (mentioned by 7 respondents) was that the tag would assist them in ‘sorting their lives out’, typified by the following statement: Six respondents stated that it would stop them from committing crime/offending or make them think about such actions. Five respondents mentioned the GPS tag was preferable to a curfew tag (in that it was less restrictive), while four respondents said wearing the tag was a preferred alternative to going to prison. Typical quotes reflecting these sentiments were: “I am a burglar and wearing this kind of tag will show probation I am not getting involved in burglaries and will remind me the police and probation will be checking on me. I think this kind of TAG is better than a curfew TAG because I can get on with my life” “I understand that I must stay away from my partner and the tag stops me going to her address. I think it will help me sort things out by reminding me I can't go to her address. I hope this and my counselling will make life better for me” 20 Other factors mentioned were that the tag would keep them out of situations where they might cause harm to others, would assist in enforcing the restrictions of their order (to stay away from their partner), and that it would show probation they were not offending. One respondent described the fitting of the tag as his ‘last chance’, but felt he was being set up to fail due to the inability of drug services to see him to address his drug dependency. Tag wearers were also asked to complete a similar questionnaire when their tag was removed. Responses from 12 individuals were analysed. Respondents were asked to rate (on a scale of 1- 7, with 1 being ‘very negative’ and 7 ‘very positive’) their overall experience of being on the tag. The results are shown in Figure 3 below, and clearly indicate respondents’ positive perceptions. Figure 3. Tag wearers’ experience of being on a GPS tag This positive outlook is reflected in responses to several related questions. For example, when asked if wearing the tag was better than they thought it would be, 10 respondents agreed, whilst only 2 disagreed. The same pattern of responses was provided to a question asking if they found the tag comfortable to wear. Respondents were also asked whether they had worried about what their friends and family thought about the tag, and only 2 agreed. The removal survey repeated the questions from the fitting survey about the impact of the tag (‘better’, ‘worse’, or ‘no impact’) on a variety of aspects of respondents’ lives. All 11 respondents thought their offending behaviour had improved because of wearing the tag, and, in terms of life in general, 9 of the 11 respondents thought the tag had made it better. Responses to the other questions are shown in Table 4 below (for comparative purposes, the responses at the point of tag removal are shown in bold, the numbers in parentheses are the responses from the tag fitting survey). Although the small sample size has to be borne in mind, what is noticeable from the table is that the positive outlook identified when the fitting survey was undertaken appears to have been maintained. At tag removal, respondents still tended to 0 1 2 3 4 5 6 1 Very negative 2 3 4 5 6 7 Very Positive Respondents “I am happy to wear this tag because I did not want to go to prison. I hope the tag and drugs help keeps me out of crime. I feel positive about wearing the GPS tag and will do my best to stick to the order” 21 feel the tag had either no impact on the quality of their life or had led to an improvement. In only one area (‘ability to socialise’) did the number of negative responses (one) outweigh the positive. Table 4. Tag wearers’ perceptions of the impact of wearing the GPS tag Impact of wearing tag on… n=11 at tag removal Better No impact Worse N/A Pre Post Pre Post Pre Post Pre Post Life in general (18) 9 (3) 1 (0) 1 (0) 0 Relationship with family (10) 6 (9) 5 (1) 0 (1) 0 Relationship with friends (7) 5 (13) 5 (1) 1 (0) 0 Money/current financial situation (6) 0 (13) 11 (0) 0 (1) 0 Physical health (7) 2 (13) 8 (0) 0 (1) 1 Mental well-being (8) 5 (11) 5 (1) 0 (1) 1 Housing situation (7) 1 (14) 10 (0) 0 (0) 0 Employment situation (2) 0 (15) 11 (1) 0 (3) 0 Ability to go out/socialise (7) 0 (12) 10 (2) 1 (0) 0 Education situation (2) 0 (15) 11 (1) 0 (3) 0 Respondents were also asked to provide general comments about their experiences of wearing the GPS tag (11 responses were received). Seven people reported the tag had helped them stay out of trouble and away from crime. Four people mentioned the tag had helped them think about or reflect upon their actions, and the same number stated they preferred the Buddi tag to a curfew tag. Although three respondents stated they would wear the tag again, the same number mentioned the problems they had charging the tag (not the act of charging the tag itself but the organisation/thought required to do so). One respondent suggested that having a charger at probation or drugs service might help in this respect. Other themes mentioned by more than one respondent were that they had tried to cut the tag off (mentioned by two respondents) and that they were conscious the police and probation could see where they were when on the tag (also two). “Wearing the tag has made me think how I behave and has helped me get into more stable accommodation. I think it’s helped me stay away from trouble and I like it better than the curfew tags because I am not restricted in my movements. It felt comfortable wearing the tag and I would wear it again”. “I found the tag surprisingly comfortable to wear. I am glad that I went for wearing the tag because I did not want to go to prison. The tag had an exclusion and wearing the tag has kept me away from the area and made me think about what I get up to. I was conscious probation and police could see where I have been.” 22 Another respondent stated that wearing the tag for three months was not long enough to change the way people think and suggested he would like to keep the tag on because he felt it was like a ‘security blanket’. A further respondent stated that initially he did not think he needed the GPS tag but had subsequently changed his views, and another that he had tried to cut the tag off but did not go through with it because there would be ‘a consequence’. “This tag has prevented me from committing further offences. I tried cutting off the strap to go and do crime but I could not get it off. Because of this it stopped me and has helped me to think about my actions. Best tag I have worn, and I have worn a lot. If I had the choice I would wear it again for longer” “Wearing the tag has stopped me from offending. I have mental health and drug problems and clearly the tag has helped me focus on staying away from trouble, keep my appointments because probation can see what I am doing and places I have been. I am trying to improve my life and the tag has helped me. I would honestly choose to wear it again”. 23 4. Discussion In the first year of its operation, a total of 73 tags were imposed as part of the GPS tagging pilot. The majority of electronic monitoring requirements (93%) were imposed alongside a rehabilitation requirement – consistent with the intended used of the tags as a package of support, monitoring and punishment. Just over half (56%) of all tag wearers complied with their order. Lack of attendance and failure to charge the tag were the two most common reasons for non-compliance. Overall, findings from the first 12 months of the pilot show the GPS tagging pilot has been well received by practitioners and tag wearers alike, although the findings are based on small samples. Feedback from the PSR staff suggested the judiciary were generally receptive to recommendations made for tagging requirements, and were happy to adopt these recommendations, although there was a perception that judges’ understanding of GPS tagging was better than magistrates’. There was a general feeling that there was scope for greater use of GPS tagging, and for its use with higher risk offenders, but that current staffing levels acted as a barrier to this being realised. In comparison to the interim evaluation report, it appears there is greater awareness of the GPS tag on the part of offender managers (at least, those that completed the survey), and a willingness to use interest and exclusion zones as part of supervision. As with the first interim report, tag wearers had positive expectations of the tag. They were confident they would successfully complete the tag order and knew what they needed to do to comply. Furthermore, tag wearers thought the tag would have a positive impact on their life and lead to them committing less crime. At tag removal, the majority of tag wearers surveyed thought being on the tag had improved their offending behaviour and had made their life better in general (although, again, the number of respondents was small). However, there were still concerns expressed (from tag wearers and practitioners) about the difficulties caused for this cohort by the requirements to keep the tag charged. It is clear from the data there are still issues around the identification of suitable individuals to be given the tag, reflected in the levels of lost equipment, and examples of early non-compliance (steps to combat are currently being taken by the implementation team). Other challenges that have emerged are the repercussions arising from the changes to PSR writing policing around making recommendations for SSOs in April 2018, which resulted in a reduction of SSOs with tagging requirements, and a reduction in eligible cases for the crime mapping strand of the pilot (which can only be undertaken on SSO tag wearers). Question marks about the effectiveness of the crime mapping, and whether the identification of crime matches should be analyst-led, as is currently the case, or should be led by IOM priorities and prompts also need to be resolved during the remainder of the pilot. Nevertheless, the overall compliance rate for the scheme (56%), and the positive views expressed by practitioners and service users about their experience with the GPS tag, provides grounds for optimism. The next evaluation report is scheduled for Summer 2019 following the completion of the pilot’s extension. This report will provide updated results for the final GPS cohort, looking at issues 24 around compliance, as well as reporting back on the experiences of tag wearers, probation and court staff. Harmful Practices Pilot Final Evaluation Report May 2017 Melissa Pepper and Helena McKinnon MOPAC Evidence and Insight 2 Executive Summary ......................................................................................................... 3 Introduction .................................................................................................................... 4 Background .......................................................................................................................................... 4 The MOPAC Harmful Practices Pilot .................................................................................................... 5 Final Evaluation Report: Building on a Year One Interim Report .......................................................... 6 Methodology ................................................................................................................... 6 Results: Reflections on the Two Year Pilot .................................................................... 7 Implementation .................................................................................................................................... 7 Early Identification and Prevention: Training Professionals .................................................................. 8 Safeguarding and Support: Advising and Guiding Professionals through the Educator Advocate Service/Surgeries ............................................................................................................................... 12 Community Engagement: Empowering Women and Girls ................................................................... 16 A ‘Partnership’ Pilot: Working Together to Tackle Harmful Practices ................................................. 18 Discussion ..................................................................................................................... 19 References .................................................................................................................... 21 Appendices .................................................................................................................... 22 3 Executive Summary The MOPAC Harmful Practices Pilot aimed to improve the way agencies identify and respond to Female Genital Mutilation (FGM), ‘Honour’ Based Violence (HBV), Forced Marriage (FM), and Faith Based Abuse (FBA), with a focus on early identification and prevention, safeguarding and support, and community engagement. The two year pilot, delivered by the Partnership to End Harmful Practices (PEHP, a consortium of seven women’s organisations) in Westminster, Kensington and Chelsea, Hammersmith and Fulham, Tower Hamlets and Waltham Forest, began in April 2015. It involved three key areas of work: training for professionals; an Educator Advocate (EA) service delivered by a member of staff from one of the PEHP organisations to provide advice and support to practitioners (via two half-day surgeries per borough per week located in statutory agencies); and community engagement activity. This report reflects on learning from the pilot, during which a number of positive messages emerged: • Training has been widely recognised as a key success with the 74 courses delivered received well by attendees in terms of quality, content, and improving practitioner knowledge around harmful practices. Respondents felt the training had improved their ability to identify and respond to harmful practices, and often highlighted examples or plans to share learning with colleagues back in the workplace. • EAs facilitated 591 surgery sessions providing advice and support in relation to 218 harmful practice and Violence Against Women and Girls (VAWG) cases. Project leads and practitioners recognised the wealth of knowledge that EAs brought around harmful practices and beyond (e.g., immigration advice, language skills), and shared examples of cases where EAs and practitioners worked together to support clients. • EAs arranged or took part in 51 community engagement events with the majority of feedback received indicating that attendees understood more about harmful practices and where to access support after events. Practitioner attendees spoke positively about opportunities to find out more about services in their area, which would help them better support clients. • A strong partnership approach underpinned the pilot, with statutory and voluntary services learning from each other around how to support those experiencing, or at risk of experiencing, harmful practices. However, the pilot has also presented challenges, particularly around the EA role/surgeries, where referrals have at times been slow, access to resources (e.g., internet, computers, and case files) limited, and EAs have been required to continually build relationships, promote the project, and ‘carve out’ their role within fast-paced, often changing surgery host organisations. While community engagement efforts were welcomed, pilot staff interviewees felt that some events could have achieved more with additional coordination and planning, including specific community engagement materials. The pilot highlights learning around the importance of comprehensive early project planning, and clear operating models, both at strategic and operational ‘on the ground’ level – challenges not unique to the Harmful Practices Pilot. The Harmful Practices Pilot has started conversations, encouraged ‘professional curiosity’ in practitioners, and highlighted new ways of thinking amongst statutory and voluntary agencies about under-researched and often misunderstood violent and abusive practices affecting thousands of individuals throughout the UK and millions across the globe. Practitioners and professionals who took part in fieldwork often commented that the pilot training package could be developed for wider rollout, together with refresher workshops to ensure that learning continues. The training ‘offer’ should be embedded within an organisational commitment to review policies to ensure that trained staff can put learning into practice back in the workplace. The pilot contributes to a limited yet developing evidence base around harmful practices. The challenge now is to continue to raise awareness and further develop partnerships amongst statutory and voluntary agencies, and communities themselves, in order to build on learning to more effectively identify and support individuals experiencing, or at risk of experiencing, harmful practices. 4 Introduction Background The limited (although developing) research around harmful practices – including Female Genital Mutilation (FGM), ‘Honour’ Based Violence (HBV), Forced Marriage (FM), and Faith Based Abuse (FBA)1 – clearly highlights the serious and long-lasting physical, emotional and psychological effects on those who experience them. The context in which harmful practices take place is complex, set against a backdrop of power, control and constructs of gender, where (actual or perceived) family love can mask abuse by multiple perpetrators and ‘normalise’ violent practices. Studies consistently point to a lack of professional awareness of harmful practices, compounded by inadequate training provision. Practitioners often have low confidence when engaging with affected individuals or communities, and are reluctant to intervene due to fear of being perceived as culturally insensitive. Furthermore, the response to harmful practices and the support available for those experiencing them is piecemeal, with limited strategic coordination and integration into wider safeguarding and Violence Against Women and Girls (VAWG) services. Specialist Black, Asian and Minority Ethnic (BAME) women’s organisations have been identified as an important factor both for immediate safety and a wider sense of empowerment; however, with few services available (and those that are, under pressure), responses are often inconsistent and opportunities to support women and girls are being missed (Bahunga, 2012; Dickson, 2014; Larasi et al., 2014; Norman et al., 2009; Payton, 2015; Roy et al., 2011; Tedam, 2014). Available data on the prevalence of harmful practices presents a patchy and incomplete picture due to issues such as under-reporting, a lack of uniform and systematic recording systems, and misunderstanding around definitions2. The World Health Organisation (2016) state that more than 200 million women and girls alive today have experienced FGM, while in England and Wales it is estimated that 137,000 women and girls are living with the consequences of FGM, with a considerable concentration in London (21 cases per 1,000 population compared to 4.8 for England and Wales as a whole) (Macfarlane and Dorkenoo, 2015). Indeed, around half (49%, n=620/1,268) of newly recorded cases and total FGM attendances (49%, n=1,140/2,332) captured by the Health and Social Care Information Centre - on behalf of the Department of Health and NHS England - in October to December 2016, relate to women and girls from the London NHS Commissioning Region (NHS Digital, 2017)3. From a policing perspective, the Metropolitan Police Service (MPS) recorded 233 FGM incidents (which included 22 offences) in 2016/17, an increase on the previous financial year (209 and 17)4. In terms of forced marriage, in 2016 the government’s Forced Marriage Unit gave support in 1,428 cases, 11 per cent of which took place entirely within the UK. London made up the largest proportion – over a fifth (21%, n=307) – of all cases (Home Office/Foreign and Commonwealth Office, 2017). The MPS recorded 226 FM incidents which included 96 offences in the most recent 1 See appendix one of https://www.london.gov.uk/sites/default/files/vawg_strategy.pdf. for definitions 2 This was recognised in the 2015 HMIC inspection The Depths of Dishonour: Hidden Voices and Shameful Crimes which recommends the development of a standardised approach to data collection. Caution is advised when interpreting harmful practices data included in this section due to low reporting levels and inconsistent recording practices. 3 ‘Newly recorded’ women and girls are those who have had their FGM information collected in the FGM Enhanced Dataset for the first time. ‘Total attendances’ refers to all attendances in the reporting period where FGM was identified or a procedure for FGM was undertaken. Women and girls may have one or more attendances in the reporting period. This category includes both newly recorded and previously identified women and girls. From 31st October 2015 mandatory FGM reporting was introduced in England and Wales, which imposes a duty on healthcare professionals to report cases in girls aged under 18 years to the police. 4 Incidents are a count of all crime records (all Notifiable Offences and Other Accepted crimes), including those which have been classified as ‘No Crime’ or a ‘Crime Related Incident’. Offences are a count of confirmed classifications of crime records (all Notifiable Offences only). MPS data is based on the presence of a relevant harmful practice flag or offence code. It is difficult to ascertain how consistently flags/offence codes are used; however, it is likely that data does not present a complete picture of all cases. 5 financial year (2016/17), a slight increase on the previous year (200 and 79 respectively). Data around HBV and FBA is considerably more limited, compounded by a lack of clear legislative framework5. Figures obtained by the Iranian and Kurdish Women’s Rights Organisation (IKWRO) indicate that across the UK in 2010 there were more than 2,800 incidents of HBV reported to police, again with the largest proportion in London (18%, n=495). The MPS recorded 349 HBV incidents which included 228 offences in the most recent financial year, slightly down on the previous year (412 incidents and 278 offences). In terms of FBA, research for the Department for Education and Skills identified 38 cases of child abuse linked to accusations of ‘possession’ or ‘witchcraft’ between 2000 and 2005, while a later study highlighted the deaths of at least six children in the UK between 2000 and 2010 following periods of abuse arising from being labelled ‘witches’ (Stobart, 2006; Tedam, 2014). According to MPS data there were 31 ‘ritual and ritualistic abuse’ incidents which included 14 offences recorded in the financial year 2016/17, a decrease on the previous year (71 incidents and a further 49 offences). The scarcity of available data and research in this area makes it difficult to fully understand the scale, impact, and context of harmful practices. The MOPAC Harmful Practices Pilot The Mayor’s Office for Policing And Crime (MOPAC) Mayoral Strategy on Violence Against Women and Girls 2013-17 developed plans to establish a multi (voluntary and statutory) agency taskforce to address harmful practices. Starting in April 2015 and running for two years, the MOPAC Harmful Practices Pilot aimed to improve the way agencies identify and respond to FGM, HBV, FM, and FBA. The initial overarching themes of the pilot were focused around: • Early identification and prevention: Through the development and delivery of training to key agencies to enable them to identify and respond to harmful practices; • Safeguarding and support: Through a specialist harmful practices Educator Advocate post to work with agencies to ensure women and girls who have experienced, or are at risk of, harmful practices are safeguarded and offered specialist support; • Enforcement and prosecutions: Through a focus on the police response to harmful practices with the ambition of securing prosecutions and ensuring that other interventions are being used effectively and appropriately6; and • Community engagement: With a focus on both empowering women and girls within affected communities, and challenging the acceptance of harmful practices by men, boys, and wider communities. 5 FGM has been a specific criminal offence in the UK since 1985, with various acts of legislation – most recently the Serious Crime Act 2015 - adding to the legal framework. The Anti-Social Behaviour, Crime and Policing Act 2014 made it a criminal offence to force someone to marry. There is no specific legislation around HBV or FBA, which are covered under a range of violence, abuse, and child safeguarding laws. 6 As learning developed throughout year one, the pilot amended in response. For example, previous research (e.g., see Larasi et al., 2014), together with pilot stakeholders and professionals, identified that enforcement may not be a positive outcome for all women and girls; therefore, an initial focus on enforcements and prosecutions was considerably scaled back with the pilot predominantly focusing on themes of early identification and prevention, safeguarding and support, and community engagement. Although not a key area of this specific pilot, MOPAC acknowledges that enforcement action is an important part of the response in some harmful practice cases, and hopes that learning will also inform these approaches going forward. 6 The programme of work was delivered by the Partnership to End Harmful Practices (PEHP)7 in two locations - the tri-borough partnership (Westminster, Kensington and Chelsea, and Hammersmith and Fulham) and the east London partnership (Tower Hamlets and Waltham Forest). There were three key areas of work: • Half-day multi-agency and two-day specialist training sessions for professionals8; • An Educator Advocate (EA) service/surgery – delivered by a member of staff from one of the PEHP specialist women’s organisations with two half-day surgeries per borough per week located in (predominantly) statutory agencies and voluntary organisations – to provide advice and support to professionals and practitioners around harmful practices; and • Community engagement activity. The first 18 months of the pilot rolled out alongside a Department for Education (DfE) Innovation Fund project, focused on addressing FGM through work with health care professionals and borough-based social care teams. Although separate and distinct in terms of processes and evaluation methods, the two pilots were conducted under the umbrella of a single Mayor of London/MOPAC harmful practices model, in the same areas with borough based project managers responsible for the delivery of both. The pilots offer opportunities for MOPAC and wider statutory and voluntary sector agencies to both learn from a proactive approach to addressing FGM, HBV, FM and FBA, and contribute to a currently limited evidence base around understanding and responding to harmful practices across London and beyond9. Final Evaluation Report: Building on a Year One Interim Report This final evaluation report builds on a year one interim report produced in June 2016, which focused primarily on implementation, views of those attending the first tranche of training sessions, and the early experiences of professionals and communities receiving services as part of pilot delivery. Year two fieldwork has developed this, while attempting to explore how training delegates have transferred learning into their day-to-day working, the influence of EA support on professionals’ ability to identify and respond to harmful practices, and the experiences of women and affected communities involved in the pilot. Methodology The Evidence and Insight (E & I) Unit – a team of social scientists based within MOPAC – were commissioned to undertake an assessment of the Harmful Practices Pilot10. The challenges of evaluation research in complex and evolving pilot landscapes such as this have been well documented (Dawson and Williams, 2009; Patton, 2002). Given the nature of harmful practices and the likely small sample sizes the pilot would generate, the most feasible design was deemed to be a ‘proof of concept’ process assessment. The evaluation aims were to: 7 The PEHP consists of seven organisations: Iranian and Kurdish Women’s Rights Organisation (IKWRO); Foundation for Women's Health Research and Development (FORWARD); Latin American Women’s Rights Organisation; Asian Women’s Resource Centre; IMECE Women’s Centre; ASHIANA Network; and the Women and Girls Network. 8 The training ‘offer’ also included a series of specialist half-day follow-up training sessions focusing on the psychological impact of harmful practices, risk assessments, and safeguarding. 9 This report comments on the evaluation of the MOPAC Harmful Practices Pilot only. The DfE FGM pilot was initially planned for one year but extended for an additional six months. A separate evaluation of the FGM pilot was conducted externally by Opcit Research with the University of Central Lancashire (UCLAN) Department of Social Work and is available at https://www.gov.uk/government/publications/female-genital-mutilation-early-intervention-model-evaluation. 10 Thanks to Barry Charleton, Christina Soderberg, and Laura Duckworth for data inputting and analytical support. Thanks also to the peer reviewer for useful comments on both the interim and final report. 7 • Critically review implementation: Has the pilot been delivered effectively (e.g., in terms of the process, services delivered, and meeting professional and individual needs)? • Capture the views of practitioners: Has the pilot improved the ability of practitioners and professionals to identify and respond to harmful practices? What are their views on the development and support they have received as part of the pilot? Have safeguarding processes been revised to ensure cases of harmful practices are responded to? • Capture the views of communities: Are individuals who have experienced harmful practices better able to identify and understand what has happened to them, and communicate their needs and views? Has there been an increase in the number of cases coming to the attention of/being reported to professionals? Adopting a pragmatic action research approach11, the evaluation fieldwork incorporated a holistic range of methods to capture both ‘the numbers and the stories’ of those involved in the pilot (Patton, 2002). This included: two stakeholder surveys (113 responses); training feedback surveys (660 responses); community engagement surveys (78 responses); 48 semi-structured face-to-face or telephone interviews with stakeholders, pilot staff, practitioners, and clients; observations at four training sessions and five community engagement events; and performance data around both individual cases supported and overall pilot delivery. An evaluation timeline and full details of fieldwork conducted is included in Appendix one. This report draws together findings across the range of evaluation methods under five headings which reflect both the overarching aims of the pilot (including each aspect of delivery – training, EA service/surgeries, and community engagement) and other emerging key themes: implementation (largely incorporating point one of the evaluation aims outlined above); early identification and prevention (point two of evaluation aims); safeguarding and support (points two and three of evaluation aims); community engagement (point three of evaluation aims); and a partnership approach to tackling harmful practices (point one of evaluation aims)12. Pilot learning outlined in this report contributes to a currently limited evidence base, and hopes to inform approaches of other organisations in their work to tackle harmful practices. Results: Reflections on the Two Year Pilot Implementation Year one of the pilot highlighted a clear need for initiatives to tackle harmful practices with the majority (86%, n=36/42) of year one stakeholder survey respondents coming into contact with a person who had experienced, or was at risk of experiencing, a harmful practice. Concerns around professionals taking the wrong course of action, individuals’ ability to identify, understand and disclose that they had experienced a harmful practice, engagement with services, and whether survivor needs were being met were also raised. Implementation proved difficult in year one, particularly in the early months, as pilot staff worked to clarify operating models, roles, and 11 For example, researchers fed back immediately after observations at training and community engagement sessions, regularly attended and contributed to pilot steering groups, produced a short report in January 2016 reflecting on practical learning from fieldwork to date in order to allow adequate time to inform plans for year two of the pilot, and delivered an interim report at the end of year one. Given the size of the research cohort, some caution should be used when considering results. Fieldwork data set out in this report reflects the views of those who took part in surveys and interviews, and researcher observations. 12 To reduce duplication, findings from all evaluation methods – including surveys, interviews, and observations – are structured within these thematic headings, rather than presenting each as a separate dataset. 8 responsibilities. This was at times compounded by the separate DfE funded FGM pilot which initially rolled out alongside the Harmful Practices Pilot. As the lifespan of the FGM pilot was shorter, there was a greater sense of urgency to ‘get going’, which perhaps impacted on the time available to carry out adequate groundwork for the Harmful Practices Pilot. Indeed, a year one interviewee raised concerns that in the early days of delivery perhaps the Harmful Practices Pilot ‘hadn’t been all it could have been as a result’. Challenges persisted at times throughout year two with interviews conducted with project leads (across MOPAC and the pilot boroughs) and EAs at the end of the pilot often highlighting a need for additional planning early on in the pilot. One interviewee felt that the ‘vision was there’ but the pilot sometimes ‘lost its direction’. Implementation problems were most commonly noted in relation to the EA service/surgery component of the pilot with an early lack of clarity around how the role would work in practice, and responsibilities for identifying and setting up surgeries. Although a document outlining the role of EAs/surgeries was developed, it would be fair to suggest that an unclear ‘offer’ early on in the pilot contributed to implementation difficulties around this part of the work. These were compounded in the second year due to a number of pilot staff changes including four of the five EAs, two project leads, and the PEHP project manager – a role that was not substantively filled throughout year two. This was highlighted as problematic by a number of year two interviewees: as one commented: ‘we needed a lynchpin who makes sure everything happens’. Rigorous implementation and delivery are essential to successful programmes, and it would perhaps be fair to suggest that ‘start up’ of this pilot commenced too early in the project planning stages. Such challenges are by no means unique to the Harmful Practices Pilot (e.g., see Dawson and Stanko, 2013) and it is positive that year two interviewees reflected on significant learning around implementation, which they felt would inform planning and delivery going forward. Early Identification and Prevention: Training Professionals Over the two-year period a total of 74 courses ran across the two pilot areas: 40 half-day multi- agency, 12 two-day specialist, and 22 half-day follow-up courses focusing on risk assessment (6), psychological impact of harmful practices (10), and safeguarding (6). It isn’t possible to state how many individuals were trained as professionals often attended more than one course; however, attendance sheets received (these were missing for 20 sessions) indicated: • 341 attendees at multi-agency courses (an average of 12, ranging from 2 to 19) • 155 attendees at specialist courses (an average of 11, ranging from 4 to 22) • 75 attendees at follow-up courses (an average of 5, ranging from 2 to 14)13. Attendees completed a feedback survey at the end of each training session (including both days of the two-day specialist course), with 660 feedback surveys received in total (surveys were missing for 15 sessions). Training session attendees came from a variety of professional backgrounds, predominantly health, social care, and the Voluntary and Community Sector (VCS) (see appendix two). There was a notably slower uptake of training in the tri-borough partnership in year two and 13 Figures based on number of attendees ‘signing in’ on attendance sheets. It is likely that some attendees did not sign in. It was anticipated that there would be approximately 15 attendees at each multi-agency and specialist training session and six attendees at follow-up training sessions. Average attendee figures are based on: 341 attendees signing in across 29 multi-agency courses for which signing in sheets were provided; 255 attendees signing in across 23 days (day one and two) of the specialist courses for which signing in sheets were provided; and 75 attendees signing in across 15 follow-up courses. Feedback analysis is based on 660 returned feedback forms. Not all respondents answered all questions. ‘N’ numbers are stated throughout. See appendix three for a full breakdown of courses, attendee numbers, and feedback forms received, together with details of courses cancelled/rescheduled, and signing in sheets and feedback forms not received. 9 a number of courses were rescheduled or cancelled (see appendix three). Issues were discussed regularly in steering group meetings and partners actively sought to understand drivers. Low take- up was linked to capacity issues (the tri-borough training team were understaffed and unable to promote courses as widely as they had in year one), wider mandatory training demands on professionals, and the possibility of reaching saturation point. The PEHP allocated some unused sessions to deliver bespoke ‘in-house’ training, and boroughs explored possibilities around offering training to neighbouring areas. Positive Views on Training Training was widely recognised by interviewees and survey respondents as a key success of the pilot, with frequent praise for the sessions’ focus on concepts of power and control, and their potential to encourage ‘professional curiosity’ and empower practitioners to recognise and respond to harmful practices. Training sessions were well received by attendees in terms of quality, content, and feeling that the pilot would improve practitioners’ ability to address harmful practices (see table one). Themes were generally the same across areas and training course types, albeit with slightly more positive responses from east specialist courses compared to tri-borough specialist courses – however some analyses were based on small numbers. Table 1: Training session survey feedback Statement14 Agree Neither agree nor disagree Disagree The training met my expectations 85% (n=532/625) 9% (n=54/625) 6% (n=39/625) The training achieved its stated objectives 86% (n=536/620) 7% (n=45/620) 6% (n=39/620) The training materials were useful 86% (n=536/623) 6% (n=39/623) 8% (n=48/623) The trainer was knowledgeable 89% (n=554/623) 4% (n=28/623) 7% (n=41/623) There was a good level of participation and interaction 88% (n=547/625) 5% (n=34/625) 7% (n=44/625) I was given enough time to ask questions 87% (n=545/625) 6% (n=35/625) 7% (n=45/625) The training session was pitched at the right level 85% (n=530/624) 7% (n=41/624) 8% (n=53/624) The environment was safe and supportive 88% (n=548/624) 4% (n=28/624) 8% (n=48/624) There was an atmosphere of trust and confidentiality 89% (n=555/621) 4% (n=23/621) 7% (n=43/621) Overall I was satisfied with the training session 87% (n=541/622) 6% (n=39/622) 7% (n=42/622) I understand the aims and objectives of the Harmful Practices Pilot 87% (n=508/585) 6% (n=38/585) 7% (n=39/585) I know where to get more information/support around the Harmful Practices Pilot 88% (n=514/586) 6% (n=35/586) 6% (n=37/586) Overall, this training session has provided me with enough information to play my part in delivering the Harmful Practices Pilot 85% (n=506/596) 9% (n=54/596) 6% (n=36/596) 14 Training attendees were asked to indicate their level of agreement with a series of statements with ‘1’ being strongly agree and ‘5’ strongly disagree. Figures used to indicate agreement to some extent are those responding ‘1’ and ‘2’. Figures used to indicate disagreement to some extent are those responding ‘4’ and ‘5’. Responses given as ‘3’ are considered to be ‘neither agree nor disagree’. 10 I think that the Harmful Practices Pilot will improve practitioners/agencies ability to address harmful practices 88% (n=531/602) 5% (n=32/602) 6% (n=39/602) I would recommend this training course to other colleagues 88% (n=531/603) 6% (n=34/603) 6% (n=38/603) In freetext fields on the training feedback survey, attendees often commented positively on the quality and content of sessions (in particular, group discussions, case studies, information on sign- posting, referrals and support services, and hearing victims’ stories), trainer knowledge, and opportunities to share experiences with a range of different agencies. However, some respondents highlighted a need for additional in-depth information, and adequate time being allocated to sessions (particularly the half-day multi-agency course), together with some logistical issues (e.g., problems with the venue or technology), which affected a small number of sessions. In terms of how the training would impact their work, most respondents reported positive changes in terms of their confidence to identify and respond to warning signs and support survivors, and plans to share learning with colleagues. ‘I am more informed and now know who to approach if I have concerns or if someone should approach me for help’ ‘I think I'm in a better position to look out for those critical signs’ ‘[The training has] given me confidence to explore/start conversations with clients and colleagues when dealing with possible victims of harmful practices’ ‘Albeit an uncomfortable subject, I found this training to be a great eye opener, very interactive and resourceful’ ‘No matter how deeply rooted a practice may be it is possible to take steps to make changes. Sessions like this will have a positive impact…education/awareness is the key to help make these changes’ Transferring Learning in to Practice An updated version of the survey used from the latter part of year one included questions on the likelihood of implementing learning in the workplace, and levels of knowledge at the start and end of training sessions. Three-quarters of respondents (75%, n=272/363) thought there was a high likelihood of implementing learning in the workplace, with almost two-thirds (64%, n=222/348) stating this would be within a month. Responses varied by course type and area, with a smaller proportion of completed multi-agency responses (65%, n=93/143) and a larger proportion of completed specialist responses (83%, n=131/158) stating there was a ‘high’ likelihood of implementing learning. A smaller proportion of responses from tri-borough training sessions stated there was a high likelihood of implementing learning (67%, n=87/129), largely driven by multi- agency training session attendees. Self-reported assessments of learning highlighted a notable shift towards proportionately more attendees rating their level of understanding as ‘excellent’ at the end of training sessions (35% compared to 6% at the start overall: 23% vs. 1% in multi- agency sessions, 39% vs. 3% in specialist, and 41% vs. 18% in follow-up). There were also sizable increases in those rating their knowledge/understanding as ‘good’ (from 30% to 57% overall) and reductions in assessments of ‘fair/just satisfactory’ and ‘poor’ (from 44% to 7% and 20% to 1% respectively overall). 11 While the training surveys completed at the end of each session gave ‘in the moment’ feedback, the year two stakeholder survey attempted to explore the influence of training on practice back in the workplace. The majority of respondents (84%, n=58/69) had attended at least one type of training session. As table two below outlines, while most respondents agreed that the training had improved their ability to identify and respond to harmful practices, a lower proportion (just over three-quarters) felt they had been able to translate their training into practice, or that there was a clear ‘fit’ between what the training said they should do and what they could actually do in their role. A year two interviewee felt that training should go beyond standalone sessions and be embedded within a workplace commitment to review policies and practices: ‘I don’t think that it [training] should just be offered. I think it should be offered and then a commitment in some way of senior management of how it’s going to be implemented…So it’s not just the learner whom the onus is on. It’s on whomever they work under’. Table 2: Year two stakeholder survey responses Statement15 Agree Neither agree nor disagree Disagree Since the training I have been able to translate learning into practice in my role 76% (n=34/45) 16% (n=7/45) 9% (n=4/45) There is a clear ‘fit’ between what the training said I should do and what I can actually do in my role 78% (n=35/45) 11% (n=5/45) 11% (n=5/45) Overall, the training has improved my ability to identify harmful practices 89% (n=40/45) 4% (n=2/45) 7% (n=3/45) Overall, the training has improved my ability to respond to harmful practices 87% (n=39/45) 7% (n=3/45) 7% (n=3/45) Reflecting on their training since being back in the workplace, year two stakeholder survey respondents often highlighted greater awareness around harmful practices and sharing learning with colleagues: ‘[The training] has given me a different perspective on how we manage risks and our approach to conversations we have with our service users’ ‘I have been able to give practical advice and guidance to other practitioners around legislation and safeguarding issues’ ‘The training has made me more mindful of the circumstances that clients may have been through on their journey’ ‘I am more confident in my understanding of harmful practices…and am more at ease when I approach these topics… I have also been able to work and respond faster to these victims and get them help with other agencies’ 15 A number of questions in the survey were based on a Likert scale with respondents asked to indicate the extent of their agreement on a scale of 1 to 7 with ‘1’ indicating ‘strongly agree’ and ‘7’ indicating ‘strongly disagree. Figures used to indicate agreement to some extent are those responding ‘1’ to ‘3’. Figures indicating disagreement to some extent are those responding ‘5’ to ‘7’. Responses given as ‘4’ are considered to be ‘neither agree nor disagree’. 12 Year two interviewees referred to the training as a ‘legacy’ of the pilot that could be developed flexibly (e.g., to incorporate specific referral pathways and local support services) for rollout beyond the pilot boroughs, in particular targeting professionals who can ‘embed and cascade learning’. Safeguarding and Support: Advising and Guiding Professionals through the Educator Advocate Service/Surgeries Performance framework data returns indicated that across the five boroughs over the course of the pilot: • There were 591 surgery sessions (approximately 81% of anticipated delivery)16. • Surgeries were held in 17 different locations: social services (including children’s, family, and adult services) (6); hospitals/medical centres (3); children’s centres (2); One Stop Shops (2); a Multi-Agency Safeguarding Hub (1); a police community safety unit (1); a school (1); and a VCS organisation (1). • EAs provided advice and support in relation to 218 cases (102 in year one, 116 in year two), covering the four strands of harmful practices in the pilot, together with almost half (49%, n=106/218) defined more widely as VAWG, domestic abuse, or mental health related (see figure one)17. Figure 1: Cases supported by EAs by harmful practice type 16 On the basis of six successive quarters from the third quarter of year one onwards (performance framework data returns for the first two quarters of year one were patchy and inconsistent). No surgery data was returned for two boroughs in quarter three of year two. Anticipated delivery is calculated on the basis of ten surgery sessions per week across five pilot boroughs for six quarters (approximately 78 weeks), excluding the two boroughs for which data was missing in quarter three of year two (n=728 surgeries in total). Reasons for surgeries not taking place included public holidays, annual leave, tube strikes, and training. Attempts have been made to make up for missed surgeries with additional surgeries or training events. 17 As outlined in the pilot model, this service is predominantly offered to practitioners to enable them to better identify and respond to harmful practices; however, in some cases the EA may provide direct support to affected women and girls. This figure includes all cases returned in the performance framework (excluding obvious duplications). Performance framework returns were patchy in quarters one and two of year one. It is difficult to unpick reasons behind the high volume of VAWG and ‘other’ cases. This may point to the ‘continuum’ of VAWG issues of which harmful practices form a part, challenges identifying harmful practices, and/or that professionals welcomed support around both harmful practices and wider VAWG and other related issues. 11 13 8 36 17 17 4 9 1 30 53 19 0 10 20 30 40 50 60 FBA FM FGM HBV VAWG Other Number of cases YR1 YR2 13 The majority of individuals in these cases were female (8 cases involved males), ranging in age from children (aged one) to older adults (up to 75 years old) with over 30 self defined ethnic backgrounds18. The majority of cases were referred by services or agencies including social care, hospitals/health, police, education, VCS, and Independent Domestic Violence Advisors (IDVAs) (77%, n=159/206) or self referred (23%, n=47/206). Advice and support offered by EAs detailed in performance framework returns included accommodation/refuge, emotional support/counselling, welfare and benefits advice, language support, advice around appropriate action to take, risk assessment and safety planning, signposting to other services/professionals, no recourse to public funds, immigration, and legal matters, and outreach including visits alongside other practitioners. Referrals specified included to other professionals (e.g., IDVAs, domestic violence workers, FGM social workers), Multi-Agency Risk Assessment Conferences (MARACs), legal support, and third sector organisations. Opportunities and Challenges of the EA Service/Surgeries Year one analysis highlighted the potential value of the EA service/surgeries, in particular supporting professionals to see harmful practices through a ‘different lens’ giving them the confidence to make decisions, and EAs making useful contributions to areas beyond (but often intrinsically linked to) harmful practices (e.g., domestic abuse, immigration issues). Similar themes emerged throughout year two fieldwork with, as one stakeholder survey respondent commented, staff being supported to ‘think outside of the box and ask more questions’. Project leads and practitioners interviewed recognised the wealth of knowledge that EAs brought around harmful practices, wider VAWG issues, human rights, and available VCS support services, together with language skills and cultural understanding, which one reflected meant that statutory services could give ‘over and above what we are normally able to offer’. A practitioner interviewee from an organisation that hosted a surgery spoke about the support the EA provided to staff in raising their awareness of harmful practices and ‘knowing what kind of gentle nudges to make, or how to make people feel safe enough to disclose’, an approach they hoped would continue after the pilot finished. However, there were some ongoing challenges around the EA service/surgeries throughout the pilot: clarity around the role and developing a clear ‘offer’ to potential hosts; being located in a consistent team that deals with appropriate referrals in which the EA can build relationships where, according to a year two interviewee, ‘they can have more impact’; logistical issues around arranging a staff/building pass, securing access to the internet, computers and case files within host organisations; and being able to effectively review and influence harmful practice policies within statutory organisations. A common theme throughout steering group discussions and interviews was that surgeries were often slow. Indeed, only around half of respondents to the year two stakeholder survey (albeit small numbers - between 49% and 53% of 41 respondents) knew who the EA was in their borough, how to contact them, and what services they offered/what they could do to help them, with a slightly lower proportion (39%, n=16/41) knowing when and where 18 Demographic information was not provided in all cases. The level of information varied from case to case, with some including just one demographic field, and others full demographic information. Self-defined ethnicities stated included Afghanistani, Albanian, Algerian, American, Bangladeshi, British, Caribbean, Chinese, Dutch, Egyptian, Eritrean, Ethiopian, French, Indian, Iranian, Iraqi, Jamaican, Kurdish, Latin American, Mauritian, Moroccan, Nigerian, Pakistani, Portuguese, Punjab, Saudi Arabian, Serbian, Spanish, Sudanese, Turkish, and Uzbekistani. 14 surgeries were held in their borough. Only a fifth (20%, n=10/50) of respondents had attended an EA surgery or accessed support from an EA (with all who answered questions (n=7) commenting positively about the service received). Of those not accessing EA support, reasons given were not knowing who the EA was, what services they could offer or how to contact them, and having no relevant cases. Building EA-Practitioner Relationships in a Busy Landscape EAs often used surgery time to offer support in other ways: training and briefings for staff; advice/signposting around wider VAWG and immigration issues; and translation services. One project lead highlighted how useful this was; however, that it relied on the initiative and proactivity of the EA to build relationships with staff and seek out these opportunities. EAs interviewed often spoke about the challenges of building and sustaining relationships and needing to continually promote the pilot, particularly when they are present for only a half day a week and within host organisations undergoing staff and structural changes. One EA interviewee commented: ‘If I had gone there three times a week or if I was even based there, I’m pretty sure I would have a different relationship altogether because I would be part of the team. Going there once a week for two or three hours doesn’t really make a difference but if I was there on a very regular basis I would be like them…working together is important… it takes a lot of courage for a survivor to make that decision to call for help and that’s a good opportunity to give them all their options’. Another likened their role to ‘a helium balloon’ ‘flying around’ asking people if they can help, while an EA referred to the service in which they were located as ‘…like a beehive. People are in and out for meetings…every day was a promotion of the project because you keep seeing new people then you don’t see the person for the rest of the year’. This scenario was often compounded by a range of different professionals (e.g., IDVAs, Healthy Relationship advisors), as one interviewee stated ‘parachuting in and out’ of services, which created confusion for practitioners. One practitioner commented on: ‘lots of different projects floating around. There’s agreements made at service level which we don’t have much to do with so we just tend to get people turn up in the office and we get told who they are and what they can be helpful with’, while another stated: ‘It’s difficult to know who’s who and who to refer to and what the right service is…you just feel a bit overloaded as to who would be the best person and what is the difference’. Indeed, EAs sometimes struggled to ‘find their place’ alongside other experts, and practitioners similarly found it challenging to recognise the difference between them. Working Together to Support Clients Despite an inconsistent flow of cases, with practitioners themselves often commenting in interview that they didn’t always have enough of the ‘right’ type of cases for EAs to work on, examples shared by practitioners during year two interviews where an EA had supported them in their work indicated positive experiences: 15 ‘Early on they were very useful in facilitating and identifying refuges for people without public funds which was really helpful. More recently they did a visit with one of my social workers which was really good…we asked the right type of questions. Being very informed, very early on without us being involved in a family’s life when it wasn’t necessary’ ‘The EA came to a meeting…and did safety planning work for a young person going overseas at potential risk of arranged marriage. Useful practical advice offered, also information about use of Forced Marriage Orders’ ‘It’s not often we have a case that falls within their remit. When we do it is superb to have them on board…they’re our safety net’ EA interviewees commented on increased awareness and changes in attitudes that they observed amongst professionals they worked with. A project lead interviewee felt that the EA had supported practitioners to contextualise their understanding of harmful practice within a theoretical framework of power and control, which had the potential to change the way they viewed cases. In terms of survivors’ experiences of the Harmful Practices Pilot, EAs highlighted perceptions of an improved service offered through increased professional awareness and knowledge, and providing a ‘bridge’ between statutory services and their own VCS organisation. A practitioner felt that the EA service/surgery ‘…being based within statutory organisations might make services easier to access…for men as well as women. If clients were going to a specific women’s centre they may feel some stigma attached to this’. As part of year two fieldwork, EAs worked with the evaluation researchers to explore views of clients who had received support from the EA and other professionals during the pilot19. Clients shared a range of experiences: lack of awareness around their rights or help available, feeling disappointed and ‘not believed’ by professionals, and the support they had received from the EA or other service providers during the pilot including legal advice, referral to counselling and medical services, and access to safe accommodation. ‘Once I spoke with the Educator Advocate and the other advisor I felt more comfortable about the situation…I felt that the Educator Advocate was listening to what I am saying, and they were knowledge[able] about forced marriage’ ‘This was the first time I got support for my FGM as I was not aware of this support available…I was able to speak about the complications that this has caused me throughout my life’ ‘I felt that the Educator Advocate had listened to what I was saying… The information I received was good as I was not aware that I was able to receive this support from services…I have been passing this information to other people and I will continue to do so’ ‘I was relieved when the help and support [from various professionals including EA] came so fast’ 19 Following advice from EAs and PEHP partners, evaluation researchers developed a detailed discussion guide (together with clear instructions and information in order to secure voluntary informed consent) for EAs to use with clients. Although not without limitations (e.g., selection bias, EAs essentially ‘self-evaluating’), this was the most practical – and least intrusive - approach available for obtaining client views. It was also an interesting way to involve pilot staff in evaluation fieldwork. EAs spoke to four clients in total (two in the east and two in the tri-borough) including one case of forced marriage, one of FGM, and two of domestic abuse. 16 The EA role/surgeries have undoubtedly been one of the more challenging aspects of the pilot – particularly around volume of referrals, access to information, being able to influence policy, and EAs ‘carving out’ their role in busy and often changing organisational landscapes. However, there are positive findings around the volume of cases that EAs have offered advice and support around, that project leads and practitioners recognised the wealth of knowledge that EAs brought and valued the opportunity to work more closely with the VCS sector, and insightful examples of joint EA-practitioner working. Interviewees reflected on learning from the pilot commenting that more time planning and developing the EA role and location of surgeries, securing buy-in and commitment at strategic level, and building relationships from within organisations would improve delivery. Some thought that an ‘as and when required’ EA service, with support offered to practitioners via telephone, email and face-to-face when needed, would work well (rather than EAs being in the office when there weren’t any relevant cases) and allow the service to be flexed across a broader range of organisations. However, initially getting to know individual experts face- to-face and establishing rapport is important. As one practitioner commented: ‘It’s good when you see the face, establishing some kind of working relationship. Because [the EA is] here we’ve already established some kind of rapport with her. I don’t think she has to come in and sit with us every week. I can contact her. Having a face-to-face conversation makes a difference’. Community Engagement: Empowering Women and Girls Providing opportunities for the community to come together to learn, discuss and be part of the response has been recognised as a crucial factor in challenging harmful practices. Furthermore, engaging and building trust is particularly important for women and girls who may be scared to come forward due to repercussions for their family and wider community (Tedam, 2014; Norman et al., 2009; Larasi et al., 2014). Performance framework data returns indicate there were 51 community engagement events throughout the pilot (17 in year one, 34 in year two) attracting over 1,000 attendees20. Events included awareness raising sessions, workshops, support groups, conferences, ‘community conversations’, youth groups, and social events (coffee mornings, a summer BBQ, a festival, and a fair). Attendees included young people, professionals, those in temporary accommodation, and a range of individuals from different ethnic backgrounds. Events often took place as part of existing forums, rather than specifically arranged for the pilot, and sometimes incorporated a ‘social side’ including food, music, and wider health, wellbeing, and personal support, in addition to advice and education around harmful practices. Fieldwork observations picked up a positive, welcoming ‘vibe’ with events held in accessible community spaces (e.g., schools, colleges, community centres, and places of worship). A year two interviewee highlighted the importance of ‘focusing on safe spaces where people go….where communities are likely to build their own network’. 20 Approximate attendance was provided for 13 events in year one (194 attendees) and 31 events in year two (825 attendees). 17 Engaging and Informing Communities The majority of respondents to feedback surveys (between 92% and 96%, n=78) collected across seven events21 agreed that after the event they understood more about harmful practices, knew more about where to get support/advice, and would feel confident asking for support/advice if they/somebody else needed it. In freetext feedback comments, attendees spoke positively about the range of speakers, some of whom presented powerful, personal accounts of harmful practices, and engaging in conversations around power and control. Professionals often commented that they valued the opportunity to find out more about the pilot and other services in the area, which would help them better support clients. ‘Forced marriage – not legal, against women rights. It is a choice. Nobody has [the] right to force anyone’ ‘Forced marriage - today I know that it became illegal. That is very good news!! This must stop’ ‘FGM is absolutely not acceptable and it is not a cultural issue, it is about humanity’ ‘An engaging and uplifting session. Extremely important for women's progression and development’ Designing a Community Engagement Model The community engagement element of the pilot developed organically, with no set model outlining target audience or materials. While this proved useful in terms of identifying and capitalising on opportunities as they arose and flexing them to fit the target audience, project lead and EA interviewees felt that the pilot could have benefitted from a more co-ordinated, planned approach at times, including developing specific community engagement materials, in order to best harness the opportunities that dialogue with potentially affected communities offered. A year two interviewee felt this could have helped the events go beyond awareness raising to asking communities about their experiences to inform practice. The fluid approach to community engagement also made it difficult to systematically capture feedback and draw out learning22. However, interviewees reiterated the importance of all community engagement efforts around harmful practices: bringing together communities, breaking down isolation, opening up dialogue, and offering support networks. The reach of community engagement events can go beyond those who attend. The spoken word is powerful. As one interviewee commented: ‘even a small event can have a bigger reach than we think’. 21 Five in the east, two in the tri-borough including a mixture of professionals/practitioners/members of staff (35) and members of the community (21), with the remainder not stated. 22 Despite early efforts to communicate dates (e.g., the PEHP set up a community engagement calendar in a shared Dropbox which evaluation researchers were given access to) and regular reminders to EAs via email and at steering group meetings, it proved difficult to systematically find out about community engagement events in advance, therefore researchers were unable to attend or disseminate feedback surveys for completion (these were also made available via the Dropbox). Researchers attended a small number of events (five – at three of which feedback was collected) and EAs collected feedback at a further four events on behalf of evaluation researchers. 18 A ‘Partnership’ Pilot: Working Together to Tackle Harmful Practices A clear message about the importance of addressing harmful practices within a partnership approach, involving the statutory and voluntary sectors and communities themselves, underpinned all aspects of the pilot. The commitment of all delivery partners – both ‘on the ground’ and at strategic level – was often highlighted as a success of the pilot by interviewees. EAs spoke positively about being part of the PEHP partnership, in particular regular opportunities to meet and share experiences and ideas, while PEHP and borough project leads valued the focus that MOPAC placed on tackling harmful practices, which supported them in prioritising issues within their organisations. The realities of negotiating partnership working across a range of agencies was tricky at times, for example, around communication, the speed at which ‘business’ operates, and being able to ‘talk the language’ of other sectors. The number of partners involved in the pilot required strong, clear leadership, which some interviewees felt was inconsistent at times. Despite this, the pilot offered an important opportunity for VCS and statutory partners to work together to address harmful practices. A number of stakeholder survey respondents and interviewees from both statutory and voluntary sectors felt that each could benefit from learning from the other. An interviewee from a statutory service commented: ‘It gave us all an experience that we will draw on in the future - that you can work together more effectively to get better outcomes…it gave the voluntary sector a stronger foothold, voice… particularly the realisation of what they are already doing in the VAWG sector… I do think they are better placed to respond in a less stigmatised way and offer forms of help that people find more acceptable’ As an EA interviewee highlighted: ‘…we don’t only speak the language, we understand the culture, we understand…people from our community’. Interviewees from both statutory and voluntary services stated that they hoped partnerships would be sustained after the pilot had ended, with practitioners continuing to refer, advise, and seek support from each other. As a year two pilot staff interviewee commented: ‘We’re here for the same thing – to save lives – so for us to be able to put their specialism and ours together is really important’ 19 Discussion Summary This report has outlined some key learning from the MOPAC Harmful Practices Pilot, an initiative to improve the response to those who have experienced, or are at risk of experiencing, Female Genital Mutilation, ‘Honour’ Based Violence, Forced Marriage, and Faith Based Abuse, through training, support, and engagement with practitioners, professionals and communities. A number of positive messages have emerged throughout the pilot: • Training has been widely recognised as a key success with the 74 courses delivered throughout the two years received well by attendees in terms of quality, content, and improving practitioner knowledge around harmful practices. Once back in the workplace, respondents felt the training had improved their ability to identify and respond to harmful practices, and often highlighted examples/plans to share learning with colleagues. • EAs facilitated 591 surgery sessions in a range of organisations throughout the pilot period, providing advice and support in relation to 218 harmful practice and VAWG cases. Project leads and practitioners recognised the wealth of knowledge that EAs brought around harmful practices and beyond, and shared examples of cases where EAs and practitioners worked together to support clients. • The 51 community engagement events delivered were welcomed, with the majority of feedback received indicating that attendees understood more about harmful practices and where to access support after events. Practitioner attendees spoke positively about opportunities to find out more about services in their area, which would help them better support clients. • A strong partnership approach underpinned the pilot, with statutory and VCS service providers learning from each other around how to support those experiencing, or at risk of experiencing, harmful practices. However, the pilot has also presented challenges, particularly around the EA service/surgeries, where referrals have at times been slow, access to resources (e.g., internet, computers, and case files) limited, and EAs have been required to continually build relationships, promote the project, and ‘carve out’ their role within fast-paced, often changing surgery host organisations. While community engagement efforts were welcomed and positively received by those attending, pilot staff interviewees felt that some events could have achieved more with additional coordination and planning, including specific community engagement materials. The pilot highlights learning around the importance of comprehensive early project planning, and clear operating models, both at strategic and operational ‘on the ground’ level – challenges certainly not unique to the Harmful Practices Pilot. 20 Learning and Concluding Thoughts Throughout year two fieldwork, survey respondents and interviewees reflected on the ‘legacy’ of the pilot and sustainability of learning going forward. Many thought that the training should be developed flexibly (e.g., to incorporate specific referral pathways, local support services, etc.) into a product for rollout beyond the pilot boroughs. Learning from the pilot highlights the importance of embedding training within a workplace commitment to review policies to ensure that trained staff can put learning into practice in a supportive environment. Survey respondents and interviewees felt that refresher training or regular workshops were necessary to keep practitioner knowledge and awareness up-to-date, particularly for those who may encounter harmful practice cases less frequently. It is also vital to be alert to harmful practices beyond those included in this pilot, which may be changing or emerging, including breast flattening, ‘corrective’ rape, ‘widowhood rituals’ levied at older women, and other forms of violence and abuse which are defended on the basis of tradition, culture or religion (Action Aid, 2013). The Harmful Practices Pilot has started conversations, encouraged ‘professional curiosity’, and highlighted new ways of thinking amongst statutory and VCS agencies about under-researched and often misunderstood violent and abusive practices affecting thousands of individuals throughout the UK and millions across the globe. Learning from the pilot contributes to a limited yet developing evidence base and hopes to inform the approaches of other organisations in their work to tackle harmful practices. The challenge now is to continue to raise awareness and further develop partnerships amongst statutory and VCS agencies, and communities themselves, in order to build on learning to more effectively identify and support individuals experiencing, or at risk of experiencing, harmful practices. 21 References Action Aid UK (2013) Harmful Traditional Practices Affecting Women and Girls. London: Action Aid UK Bahunga, J. (2012) ‘Tackling Child Abuse Linked to Faith or Belief’. Every Child Journal 3 (3): 14- 19 Dawson, P. and Stanko, B. (2013) ‘Implementation, Implementation, Implementation: Insights from Offender Management Evaluations’. Policing: A Journal of Policy and Practice 7 (3): 289-298 Dawson, P. and Williams, E. (2009) ‘Reflections from a Police Research Unit – An Inside Job’. Policing: A Journal of Policy and Practice 3 (4): 373-380 Dickson, P. (2014) ‘Understanding Victims of Honour’ Based Violence’. Community Practitioner 87 (7): 30-33 Home Office/Foreign and Commonwealth Office (2017) Forced Marriage Unit Statistics 2016. London: Home Office/Foreign and Commonwealth Office Iranian and Kurdish Women’s Rights Organisation (IKWRO) (2011) Nearly 3000 cases of ‘honour’ violence every year in the UK. London: IKWRO Larasi, M., Roy, S. and Tweedale, R. (2014)”This is Not My Destiny” Reflecting on Responses to Forced Marriage in England and Wales. London: Imkaan/Rights of Women Macfarlane, A. and Dorkenoo, E. (2015) Prevalence of Female Genital Mutilation in England and Wales: National and Local Estimates. London: City University and Equality Now NHS Digital (March 2017) Female Genital Mutilation (FGM) Enhanced Dataset October 2016 to December 2016, England, Experimental Statistics. London: NHS Norman, K., Hemmings, J., Hussein, E. and Otoo-Oyortey, N. (2009) FGM is Always With Us: Experiences, Perceptions and Beliefs of Women Affected by FGM in London. London: Options and Forward Patton, M. (2002) Qualitative Research and Evaluation Methods. Third Edition California: Sage Publications Payton, J. (2015) ‘Honour-Based Violence: Policing and Prevention’. Policing and Society: An International Journal of Research and Policy 25 (1): 126-128 Roy, S., Ng, P. and Larasi, I. with Dorkenoo, E. and Macfarlane, A. (2011) The Missing Link: A Joined Up Approach to Addressing Harmful Practices in London. London: Imkaan/Equality Now/City University Stobart, E. (2006) Child Abuse Linked to Accusations of ‘Possession’ and ‘Witchcraft’. Research Report EE750 London: Department for Education and Skills Tedam, P. (2014) ‘Witchcraft Branding and the Abuse of African Children in the UK: Causes, Effects and Professional Intervention’. Early Child Development and Care 184 (9-10): 1403-1414 World Health Organisation (February 2016) Female Genital Mutilation Factsheet. http://www.who.int/mediacentre/factsheets/fs241/en/ 22 Appendices Appendix 1: Methodology – evaluation timeline and fieldwork conducted Performance monitoring Process Data captured around both individual cases and pilot delivery (returns were more complete from quarter three of year one onwards). Review of existing literature; Interviews with practitioners/ key partners involved in development TIMELINE Apr –Jun Jul-Sep Oct-Dec | Jan-Mar Apr-Jun Jul-Sep Oct-Dec | Jan-Mar 2015 2016 2017 Capture views of practitioners accessing EA services/surgeries via short surveys Interviews with EAs and project leads; Survey of practitioners/ key partners Interviews with EAs and project leads; survey of practitioners/ key partners Evaluation deliverables Regular updates provided to boards/meetings throughout pilot Learning update Mid-pilot report Final report (Jan 2016) (Spring/Summer 2016) (Summer 2017) Feedback from training attendees; observations at training sessions Capture views of services users (to explore feasibility with PEHP partners/practitioners) Feedback from community engagement event attendees; observations at community engagement events 23 Method Participants Date Performance framework and contract management data returns Provided by the PEHP From quarter 2 2015/16 onwards Attendance at MOPAC and borough steering group meetings 29 meetings attended April 2015 - May 2017 Theory of change interviews (face- to-face) MOPAC and borough project leads x 5 interviews/6 interviewees May – June 2015 Training feedback surveys 660 feedback surveys received August 2015 – February 2017 Observation at multi-agency and specialist training sessions Multi-agency session x 2 Specialist session x 2 September 2015 – January 2016 Year one key stakeholder interviews (face-to-face) MOPAC, borough and PEHP project leads x 7 interviews/8 interviewees October – November 2015 Year one online stakeholder survey 44 responses Social services/social work (14), VCS (10), health (9), children’s centre (4), housing (3), local authority (3), Community Rehabilitation Company (1). Waltham Forest (19), Tower Hamlets (8), Westminster (4), Kensington and Chelsea (3), Hammersmith and Fulham (7), Kensington and Chelsea and Westminster (1), across tri-boroughs (2). Five of these respondents worked across wider boroughs including Barnet, Brent, Croydon, Harrow, Merton, Newham, Redbridge, Richmond, Sutton and Wandsworth. November 2015 Observation at community engagement events Events x 5 November 2015 – December 2016 Year one Educator Advocate interviews (face-to-face) Educator Advocates x 5 November 2015 Community engagement feedback surveys 78 feedback surveys received from 7 community engagement events Professionals/practitioners/members of staff (35); Members of the community (21); remainder not stated. January 2016 – March 2017 End of year one Educator Advocate ‘catch up’ interviews (telephone) Educator Advocates x 5 April 2016 Year one surgery host/practitioner interviews (telephone) Surgery hosts/practitioners x 5 March- April 2016 24 Online EA/surgery attendee feedback The evaluation research team designed an online feedback survey to capture views of practitioners after they accessed support from the EA and/or attended a surgery. Success of this method was dependent on EAs remembering, and being able to (e.g., having contact details and access to Wi- Fi/a computer), send the survey link on to practitioners, and practitioners themselves choosing to complete it. Regular reminders were emailed to EAs and raised in focus groups. The survey received only two responses. These were not included in analysis for this report. June 2016 – April 2017 Service user feedback (collected by Educator Advocates) Clients x 4 November 2016 – January 2017 Year two online stakeholder survey 69 responses Social services/social work (16); VCS (16); health (12); Police/CJS (6); local authority (5); housing (4); education (2); children’s centre (1); remainder not stated. Waltham Forest (23); Tower Hamlets (11); Westminster (10); Hammersmith and Fulham (5); Kensington and Chelsea (9); Pan London/multiple boroughs (9); remainder non-pilot boroughs/not stated. March – April 2017 Year two key stakeholder interviews (face-to-face) MOPAC, borough and PEHP project leads x 5 interviews/7 interviewees March – April 2017 Year two Educator Advocate interviews (face-to-face) Educator Advocates x 5 March – April 2017 Year two surgery host/practitioner interviews (face-to-face and telephone) Surgery hosts/practitioners x 7 interviews/8 interviewees March – April 2017 25 Appendix 2: Professional backgrounds/organisations of training attendees (who stated their job role/organisation on training feedback surveys) Multi-agency course Number % Voluntary Community Sector (VCS) 60 18 Health - Other 53 16 Social Care - Children 42 13 Police/Criminal Justice 37 11 Social Care - Social Worker 24 7 Education 23 7 Housing 20 6 Social Care - General 4 1 Health - Sexual Health 3 1 Social Care - Other 3 1 Health - Midwife/Gynaecological 2 1 Not stated/unclear 65 19 Total 336 100 Specialist course, day 1 Number % Health - Other 26 20 Social Care - Social Worker 17 13 Voluntary Community Sector (VCS) 13 10 Social Care - Children 9 7 Health - Midwife/Gynaecological 4 3 Police/Criminal Justice 3 2 Housing 2 2 Social Care - Other 2 2 Education 1 1 Not stated/unclear 54 41 Total 131 100 26 Specialist course, day 2 Number % Health - Other 23 19 Social Care - Social Worker 12 10 Social Care - Children 11 9 Voluntary Community Sector (VCS) 11 9 Police/Criminal Justice 4 3 Health - Midwife/Gynaecological 3 3 Housing 2 2 Health - Sexual Health 1 1 Not stated/unclear 51 43 Total 118 100 Follow-up course Number % Voluntary Community Sector (VCS) 20 27 Health - Other 12 16 Social Care - Children 6 8 Social Care - Social Worker 6 8 Education 4 5 Health - Midwife/Gynaecological 2 3 Social Care - Other 1 1 Not stated/unclear 24 32 Total 75 100 27 Appendix 3: Training courses, attendees and feedback forms received; courses cancelled/rescheduled Count East (Tower Hamlets and Waltham Forest) Tri-borough (Westminster, Kensington and Chelsea, and Hammersmith and Fulham) Total Year 1 Year 2 Year 1 Year 2 Year 1 Year 2 Total Number of courses Multi: 9 Specialist: 3 Follow: 4 Risk: 0 Psych: 3 Safe: 1 Multi: 8 Specialist: 3 Follow: 9 Risk: 3 Psych: 3 Safe: 3 Multi: 12 Specialist: 3 Follow: 6 Risk: 2 Psych: 2 Safe:2 Multi: 11 Specialist: 3 Follow: 3 Risk: 1 Psych: 2 Safe: 0 Multi: 21 Specialist: 6 Follow: 10 Risk: 2 Psych: 5 Safe: 3 Multi: 19 Specialist: 6 Follow: 12 Risk: 4 Psych: 5 Safe: 3 Multi: 40 Specialist: 12 Follow: 22 Risk: 6 Psych: 10 Safe: 6 Number of attendees registered Multi: 126 SP1: 41 SP2: 26 Follow: 8 Risk: 0 Psych: 3 Safe: 5 Multi: 136 SP1: 65 SP2: 64 Follow: 60 Risk: 20 Psych: 29 Safe: 11 Multi: 114 SP1: 46 SP2: 46 Follow: 26 Risk: 6 Psych: 12 Safe: 8 Multi: 37 SP1: 34 SP2: 20 Follow: 8 Risk: 0 Psych: 8 Safe: 0 Multi: 240 SP1: 87 SP2: 72 Follow: 34 Risk: 6 Psych: 15 Safe: 13 Multi: 173 SP1: 99 SP2: 84 Follow: 68 Risk: 20 Psych: 37 Safe: 11 Multi: 413 SP1: 186 SP2: 156 Follow: 102 Risk: 26 Psych: 52 Safe: 24 Number of attendees attended/signed in Multi: 99 SP1: 31 SP2: 18 Individuals across SP days: 34 Follow: 4 Risk: 0 Psych: 0 Safe: 4 Multi: 113 SP1: 52 SP2: 48 Individuals across SP days: 56 Follow: 46 Risk: 16 Psych: 22 Safe: 8 Multi: 93 SP1: 34 SP2: 31 Individuals across SP days: 35 Follow: 17 Risk: 0 Psych: 9 Safe: 8 Multi: 36 SP1: 27 SP2: 14 Individuals across SP days: 30 Follow: 8 Risk: 0 Psych: 8 Safe: 0 Multi: 192 SP1: 65 SP2: 49 Individuals across SP days: 69 Follow: 21 Risk: 0 Psych: 9 Safe: 12 Multi: 149 SP1: 79 SP2: 62 Individuals across SP days: 86 Follow: 54 Risk: 16 Psych: 30 Safe: 8 Multi: 341 SP1: 144 SP2: 111 Individuals across SP days: 155 Follow: 75 Risk: 16 Psych: 39 Safe: 20 Number of feedback forms received Multi: 90 SP1: 30 SP2: 27 Follow: 11 Risk: 0 Psych: 7 Safe: 4 Multi: 98 SP1: 44 SP2: 42 Follow: 34 Risk: 13 Psych: 16 Safe: 5 Multi: 115 SP1: 31 SP2: 28 Follow: 23 Risk: 6 Psych: 9 Safe: 8 Multi: 34 SP1: 25 SP2: 21 Follow: 7 Risk: 0 Psych: 7 Safe: 0 Multi: 205 SP1: 61 SP2: 55 Follow: 34 Risk: 6 Psych: 16 Safe: 12 Multi: 132 SP1: 69 SP2: 63 Follow: 41 Risk: 13 Psych: 23 Safe: 5 Multi: 337 SP1: 130 SP2: 118 Follow: 75 Risk: 19 Psych: 39 Safe: 17 Specialist day 1 (SP1); Specialist day 2 (SP2) 28 Area Courses cancelled Courses rescheduled (but went ahead) Sign in sheets/feedback forms not received East 2 x follow-up risk assessment (Y1) 1 x multi-agency (sign in sheet and feedback missing) (Y1) 1 x specialist day 2 (sign in sheet missing) (Y1) 2x follow-up psychological impact (sign in sheet and feedback missing; sign in sheet missing) (Y1) 1 x follow-up risk assessment (sign in sheet and feedback missing) (Y1) 1 x follow-up risk assessment (sign in sheet and feedback missing) (Y2) 1 x follow-up psychological impact (sign in sheet and feedback missing) (Y2) 1 x follow-up safeguarding (sign in sheet and feedback missing) (Y2) Tri-borough 1 x follow-up risk assessment (Y1) 9 x multi-agency (Y2) 6 x follow-up risk assessment (Y2) 3 x multi-agency (Y2) 1 x specialist (Y2) 4 x multi-agency (sign in sheet and feedback missing) (Y1) 2 x multi-agency (sign in sheet missing) (Y1) 1 x multi-agency (sign in sheet missing) (Y2) 4x multi-agency (sign in sheet and feedback missing) (Y2) 1 x follow-up risk assessment (sign in sheet and feedback missing) (Y2) 1 A Randomised Control Trial of Mental Health Awareness and Safeguarding Training (MAST) March 2016 Emily Southall, Lynne Grossmith & Paul Dawson MOPAC Evidence and Insight 2 Executive Summary Mental health is a core business for MOPAC, policing and partners. In 2014 MOPAC set out key strategic ambitions to reduce the harm caused by gangs, split into the three elements of prevention, intervention and enforcement. The MAST programme feeds into these ambitions by consistently identifying and addressing mental health/emotional trauma of gang members. Furthermore, there is recognition of the need for frontline practitioners to be able to identify and address mental health issues in relation to gangs. MPS officers are often required to act as gatekeepers for those in crisis and regularly face the challenge of having to respond appropriately. Effective police intervention (i.e., identification, timely referrals and a partnership response) can play a significant role in managing mental health needs. Over the past decade there has also been an increase in both awareness and a concentrated effort to tackle gang crime – with London identifying approximately 3,500 offenders across 182 gangs who are responsible for 9% of all personal robbery; 16% of serious youth violence; and 40% of shootings. Individuals involved with gangs and violence are well recognised to have complex needs, experiencing higher levels of victimisation and a higher incidence of mental health needs, above both the general population and other entrants to the criminal justice system. In April 2014 the MPS and MOPAC introduced MAST (Mental Health Awareness and Safeguarding Training), with the core aim of promoting the safety of London’s youth, by ensuring practitioners take appropriate and timely action when there are signs of mental health needs and/or emotional trauma. Tasked with evaluating MAST, Evidence & Insight—the research capacity of MOPAC—used an innovative Randomised Controlled Trial (RCT) design, so any differences in responses to a knowledge survey could be directly attributable to the impact of the training. Over one thousand practitioners were included in the research. Key Findings  Results indicate that general practitioners (i.e., regardless of treatment or control) appeared to have a good base level understanding on issues surrounding mental health. However, feedback about the training was very mixed; less than a third of practitioners were satisfied with the MAST training and negative feedback was received in terms of delivery, venue, equipment and information. Practitioners were most positive towards the use of discussions within the training.  Despite the negative feedback, MAST was found to have a significant positive impact upon practitioner learning - in particular around understanding mental health, stereotypes, service delivery and overall confidence. Given the RCT methodology employed, we can have confidence in stating it was attending MAST that led to these improved learning outcomes.  The results indicate clear potential in MAST, which potentially could be enhanced if delivery issues were addressed. 3 1. Introduction Policing and Mental Health Mental health is a core business for policing. It is estimated that within London more than one million residents live with mental health needs, a higher prevalence than in other counties across the UK.i As a result, the MPS routinely come into contact with members of the public, victims of crime and offenders who are experiencing mental health needs.ii The MPS Territorial Policing Mental Health team estimate that mental health accounts for over 40% of policing workiii and the college of policing estimate that a typical officer might deal with an average of 14 incidents associated with mental health per day. To illustrate this demand, between April 2014 and March 2015, the MPS dealt with 79,811 emergency (999) or non-emergency (101) calls receiving a mental health qualifying code at an average of 14 incidents per day. In addition, the number of mental health related CRIS records captured by the MPS between 13/14 and 14/15 increased by 31% (from 7605 – 9978)iv indicating the increasing demand. This is a challenge that needs to be addressed. Within such encounters police often act as gatekeepers for those experiencing a mental health crisis and regularly face the challenge of having to respond appropriately. Effective police intervention (i.e., identification, timely referrals, strong partnerships etc.) can play a significant role in managing mental ill health. However, historically this has been a difficult challenge for policing to meet, with many notable cases of mismanagement and numerous government reviews/official reports calling for the police to improve their understanding and treatment of people with mental health needs.v The most recent of these being the Independent Commission led by Lord Adobowale subsequent to the death of Sean Rigg in police custody in 2008. These reviews consistently highlight challenges around officer training, knowledge, data capture and partnership working. Subsequent to the Adobowale commission there have been a number of innovations seeking to enhance police (and partners) working with mental health such as the Street Triage, recommendations from the Crisis Care Concordatvi and Liaison and Diversionvii - indicating positive strides on this critical issue. The Specific Needs of Gangs and Mental Health Over the past decade there has been an increase in both awareness and concentrated effort to tackle gang crime. To illustrate, this includes Government initiatives (e.g., The Home Offices' Tackling Gangs Action Programme, 2007), Ending Gang and Youth Violence (2015), MOPAC Strategic Ambitions document (2014), Policy changes (e.g., mandatory sentences for knife and gun crimes), and the launch of the MPS's Trident in 2012, along with a wealth of original academic research.viii The MPS estimates that in London, there are approximately 3500 gang offenders and 182 gangs and that gang members are responsible for 9% of all personal robbery, 16% of serious youth violence and 40% of shootings.ix Offenders are almost exclusively male (97%), of black ethnicity (72%) and the majority are under the age of 19 (47%).x Criminal career analysis, conducted by MOPAC Evidence & Insight indicated a typical gang member had on average 9 4 proven offences in their history, and were on average 15 years old when they committed their first offence.xi Other research demonstrates gang individuals are more likely to be stopped and searched by the Policexii and experience higher levels of victimisation. These findings indicate the increased potential number of encounters between police officers and other practitioners and gang individuals. Following a reduction in gang crime in 2012, the figure has since been increasing (1579 gang flagged offences in 2013 (CY), compared to 2094 in 2015 (CY)).xiii Research shows that gang members have a higher incidence of mental health needs, not just above the general population, but also above other entrants to the criminal justice system.xiv Broader research illustrates wider complex mental health needs associated with youth violence and gang membership, showing high levels of psychiatric morbidity, post-traumatic stress disorder, anti-social personality disorder, psychosis, suicide attempts and anxiety disorders.xv To further illustrate, within a sample of 100 young gang members, it could be expected thatxvi:  86 will have conduct problems (<18 years) or antisocial personality disorder (18+ years)  59 will have anxiety disorders (including post-traumatic stress disorder)  34 will have attempted suicide  25 will have psychosis  20 will have depression The relationships between gang-affiliation and poor mental health operate in both directions, with many of the factors that push or pull young people towards gangs relating to their mental wellbeing and involvement in gang-related activities potentially damaging mental health.xvii Such complex needs place a heavy burden on mental health services and highlight the need for an effective partnership response. What is MAST? In April 2014 MOPAC and the MPS decided to address the need for specific gang-related mental health training by introducing MAST (Mental Health Awareness and Safeguarding Training). The core aim was to promote the safety of young people in London, by ensuring practitioners take appropriate and timely action when there are signs of mental health needs and/or emotional trauma. The provision of joint agency training for practitioners1 aimed to break down a lack of understanding that can exist between professionals. There was a particular focus on the link between mental health, offending in general and the harm caused by gangs – both to gang members themselves and to vulnerable victims. In July 2014, MOPAC and the MPS secured the bid from the Home Office. The MAST programme was set to cost £1.41m over two years (£0.84m funded from Innovation Fund). The training consisted of structured delivery sessions, with a mixture of individual and group work, along with short facilitator inputs through video clips, testimonials, contextualised scenarios and case studies demonstrating relevant theory and practice linked to gangs and mental health issues.xviii The MOPAC Evidence & Insight team were tasked with evaluating MAST. 1 This includes; police, teachers, gang workers, Youth Offending Service, community safety officers, health practitioners, housing officers, young offender institutions, third sector organisations, other emergency services, Child Adolescent Mental Health Services (CAHMS). 5 2. Methodology A Randomised Controlled Trial design (RCT) was selected to evaluate MAST. In this type of design, the allocation of individuals (or groups/areas) to receive an intervention is determined by random chance, with analysis then comparing those that (randomly) did and did not receive the intervention. RCTs are conceptualised as the 'gold standard' in determining whether a cause and effect relationship exists between the ‘treatment’ and ‘outcome’.xix Previous research has successfully demonstrated the feasibility of using this design in evaluating training.xx The research randomly allocated practitioners to either control or treatment conditions. Surveys were split into two sections: 1) general feedback about the content and delivery of the training and 2) a series of knowledge based ('true' or 'false') statements and scaled attitudinal questions. The control group received section two before the training and section one after. The experimental group received section one and two together post-training. Due to the random nature of selection, any differences in responses to section two could be directly attributable to the impact of the training. Figure 1 below shows the process of when the surveys were sent out, in relation to the training. Figure 1. When surveys were sent out and randomisation for the MAST evaluation Randomiser – to assign practitioners to groups Experimental: Joining emails > NO survey Control: Joining emails > survey link (Section 2) Experimental: survey link (Sections 1 and 2) Control: survey link (Section 2) Before Training After Training 6 3. Results This section summarises findings from the trial including demographics of practitioners, feedback on the experience of attending MAST training, and the impact of the training. All practitioners who received MAST between 1st June 2015 and 10th January 2016 were included within the current study and randomly allocated to treatment or control. The sessions lasted for a whole day, and each participant only attended one session. Of the total 1020 practitioners, the majority of participants were female (70.8%, n=738), aged between 35-44 years (29.7%, n=305) and involved with policing (34.3%, n=273), followed by those in the education sector (18.2%, n=145). There was no difference between demographics comparing control and treatment groups indicating the randomisation was successful. The MAST Experience Practitioners reported far more negatives than positives when giving feedback on the MAST training. To illustrate, less than a third of practitioners reported being satisfied with the session (28%, n=280). In terms of delivery techniques, there was limited support for the PowerPoint slides (35%, n=367 reported them as useful), hand-outs (43%, n=458, reported them as useful) and exercises (51%, n=543, reported them as useful). Few practitioners reported the event materials being clear and easy to understand (15%, n=146). Only 32% (n=324) of practitioners felt the training was pitched at the right level, the majority felt the training was delivered too quickly (64%, n=648) and very few thought they were given time to ask questions (7%, n=68). The practitioners also gave negative feedback regarding the venue of the training itself, with only 14% (n=139) agreeing it was satisfactory with specific comments on the venue focused upon the lack of refreshments making the day unpleasant. In terms of the positives, the majority of practitioners reported the 'group interactions' were positive (70%, n=745) and were generally positive towards the facilitator (e.g., they were engaging (73%, n=767) and knowledgeable (66%, n=692)). Practitioners suggested the trainers were more akin to 'facilitators' with most learning coming from discussions with one another. ‘The training was not specific to anyone’s role, therefore it felt ineffective’ ‘The training was basic and not pitched to the right level of service’ ‘Discussions/sharing practice was very informative’ ‘Excellent for meeting other services and building links for the future’ ‘There was a lot of time for discussion and coming up with ideas together but very little actually taught’ 7 Of note for future sessions, around half of respondents (54%, n=535) said that learning about the 'resources available to individuals' would be beneficial and that they found the topic ‘gangs and associated mental health needs’ to be the most useful (60%, n=675). Finally, the understanding of cross-partner working is clearly something practitioners in both groups are keen to develop, as the majority (61%, n=535) stated ‘joint agency training’ would be useful. Whilst the practitioner experience of MAST appears to be more negative than positive, the above should be seen as learning opportunities. Each of the negatives raised by practitioners could be designed out and improved upon by revising the MAST training manual. Impact of MAST training It should be stated upfront that practitioners in general (i.e., regardless of treatment or control group) appeared to have a good base level understanding on issues surrounding mental health. In 8 of the 18 ‘true’ or false’ questions (all survey questions and % in appendix) there were no difference between treatment and controls - both were equally likely to answer correctly. These questions focused on the practical and common sense aspects of mental health knowledge, such as risk factors and types of mental health need. This was a positive and welcome finding given what has been written before around mental health training. However, analysis suggested there are significant differences between the experimental and control groups on a number of other questions relating to knowledge and confidence. Results are therefore outlined in two thematically based areas: 1) Understanding mental health and stereotypes and 2) Service delivery and confidence. Understanding Mental Health and Stereotypes The experimental group (i.e., those to receive the survey after the training) were significantly more likely to answer correctly to the following questions:  ‘people with mental health issues cannot be cured’ (more likely answer false: 94%, (n=413) vs 88% (n=443), p=.004).  ‘people with mental health needs are more likely to be violent’ (more likely to answer false: 81% (n=355) vs 73% (n=367), p=.006)  '… can identify warning signs that someone is gang affiliated’ (more likely to agree, 49% (n=212) vs 28% (n=138), p=.000). We see the majority of all practitioners answer correctly to the above questions; however, the treatment group do significantly better - indicating the training has successfully increased knowledge and reduced stereotypes. 8 Service Delivery and Confidence The experimental group were significantly more likely to answer correctly to the following questions:  'I know how best to help' (more likely to agree, 34% (n=151) vs. 16% (n=79), p=.000).  ‘I understand the services on offer from different organisations’ (more likely to agree, 45% (n=192) vs. 20% (n=97), p=.000).  The experimental group felt more confident in making a referral to the appropriate organisation (54% (n=233) vs. 32% (n=156), p=.001).  The experimental group were also less likely to feel worried about taking the wrong choice of action (34% (n=147), vs. 24% (n=118) (p=.001).  The experimental group were significantly more confident in understanding the most common types of mental health and emotional trauma (69%, (n=331) vs. 32% (n=152), p=.000).  The experimental group were significantly more confident in understanding how mental health issues may present themselves in young people (71%, (n=341) vs. 29% (n=141), p=.001) These findings indicate that MAST has had a positive impact upon service delivery and overall confidence for practitioners working with mental health and trauma. 9 4. Discussion Many reports have highlighted the complex relationship between mental health and the police alongside the challenges officers (and other practitioners) face. Evidence indicates that the police are facing an increasing mental health demand and often report feeling unprepared to meet this challenge. In this instance the development of MAST training is timely. In terms of the results - a decidedly mixed picture emerges. Less than a third of practitioners were satisfied with the MAST training, and there was negative feedback in terms of delivery, venue, equipment and information. The only aspect of MAST that practitioners were mostly positive about was the use of discussions within the training. These design issues are obviously important, although are all able to be designed out in subsequent iterations of MAST. However, such negative feedback did not appear to impact upon practitioners learning, as the research found that MAST had a positive impact upon practitioner learning - in particular around understanding mental health, stereotypes, service delivery and confidence. Given the methodology employed, that of a Randomised Control Trial, we can have confidence in stating that it was attending the MAST that led to these improved learning outcomes. Stage two of MAST was originally planned as part of the evaluation - although delays in rollout of the second part made further follow up questions impractical. However, further work would be welcome on this second stage to explore impact and issues of sustainability around the learning. Indeed, whilst this evaluation suggests knowledge has increased, this is no guarantee it will be retained or how it can be practically used. Nonetheless, the findings are most encouraging at this stage. Given there is a significant impact upon learning at this stage, especially considering the generally negative views of participants - imagine the results if these delivery issues were designed out. There is clear room for increased impact. Finally, the methodology employed here is also noteworthy. Many authors describe Randomised Controls Trials as complex, time consuming and/or expensive.xxi The current research demonstrates this is not the case and RCTs can be used in a cost free and quick time environment. Further opportunities should be sought to utilise this design when appropriate. 10 Appendix A Question Measure Experimental % (n) Control % (n) x2 p Resilience to developing mental health needs is not something that can be learnt False 84.6% (374) 79.9% (401) 3.585 .058 People with mental health needs are more likely to be violent False 80.5% (355) 73% (367) 7.412 .006 1/20 adults will be diagnosed with a mental health need at some point in their life time False 45.1% (199) 29.7% (149) 24.046 .000 Just under 50% of people don't mention their mental health needs on forms False 23.6% (104) 14.9% (75) 11.402 .001 Just over 60% of young people who reported having a mental health need stopped hanging out with friends False 28.6% (126) 21.5% (108) 6.267 .012 Just over 40% of young people who reported having a mental health need did not tell their families True 93% (410) 86.7% (435) 10.060 .002 Having a clear sense of identity promotes resilience to developing mental health needs True 88.9% (392) 80.9% (406) 11.583 .001 People with mental health needs cannot be cured False 93.7% (413) 88.2% (443) 8.186 .004 Just under 80% of those in a gang will have a drug dependence (mainly cannabis) False 40.7% (179) 33.9% (170) 4.672 .031 Good educational experience is a key factor in promoting resilience to developing mental health needs True 88.2% (388) 81.1% (407) 8.991 .003 80% of gang members will have conduct problems True 74.4% (328) 73.9% (371) 0.027 .869 Lack of social inclusion is a joint risk factor for those who have mental health issues and are gang affiliated True 93.4% (412) 95% (478) 1.123 .289 Mental health issues frequently affect people's ability to function True 78% (344) 79.3% (399) 0.244 .621 Risk factors are cumulative True 78% (344) 73.2% (368) 2.974 .085 Domestic violence is a joint risk factor for those who have mental health issues and are gang affiliated True 82.1% (362) 84.5% (424) 0.955 .328 A lack of positive interactions with others can be a risk factor in mental health issues True 91.4% (459) 91.8% (408) 0.050 .824 People with mental health issues are less likely to be a victim of crime False 97.2% (489) 96.1% (424) 0.850 .357 11 9/10 young people who describe mental health problems experience stigma and discrimination True 89.3% (394) 92.4% (465) 2.761 .097 I know how best to help Agree (scoring 7-10) 34.6% (151) 16% (79) 66.448 .000 I am worried about taking the wrong choice of action Disagree (scoring 1-4) 33.7% (147) 23.9% (118) 15.100 .001 I am concerned the individual will lash out and be violent Disagree (scoring 1-4) 53.2% (232) 49.2% (243) 1.864 .394 I am confident my colleagues will know what to do Agree (scoring 7-10) 46.6% (203) 24.7% (122) 51.072 .000 I can identify warning signs that someone is gang affiliated Agree (scoring 7-10) 49.3% (212) 28.3% (138) 54.537 .000 I can understand the services on offer from different organisations Agree (scoring 7-10) 44.7% (192) 19.9% (97) 79.889 .000 I am confident in making a referral to the appropriate organisation Agree (scoring 7-10) 54.2% (233) 32% (156) 56.398 .000 There is a good integration between services with those working with mental health needs Agree (scoring 7-10) 25.1% (108) 12.9% (63) 26.467 .000 There is confusion as to what responsibilities lie with which organisation Disagree (scoring 1-4) 22.1% (95) 12.3% (60) 17.553 .000 Q10: How confident are you in your understanding of the most common types of mental health and emotional trauma? Confident (scoring between 7-10) 68.5% (331) 31.5% (152) 165.25 9 .000 Q11: How confident are you in your understanding of how mental health issues may present themselves in young people Confident (scoring between 7-10) 70.7 (341) 29.3% (141) 202.22 9 .000 I am worried about the legal reprisals of the actions I take Agree (scoring between 7-10) 15.1% (66) 15.65 (77) 3.672 .159 There are barriers to timely and appropriate sharing of information between organisations Disagree (scoring between 1-4) 55.3% (241) 49.4% (244) 3.672 .159 There are lots of unknown circumstances Disagree (scoring between 1-4) 8.5% (37) 7.7% (38) 0.455 .796 12 References i Hobson, Z. (2015) Mental Health Street Triage Pilot. London (Internal MOPAC report) ii Independent Commission on Mental Health and Policing Report (2013) iii London Councils, January 2014 iv Hobson, Z. (2015) Mental Health Street Triage Pilot. London (Internal MOPAC report) v Bradley, K. (2009). The Bradley Report: Lord Bradley’s Review of People with Mental Health Problems or Learning Disabilities in the Criminal Justice System. London: DH; Flanagan, R. (2008). Final report of the independent review of policing commissioned by the Home Secretary and conducted by Sir Ronnie Flanagan. London: Home Office; Cabinet Office (2007). The Cabinet Office Policy Review, Building on Progress: Security, Crime, and Justice. London: The Cabinet Office; IPCC (2006). Deaths During or Following Police Contact. Available at: www.ipcc.gov.uk; including the Independent Commission on Mental Health and Policing chaired by Lord Adebowale, the subsequent Mental Health and Policing report the Mental Health Crisis Care Concordat and the Home Affairs Select Committee Mental Health and Policing Inquiry. vi Mental Health Crisis Care Concordat (2014). http://www.crisiscareconcordat.org.uk/ vii NHS Commissioning: Liaison and Diversion (2015). https://www.england.nhs.uk/commissioning/health-just/liaison-and- diversion/ viii http://www.centreforsocialjustice.org.uk/UserStorage/pdf/Pdf%20reports/DyingtoBelongFullReport.pdf; https://www.ncjrs.gov/pdffiles1/nij/188741.pdf ix Specialist Crime & Operations Intelligence Department (April 2014 – March 2015) x MetMis xi Strategic Ambitions for London: GANGS AND SERIOUS YOUTH VIOENCE. 2014. xii Stanko, B., Dawson, P. (2012) Reflections on the offending histories of those arrested during the disorder. xiii MetMis xiv http://psychnews.psychiatryonline.org/newsarticle.aspx?articleid=1726661; ‘A need to belong: What leads girls to join gangs’ Lorraine Khan, Helena Brice, Anna Saunders & Andrew Plumtree. Centre for Mental Health; http://www.centreformentalhealth.org.uk/pdfs/Bradley_Commission_briefing2_youngadults.pdf xv Coid, (2013). Gang membership, Violence, and Psychiatric Morbidity. xvi Madden, V (2013) Understanding the Mental Health Needs of Young People Involved in Gangs: A Tri Borough Public Health Report produced on behalf of the Westminster Joint Health and Wellbeing Board. The Westminster Joint Health and Wellbeing Board. http://www.mac-uk.org/wped/wp-content/uploads/2013/03/Mental-Health-and-Gangs-Report-2013.pdf xviihttps://www.gov.uk/government/uploads/system/uploads/attachment_data/file/398674/The_mental_health_needs_of _gang-affiliated_young_people_v3_23_01_1.pdf xviii http://www.masttraining.co.uk/mast-awareness.html xix The Magenta Book, 2001. Available at: https://www.gov.uk/government/publications/the-magenta-book. xx Skogan, W (2014). Training police for procedural justice. Journal of experimental criminology. xxi Gondolf, E. W. (2001). Limitations of Experimental Evaluation of Batterer Programs. Trauma, Violence and Abuse, 2, 79– 88; Hollin, C. (2008). Evaluating Offending Behaviour Programmes: Does Only Randomization Glister? Criminology & Criminal Justice, 8, 89–106. 1 Police Now Cohort 1 Final Evaluation Report December 2017 Julia Yesberg & Paul Dawson MOPAC Evidence and Insight 2 Contents Executive Summary ......................................................................................................... 3 1. Introduction ................................................................................................................ 5 Background to Police Now: A Changing Policing Landscape ................................................... 6 The Development of Police Now: The First Cohort .................................................................. 7 The 2015 Police Now Graduate Leadership Development Programme Design ........................ 8 The MOPAC Police Now Evaluation....................................................................................... 12 2. Methodology ............................................................................................................. 13 Evaluation Aims and Objectives ............................................................................................. 13 Methodology and Data Sources............................................................................................. 13 Structure of Report ................................................................................................................ 17 3. Results ...................................................................................................................... 18 The First Police Now Cohort .................................................................................................. 18 Delivery of the Police Now Graduate Leadership Development Programme .......................... 21 Life in the Met Police ............................................................................................................ 27 Participants’ Attitudes Toward the MPS and Serving the Public ............................................ 32 Evidence Based Policing and Police Now ............................................................................... 35 Police Now’s Impact on Public Perceptions and Crime .......................................................... 40 4. Discussion ................................................................................................................. 44 Delivery of the First Police Now Graduate Leadership Development Programme .................. 44 Transition to Front-line Policing ............................................................................................ 44 Experience of the MPS as an Organisation ............................................................................ 45 Positive Attitudes Toward Serving the Public ........................................................................ 46 Engagement with Evidence Based Policing Practices ............................................................. 46 Measuring Police Now’s Impact ............................................................................................. 47 Conclusion ............................................................................................................................. 48 Appendices .................................................................................................................... 49 References .................................................................................................................... 58 3 Executive Summary About Police Now and the Graduate Leadership Development Programme Previously incubated in the Metropolitan Police Service (MPS), Police Now became an independent charitable social enterprise in April 2016. Police Now’s aim is to “transform communities, reduce crime and increase the public’s confidence in policing” by recruiting graduates and placing them on the policing frontline. The Police Now Graduate Leadership Development Programme is an innovative programme that seeks to transform traditional police training, focussing upon recent graduates initially within an intensive 6-week Summer Academy before these new officers are placed in some of the most disadvantaged communities. The development of Police Now can be understood within the context of a broader development in policing, including the wholesale adoption of neighbourhood policing, the rise of evidence based policing (EBP), improvements to the quality of police training, the recasting of the public as police ‘customers’ and the materialisation of a new politics of police workforce diversity. Alongside these developments has been the growing acknowledgment that policing should be recognised as a profession of degree level complexity, rather than an occupation or craft. The MOPAC Police Now Evaluation A two-year evaluation of the first cohort of the Police Now Graduate Leadership Development Programme (henceforth referred to as the ‘Police Now programme’) has been conducted by the Evidence and Insight team at the Mayor’s Office for Policing And Crime (MOPAC), supported by funding from the Home Office Police Innovation Fund. Using a longitudinal mixed-methods approach, the evaluation had five main objectives: to explore the delivery and processes of the first Police Now programme; to understand how Police Now participants experience the job of policing; to explore Police Now participants’ attitudes and how they change over time; to examine participants’ willingness to engage in innovative evidence based policing practices; and to measure the indicative impact of the programme on public perceptions and crime. This is the final report of the two-year period. It is worth noting upfront that findings were shared with Police Now throughout the lifetime of the evaluation that led positively to improvements within the programme. These are acknowledged throughout the report, although it is beyond the remit of the current evaluation to assess such changes upon later cohorts. For interested readers, an evaluation of the second cohort is being conducted by academics at UCL and Birkbeck with published products expected mid-late 2018. The First Police Now Cohort • The first cohort of Police Now participants included 67 individuals. Almost half of participants in the cohort were female (45%, n=30) and overall 12% (n=8) were from Black, Asian and Minority Ethnic (BAME) groups. • They joined the MPS in July 2015, completed the Summer Academy in August 2015 and from September 2015 were placed into wards as Dedicated Ward Officers (DWOs). Of the 67 officers who were placed into wards, 55 graduated from the two-year programme. • Participants were posted in some of the most disadvantaged communities across London, with the majority posted in wards within the top 20% of vulnerable localities. 4 Key Findings: The Police Now Summer Academy • The Police Now participants were highly satisfied with the level and quality of training they received as part of the Summer Academy. The syndicate structure emerged as an important feature, along with the positive learning environment and the quality of visiting fellows. However, the majority of participants thought the Summer Academy was too short and both participants and syndicate leads thought the training should have had a larger focus on the practical aspects of the job and longer, more varied shifts during field training. • Fewer than half of participants felt prepared to begin their roles as police officers (43%, n=19) and one third felt confident (36%, n=16). However, confidence and preparedness also appeared to be an issue for recruits who joined via the traditional MPS entry route, suggesting it is not a problem unique to the Police Now officers but a wider issue for policing training. • Overall, satisfaction with the delivery of the Police Now programme was high and the overwhelming majority of participants would encourage other graduates to join the programme. Life in the Met Police • After a few months on the job, the majority of participants surveyed felt they had been accepted by colleagues; however, at the same time, around half reported a stigma associated with being part of the Police Now programme and a similar proportion felt uncomfortable disclosing they are part of the programme when meeting new colleagues. Participants and syndicate leads thought part of this was attributable to a lack of awareness and misconceptions about the programme. The proportion of Police Now participants who reported a stigma was lower than previous graduate programmes in the MPS (54% for Police Now vs. 73% for the MPS 2013 Graduate Entry Programme) and it is possible this stigma could reduce in later cohorts once Police Now has had time to embed within the organisation. • After a few months on the job, participants’ satisfaction with their role as a DWO was high and the majority reported being motivated at work. However, consistent with previous research into new police recruits, the officers’ job satisfaction and motivation at work declined over time. • Of concern, the majority of participants felt they were not given the resources needed to do their job and approximately half of participants felt they were not achieving a good work-life balance. These are not Police Now programme related issues but wider issues for the MPS. The Police Now cohort’s ratings were also reasonably consistent with other MPS staff with similar lengths of service. Attitudes Toward the MPS and Serving the Public • Although participants started their jobs with positive attitudes toward the MPS, over time, these attitudes declined. After almost two years working for the MPS, around one third of participants were confident of receiving a good service if they contacted the MPS as a member of the public compared to 75% at the start of the programme. This decline is not unique to Police Now participants and reflects overall findings from the MPS Staff Survey where satisfaction with the organisation decreases as service length increases. Overall, Police Now participants expressed more positive attitudes toward the MPS than all MPS staff combined. • Participants were less likely to want to spend their whole career in the MPS than a sample of other police officers surveyed as part of this evaluation. This finding is consistent with the recruitment model of Police Now - participants are encouraged to use the programme as a platform for developing their leadership career. • Encouragingly, the Police Now participants demonstrated a strong orientation toward serving the public and neighbourhood policing. Nearly all participants agreed that customer service was an 5 important part of a police officer’s role and that all members of the public should be treated with respect. A comparison of the Police Now cohort with other police officers showed Police Now participants held more favourable views toward serving the public. Evidence Based Policing Practices • Police Now participants appeared to have a good understanding of evidence based policing (EBP) although many participants thought their training did not effectively teach them how to use and apply EBP in their work and a number of participants spoke of the challenges they have faced implementing EBP in their ward. • The Police Now 100 day impact assessments represent an innovative process through which participants are required to feed back about the work they have done in the preceding 100 days to reduce crime and increase confidence in the police. As far as we know, there is no comparable work in UK policing currently. • Although these assessments are a move in the right direction, an inspection of the Police Now ‘Impact Library’ showed a lack of practical use of EBP evident within the assessments. To illustrate, out of a sample of 70 impact assessments, only 6 incorporated any previous learning from the literature and only a third sought to empirically demonstrate success; of these, only 4 used a comparison group. This finding illustrates that more work is required to support officers to practically embed EBP within their roles. Impact on Public Confidence and Perceptions of the Police and Crime • Key aims of Police Now are to reduce crime and increase the public’s perception of 'how good a job' the police are doing (i.e., confidence in the police). We explored these outcomes and a number of key drivers of public confidence and other questions related to neighbourhood policing using the MOPAC Public Attitude Survey (PAS) and ward level crime data. • Based on the total number of PAS respondents, across all the public perception questions explored (including public confidence aka ‘how good a job’), on average, the Police Now wards showed no change in public perceptions in the year subsequent to the programme compared to a matched set of wards. To illustrate, there was a 2% increase in ‘how good a job’ people thought the police were doing in Police Now wards compared to a significant 3% increase in comparison wards. This analysis incorporated 2,732 respondents from Police Now wards and 4,218 respondents within statistically matched comparison wards. • Looking deeper, there were inconsistent results. For example, improvements were observed in 'how good a job' young people (aged 16-24) thought the police were doing in Police Now wards (a significant 15% increase vs. a non-significant 4% increase in comparison wards). However, there were other drivers of public confidence that showed the opposite pattern; for example, there was a significant 11% increase in comparison wards in the proportion of young people who thought the police were 'dealing with the things that matter' compared to a non-significant 6% increase in Police Now wards. These findings relating to young people are based on relatively small numbers (i.e., 206 young respondents in Police Now wards in the year prior to the programme and 202 in the year after) and preclude firm ‘cause and effect’ statements on impact. • Police Now wards had higher crime volumes than comparison wards in the year prior to the programme, potentially revealing a limitation to the matching procedure. None-the-less, there were no differences in crime trends subsequent to the introduction of Police Now; both sets of wards showed comparable increases. 6 1. Introduction Police Now is an independent charitable social enterprise whose goal is to “transform communities, reduce crime and increase the public’s confidence in policing” by recruiting graduates and placing them on the policing frontline. This report presents findings of a two-year evaluation of the Police Now Graduate Leadership Development Programme in which the first cohort of participants had centre stage. It covers learning around the design, implementation and indicative impact of the Police Now programme. It is worth noting upfront that findings were shared with Police Now throughout the lifetime of the evaluation that led to subsequent improvements within the programme. These are acknowledged within the report, although it is beyond the remit of the current evaluation to assess such changes upon later cohorts. For interested readers, an evaluation of the second cohort is being conducted by academics at UCL and Birkbeck with published products expected mid-late 2018. The current report aims to present a major contribution to the policing evidence base, developing knowledge in key areas such as police recruitment, workforce development, police training and police culture. Background to Police Now - A Changing Policing Landscape Over recent years, policing has been undergoing a series of changes. These changes include the wholesale adoption of neighbourhood policing practices, the rise of evidence based policing (EBP), improvements to the quality of police training (e.g., following the Neyroud review), the recasting of the public as ‘customers’ and the materialisation of a new politics of police workforce diversity (i.e., recruiting more black and minority ethnic officers, females and university educated officers). 1 Alongside these changes, there has been a demand for officers to undertake increasingly complex, non-crime related work and a growing recognition that policing is a profession requiring degree level complexity.2 Despite this recognition, the police service has faced challenges attracting graduates who continue to have low perceptions of policing as a credible profession. In the UK Graduate Careers Survey (2015),3 83% of final-year student job hunters had never considered a career in policing. Notably, despite 75% of students agreeing a career in policing would be a rewarding role, 66% felt it ‘would be a waste of my degree’. In terms of making policing a more attractive career path, students were clear that a higher salary would help, coupled with a more clearly defined career path and increased chances of promotion. However, the current financial climate and challenges facing the police service pose a significant challenge in terms of attracting graduates to the police force. Interestingly, the public’s perceptions of whether police officers should be educated to degree level are mixed. Out of a sample of 3,195 representative Londoners, 47% thought it was important for police officers to have a degree, but the same proportion (47%) did not think it was important. A further 5% thought it depended on the role.4 1 Loftus (2009); Neyroud (2011); Paoline (2003) 2 Karn (2013) 3 The survey was based on face-to-face interviews with 18,412 final year students studying at thirty leading universities in the UK in February 2015. 4 From MOPAC’s Public Attitude Survey - administered to approximately 12,800 Londoners each year. 7 To make policing a more attractive option for graduates, flexible recruitment initiatives have emerged that are challenging the idea that policing is a ‘job for life’ (e.g., Direct Entry schemes). The development of Police Now can be understood within this context. In a survey with final year students, 61% stated the opportunity to join a programme like Police Now would make a career in policing more attractive, acknowledging the appeal of the leadership skills, employability and training the programme would offer them.5 Concurrent with the above developments, the College of Policing is establishing the Policing Education Qualifications Framework (PEQF) due for roll-out in September 2018.6 The PEQF is “intended to support the development of policing as a profession through the provision of a coherent national approach to raising and recognising educational standards in policing”.7 The PEQF is being designed to enable officers to obtain publicly recognised qualifications at the education level relevant to their role. Three entry routes into UK policing are currently being developed for new recruits: an undergraduate Bachelor’s degree in Policing, a graduate entry/conversion route for graduates with a degree in something other than policing, and a higher apprenticeship route. As part of the PEQF, the national training curriculum is being given an overhaul to address the shortcomings identified in police officer training in England and Wales.8 The Development of Police Now: The First Cohort Police Now was originally developed by frontline police officers as an innovative, two-year programme putting graduates on the frontline of policing. Commencing in 2015, Police Now takes inspiration from graduate recruitment models in other parts of the public sector; for example, the Teach First programme founded in 2002 and Frontline founded in 2013 (a recruitment scheme for social workers). Teach First is an independent social enterprise that was established as a new strategy for teacher recruitment in England and Wales. It sought to address the problem of educational disadvantage by recruiting the brightest graduates and placing them into schools within the most challenging environments. Overall, the evidence indicates Teach First has demonstrated a positive impact on GCSE results for pupils in low-income communities. Furthermore, schools in London where Teach First teachers are placed have moved from being the lowest performing in England to the highest performing.9 Like Teach First, Police Now aims to address the problem of disadvantage. Neighbourhood policing is “at the heart of the Police Now model”.10 Police Now brings graduates into policing with the purpose of transforming some of the most disadvantaged communities. Initially incubated in the Metropolitan Police Service (MPS), Police Now became an independent 5 Furthermore, almost half of the most recent intake of Police Now participants stated they would not have considered a career in policing without Police Now (based on survey data collected by Police Now; see their Impact Report 2017) 6 See http://www.college.police.uk/What-we-do/Learning/Policing-Education-Qualifications-Framework/Pages/Policing- Education-Qualifications-Framework.aspx for more information. 7 College of Policing, 2016 8 Neyroud (2011) 9 Teach First Our Impact https://www.teachfirst.org.uk/what-we-do/our-impact 10 Police Now Case for Change (2014) 8 charitable social enterprise in April 2016 and expanded to work with seven police forces across the UK: the MPS, Cheshire Constabulary, Lancashire Constabulary, Northamptonshire Police, Surrey Police, Thames Valley Police and West Midlands Police. In 2017 there was a further expansion of the programme and Police Now now works with just under half of the country’s police forces.11 The Police Now programme encompasses a number of significant innovations in the recruitment, development and deployment of police officers. It differs significantly from existing entry routes into the police force in that participants are recruited as police officers for an initial two-year period. At that point, they have the option to either remain in policing or to leave the police service to pursue a career elsewhere. This progressive recruitment initiative challenges some of the traditional underlying assumptions that policing is a ‘job for life’. The 2015 Police Now Graduate Leadership Development Programme Design Entry into the 2015 Police Now programme was dependent on participants passing an in-depth recruitment process (see Appendix B) consisting of online testing, a full-day at an assessment centre (including a competency based interview, role plays and presentations), along with passing the standard police tests (e.g., fitness testing, vetting). Successful applicants for the first year of the Police Now programme were then invited to attend a weekend induction in the spring of 2015 (see Figure 1 for the participant journey). Figure 1. Police Now ‘Participant Journey’ Spring Induction The Spring Induction was a one-day event that provided participants an opportunity to meet their cohort and members of the Police Now team, as well as giving them a taste of what they could expect from the Summer Academy and the two-year programme. The Spring Induction introduced participants to the realities of frontline policing and what they could expect from the police service as an employer. 11 The second cohort of Police Now consisted of 108 participants; at the time of writing this report 10 participants had left the programme. The third cohort consisted of 228 participants who completed the Summer Academy in August 2017. These cohorts are beyond the scope of the current evaluation. March '15 Spring Induction March-July '15 Pre-Learn July-Aug '15 Summer Academy Sept-Oct '15 28 Day DWO Immersion Period Oct '15-July '17 DWO Posting Oct '15- July '17 Ongoing Skills Sessions and 100 Day Impact Events July ‘16-July '17 Internal or External Secondment July '17 Programme Graduation 9 Pre-Learn Course Prior to the Police Now Summer Academy, participants completed a comprehensive pre-learn course. The pre-learn course—written by serving police officers and subject matter experts to complement the Blackstone’s Handbook for Policing Students—was designed to ensure participants had all the legal and procedural knowledge needed at the start of the Summer Academy. The pre-learn course was assessed on the first day of the Summer Academy. The 2015 Summer Academy The intensive six week long Summer Academy was developed to represent an innovation to police officer training and leadership development and aimed to build on and enhance traditional police training. The Police Now training was based on National Standards (e.g., College of Policing Learning Objectives) and included all mandatory training (e.g., Occupational Safety and Emergency Life Support training) plus additional Police Now training (see Figure 2). The four specified objectives of the Summer Academy developed by Police Now were: - To enable participants to excel at delivering the Police Now aim of transformation in communities. - To enable participants to be exceptional police officers. - To enable participants to be leaders on the policing frontline and in wider society. - To deliver the gold standard in policing and leadership training and development. Figure 2. Hours breakdown of the Summer Academy The Summer Academy delivered in July-August 2015 consisted of a mixture of classroom and practical training and was highly intensive; participants attended 11-hour days over 6 weeks consisting of approximately 300 hours of learning (see Figure 3 for the Summer Academy timetable).12 Throughout the Summer Academy, visiting fellows drawn from inside and outside of policing delivered lectures, seminars and practical training on their areas of expertise. In the last three weeks of the Summer Academy, participants took part in Borough Field Training to 12 Compared to 480 hours in standard entry training. 36 27 13 17 7 0 5 10 15 20 25 30 35 40 45 50 Assessment & Immersive Learning Traditional Police Training Police Now Additional Training Mandatory Training Administration % of Summer Academy 10 develop their skills and knowledge in a real life environment. The field training consisted of 15 shifts (5 Emergency Response Patrol Team shifts, 5 Safer Neighbourhood Team shifts and 5 Criminal Investigation Department/Crime Prevention Unit shifts) under the 1:1 supervision of an experienced officer. Throughout the Summer Academy, participants underwent regular theoretical and practical assessments that meet and exceed current MPS foundation course assessments. Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 0800-0850 Written Assessment, Practical Assessment or Reflection Session 0900-1300 Classroom and Practical Exercises 1300-1400 Lunch and/or Travel 1400-1900 Mandatory Officer Safety Training/Emergency Life Support and Practical Exercises Practical Exercises and Borough Field Training Sessions Figure 3. Police Now Summer Academy Timetable The syndicate structure was an integral part of the Summer Academy and the Police Now programme. Police Now participants were broken into 8 syndicates of 8-9 participants led by a Syndicate Lead who was a substantive police officer with recent frontline policing experience (see Figure 4). The syndicate lead’s role was to inspire and guide participants through the Summer Academy while also preparing and delivering training content. They were responsible for monitoring participant performance, providing advice, and identifying further development needs. Syndicate leads were also responsible for identifying and responding to individual welfare needs and resolving problems when they arose. Figure 4. Police Now Summer Academy Staff Structure Gold: Programme Lead Silver: Course Director Bronze: 8 Syndicate Leads 8 Syndicates: Each with 8-9 Police Now Participants 11 The 28 Day DWO Immersion Period Following the Summer Academy, the 28 day DWO immersion period was designed to introduce participants to the DWO role under the 1:1 supervision of an experienced DWO mentor based on a nearby or neighbouring borough. As part of the immersion period, participants were required to complete the Student Officer Record of Competence (SOROC) assessments and were also expected to commence their first 100-day plan for their community (see below). Two-Year DWO Posting Throughout the first year, Police Now participants attended training sessions focused on professional policing and their own leadership journey. They were also allocated a mentor to support them and aid their professional development throughout the process. During year two, participants continued to develop their skills as a police officer. As well as continuing to attend skills sessions throughout the year, participants were also given the opportunity to apply for a secondment either internal to policing or with one of Police Now’s ‘graduate partner’ organisations. 13 The secondment was designed to increase participants’ understanding of operational policing in external organisations and broaden their skillsets within and beyond the policing environment. Participants were also assisted to find a mentor to support their personal development. It was envisaged that these mentors would share their knowledge with the participants and act as a safe forum for any ideas, concerns or questions. With their guidance, participants would be able to identify the best way to develop their careers. 100 Day Impact Assessments During their two-year postings, participants completed 100 Day Impact Assessments. In these assessments, participants used a range of visualization and presentation techniques (e.g., videos, blogs, oral presentations, posters) to answer the question: What is the most impactful thing you have done on your Ward to reduce crime and increase the public’s confidence in policing during the last 100 days? Participants were encouraged to focus on measurable impact and to describe 13 Participants were able to apply for a secondment with one of the partner organisations during their second year for a period of 2 to 4 weeks. Over 20% of the 2015 cohort successfully completed industry secondments with corporate partners, including KPMG, WPP, the Home Office, National Crime Agency, Accenture and PA Consulting. Approximately half of the 2015 cohort completed internal attachments with specialist crime teams and investigation units. The Dedicated Ward Officer role Each ward in London has at least one Dedicated Ward Officer (DWO) who is situated within a Safer Neighbourhood Team (SNT) - led by a Sergeant and comprising two Police Constables and at least three Police Community Support Officers. The DWO role provides a vital link with the community and is solely dedicated to the ward. Other officers within the SNT may work flexibly across a larger area. In the Mayor’s 2017-2020 Police and Crime Plan he pledged to increase the number of DWOs to at least two in every ward, acknowledging their important role in engaging with the community. 12 innovations they had made over the preceding 100 days. Participants fed back their work at Impact Events throughout the two years which were attended by a range of people, including serving police officers, academics, civil servants, charities, and councilors. The work was assessed by Police Now with a set of criteria, including their initiative in taking action, use of evidence based policing, and how they communicated their work. Outside of Police Now, as far as we are aware, no other neighbourhood police officers are held to account in this way. The assessments therefore represent an attempt by Police Now to ‘raise the bar’ through which officers evidence the impact of their work. The MOPAC Police Now Evaluation Evidence and Insight (E&I) are the dedicated analysis and research function within the Mayor’s Office for Policing And Crime (MOPAC). The team is an amalgamation of a long-established analysis function within MOPAC and a team of social researchers that moved across from the MPS in 2014. E&I conducted a large-scale two-year evaluation, explicitly focussing upon the first cohort of the Police Now programme, with a priority to achieve a balanced, independent and factual report. The research used a longitudinal mixed-methods approach to explore both the processes and indicative impact of the Police Now programme. This report pulls together two years of learning and in doing so provides valuable learning for Police Now and wider stakeholders and interested parties, as well as making a significant contribution to the policing evidence-base, developing knowledge in key areas such as police recruitment, workforce development, police training and police culture. 13 2. Methodology Evaluation Aims and Objectives The evaluation of the first Police Now cohort had five key objectives: 1. To explore the delivery and processes of the first Police Now Graduate Leadership Development Programme, particularly focusing on the Summer Academy training and the 28 day DWO immersion period. 2. To understand how Police Now participants experienced the job of policing and the DWO role. 3. To explore how Police Now participants’ attitudes changed over time, including their attitudes toward the MPS as an organisation and serving the public. 4. To examine participants’ willingness to engage in innovative evidence based policing practices and assess how well the Police Now programme equipped them to be champions of evidence based policing. 5. To measure the indicative impact of Police Now on public perceptions and crime. Methodology and Data Sources The evaluation used a longitudinal mixed-methods approach. It combined a number of quantitative and qualitative research methods and drew on several data sources to explore the processes and indicative impact of Police Now (see Figure 5 below for a timeline of the evaluation). Figure 5. Police Now Evaluation Timeline Jul-15 Aug-15 Sep-15 Oct-15 Nov-15 Dec-15 Jan-16 Feb-16 Mar-16 Apr-16 May-16 Jun-16 Jul-16 Aug-16 Sep-16 Oct-16 Nov-16 Dec-16 Jan-17 Feb-17 Mar-17 Apr-17 May-17 Jun-17 Jul-17 Aug-17 Sep-17 Oct-17 Nov-17 Dec-17 Tracker Survey 1 Tracker Survey 2 Tracker Survey 3 Tracker Survey 4 Cohort, Line Manager Interviews Leadership Team Interviews Comparison Officer Interviews Comparison Officer Survey Comparison Officer Survey Summer Academy DWO Immersion Period DWO Posting Programme Graduation Interim Report Year 1 Report Interim Report Final Report Evaluation Reports Comparison Surveys Interviews Tracker Surveys Police Now Programme Timeline 14 Tracker Surveys capturing Police Now officers’ views over time Participant ‘tracker surveys’ formed the central thread of the feedback loop in this evaluation. The tracker surveys were administered longitudinally to the Police Now cohort to capture how their motivations, job satisfaction and attitudes associated with the police culture and job of policing changed over time. Over the course of the two year programme, a total of four tracker surveys were administered to the participants. The first tracker survey was administered to the cohort in July 2015 during the induction week at the Summer Academy. The survey was designed to ascertain participants’ previous experience of working with or for the police and to measure and benchmark their motivations for joining the police, their career aspirations, their expectations of the MPS as an employer, and their early career attitudes toward service-based policing philosophies. Each participant was provided with a paper copy of the survey to complete, which included a cover page outlining the purpose of the survey and ethical considerations (e.g., confidentiality), with a response rate of 100% (n=69). The second tracker survey was administered in October 2015 (n=45; response rate of 67%), the third in July 2016 (n=38; response rate of 60%), and the fourth in February 2017 (n=23; response rate of 40%).14 These surveys were completed online by participants. They included identical questions to capture change in attitudes over time.15 An additional section was added to the third tracker survey to capture participants’ understanding of—and experience implementing— evidence based policing. For each tracker survey, participants were emailed a link to complete an online survey and were informed of confidentiality and anonymity. Participants were asked to provide their warrant number so that their experiences and attitudes could be measured longitudinally. Interviews capturing experiences of Police Now A total of 18 face-to-face interviews were conducted with syndicate leads and the Police Now leadership team in October – November 2015. The interviews were designed to capture their experiences of the Summer Academy (including what worked and what could be improved upon), insights into their own career and how Police Now differs to the training they received as new recruits, and the current state of policing and the MPS. Interviews were recorded, transcribed and thematically analysed. Telephone interviews were conducted with a sample of Police Now participants (n=22) and a sample of participants’ line managers (n=25) to further explore themes uncovered in the tracker surveys.16 Interviews were conducted in February – March 2016 and lasted approximately 40 minutes each. Interviewees were recruited via email or in person at the first 100 Day Impact Event and were asked to sign a consent form outlining confidentiality and anonymity. 14 The survey was developed and hosted using the Opinion Research Services (ORS) secure online survey portal. Response rates are adjusted for programme exits. 15 The online nature of the follow-up tracker surveys may explain the attrition rate in the number of participants completing the follow-up surveys. 16 These interviews were completed by ORS’s qualitative research team. 15 Lastly, telephone interviews were conducted with a sample of police officers (n=13) who joined the MPS in July 2015 via the ‘traditional’ entry route.17 Interviews were completed in June 2016 and lasted approximately 40 minutes each. Other Surveys A number of additional surveys were administered to capture the experiences of police officers not on the Police Now programme. An online survey was sent to: (1) the general (i.e., not Police Now) MPS intake from July 2015 (n=111; response rate 31% or n=34); (2) the general MPS intake from July 2016 (n=70; response rate 14% or n=10); and (3) a sample of non-Police Now DWOs (n=113; response rate 27% or n=31).18 Because of the low numbers completing the surveys, the three comparison groups were combined to give a total sample of 75 police officers (a response rate of 26% from the 294 officers invited to participate): whilst a relatively small sample, it is enough to draw tentative insights and learning.19 The evaluation also draws on data collected from surveys administered to participants by the Police Now operational team at fortnightly intervals throughout the Summer Academy and a short survey administered to the line managers of participants following the Summer Academy. Comparisons to other surveys (e.g., conducted with recruits from the 2013 MPS Graduate Entry Programme and the 2015 MPS staff survey) are also made. MOPAC’s Public Attitude Survey measuring public confidence and perceptions of the police Part of Police Now’s aim is to increase public confidence in the police. Historically, public confidence has been measured using a single question: How good a job do you think the police in your local area are doing? Previously, MPS researchers along with academic support developed a ‘confidence model’ identifying key aspects that influenced ‘how good a job’ people thought the police were doing (the two strongest drivers being fair treatment and effective engagement).20 The present analysis looked at the ‘how good a job’ question, but importantly it also assessed a number of the key drivers of public confidence and other questions related to the public’s perceptions of neighbourhood policing and the DWO role (see Appendix C for a list of the questions used). These broader questions on public perceptions are arguably more actionable and measurable than the ‘how good a job’ question and are all aligned with neighbourhood policing and Police Now’s aims. To explore these areas, data from the MOPAC Public Attitude Survey (PAS) was used.21 Wards in London where a Police Now officer was posted (‘Police Now wards’) were compared to a group 17 These interviews were completed by ORS’s qualitative research team. 18 55% (n=17) of DWOs had less than 5 years service, 35% (n=11) had between 5 and 15 years service and 10% (n=3)had more than 15 years service. 19 Gender and ethnicity information was collected for the 2016 MPS intake and DWOs. Compared to the Police Now cohort, the comparison group included more Males (80% vs. 55%) and BAME (29% vs. 12%) respondents. Along with these differences, there could be other measurable differences between the two groups that we are unable to account for; therefore, any comparisons made should be interpreted with caution. 20 Stanko & Bradford (2009) 21 The PAS consists of 12,800 face-to-face interviews with Londoners each year and is representative at the London level. The methodology broadly mirrors the Crime Survey for England And Wales. 16 of wards without a Police Now officer. The 60 Police Now wards22 were statistically matched to non-Police Now wards using the Vulnerability Localities Profile (VLP).23 Comparison wards were selected if they were nearest to one of the Police Now wards on the VLP (i.e., wards most similar in their vulnerability score). A total of 91 comparison wards were identified (out of 537 possible comparison wards). The Police Now wards’ average VLP score (M=130.7) was equivalent to the comparison wards’ (M=130.3)24, confirming the success of the matching procedure. The average VLP score was in the top 20% of all wards in London for both the Police Now and comparison ward groups. PAS data was gathered for these 151 wards (60 Police Now, 91 comparison) from the year prior to Police Now (October 2014-September 2015) and the year subsequent to the introduction of Police Now (October 2015-September 2016). The total number of respondents for Police Now wards for the year ending September 2015 was 1,450, and 1,282 for the year ending September 2016. The total number of respondents for the comparison wards for the year ending September 2015 was 2,319, and 1,899 for the year ending September 2016. Crime Data The same comparison wards were used to explore the impact of Police Now on ward crime levels. Crime data was extracted for the year prior to Police Now (October 2014-September 2015) and the year subsequent to the programme (October 2015-September 2016) for both the Police Now and comparison wards. Total notifiable offences were examined (all offences recorded during the year), along with specific crime types (e.g., robbery, burglary, theft, motor vehicle offences, criminal damage, knife crime). Officially record crime statistics do not capture the large number of crimes that go unreported; however, using official crime data is commonplace in criminal justice research. Future evaluations of Police Now could incorporate unofficial crime data, including information from public surveys (e.g., the Crime Survey of England and Wales) and data from the charity and voluntary sector. Police Now Impact Library Police Now’s Impact Library is an innovative prototype designed to communicate the work that Police Now participants are doing in their neighbourhoods. The Impact Library showcases participants’ 100 day impact assessments and was used in this evaluation to assess participants’ levels of implementing effective EBP. A set of criteria was developed to measure the cohort’s impact assessments against a variety of evidence based workings (see Appendix D for the coding criteria). For example, the evaluation assessed whether they incorporated learning from previous research, whether a clear problem was identified through the use of data, whether success was clearly defined and outcomes measured and whether caveats were included in the reporting of 22 Two wards had more than one Police Now officer posted. Some participants who exited the programme were included in the analysis because they left the force after the relevant time period. 23 The Vulnerable Localities Profile is an academic tool developed originally by University College London to explore the relative safety of locations in London at ward level. It integrates data on crime (burglary and criminal damage rates), deprivation (claimant count rate, GCSE scores, average household income), and population and is accessible via the London Landscape: https://www.london.gov.uk/what-we-do/mayors-office-policing-and-crime-mopac/data-and-research/london-landscape. The VLP score used in this report was for the financial year 15/16. 24 According to an independent samples t-test; t(149)=1.98, p=.92 17 the results. A total of 70 impact assessments (65 posters and 5 presentations) were randomly selected to be assessed using the coding scheme.25 There are a number of important caveats to note up-front here: (1) the purpose of the impact assessments and how they were assessed by Police Now included a wider focus than EBP; (2) we can only go on what was written down by participants on the poster or presentation; and (3) for time considerations, our analysis focussed on predominantly posters rather than presentations, which may mean we are missing more detailed information that may be included in presentations. Despite these caveats, the impact assessment posters remain a key record of the officers’ workings and provide valuable insight into their practical use of evidence. Structure of Report This report follows the journey of the first cohort of Police Now participants from induction into the Summer Academy through to completion of the two year programme. It starts by introducing the first cohort of participants, including their background and expectations of the programme before exploring the delivery of the programme from the perspective of the participants, their line managers, and syndicate leads. Participants’ experiences of being a police officer and their attitudes and motivations are then examined, with a particular emphasis on how their attitudes change over time and how they differ to police officers not on the programme. The last two sections of the report explore participants’ orientation toward and use of EBP and Police Now’s indicative impact on public perceptions and crime. Although this report is designed to provide a comprehensive overview of the findings from the evaluation, further detail and additional findings from the research can be found in the three previous interim reports.26 25 More posters were chosen over presentations purely for ease of analysis and time considerations. 26 Year 1 report available here: https://www.london.gov.uk/what-we-do/mayors-office-policing-and-crime-mopac/data-and- research/academic-research. 18 3. Results The First Police Now Cohort Before addressing the five key objectives of the evaluation, this first section introduces the first Police Now cohort, including who they are, where they were posted as police officers across London, their background and motivation for joining the police and their expectations of being a police officer in the MPS. Who are the first Police Now participants? The launch of Police Now in 2015 attracted 2,200 registrations of interest and 1,248 applicants.27 The first cohort of 69 Police Now participants was selected from this applicant pool and joined the MPS in July 2015. Almost half of the 67 participants who completed the training were female (45%; n=30) and 12% (n=8) were from Black, Asian and Minority Ethnic (BAME) groups (see Figure 6). The proportion of females in the Police Now cohort is larger than the overall proportion of female police officers in the MPS (25%) 28 and the 2013 Police Officer Graduate Entry programme (40%). 29 The proportion of participants from BAME groups is double the proportion of BAME recruits in the 2013 graduate cohort (6%), but similar to the overall proportion of BAME police officers in the MPS (11%). The proportion of BAME officers in the 2016 cohort increased to 20% and the proportion of females increased to 54%. Police Now participants30 45% female from a BAME background Figure 6. Police Now participant demographics Police Now Participant Ward Postings After completing the Summer Academy in September 2015, participants were posted in wards in London as Dedicated Ward Officers (DWOs; see Figure 7 for participant ward postings). The cohort was posted in a total of 15 boroughs across London. Of the 60 Police Now wards, 63% were in the top 20% of the most vulnerable wards in London (according to the Vulnerable 27 17 applicants for every place offered on the 2015 programme. A diverse applicant pool was achieved: 18% BAME and 49% female. 28 MPS Staff Survey (2015) 29 Fenn (2014) 30 69 participants started the programme but 2 left during the Summer Academy. Over the course of the two year programme, 12 participants exited the Police Now programme and left the MPS. All individuals who exited the programme were invited to take part in an interview about their experiences of the programme; one person took part and therefore these results will not be presented in the report. 67 12% 19 Localities Profile), suggesting Police Now has selected wards with some of the most disadvantaged communities in London.31 Figure 7. Police Now participant ward postings (August 2015) Background and motivation for joining the police In Tracker Survey 1 (TS1), 22% (n=15) of participants noted they had prior experience of working with or for the police, either as Special Constables, Volunteer Police Cadets, or as a member of Police Staff. Almost one third of participants (32%; n=22) had a family member or close friend with experience of working for the police, the majority (82%; n=18) as a police officer. These proportions are significantly lower than the MPS 2013 Graduate Entry Programme cohort. In this cohort, 40% (n=43) had prior experience of working with or for the police and 49% (n=53) had a family member or close friend with experience. When asked to select three main reasons for pursuing a career as a police officer, the most common reason for Police Now participants was that it was ‘an exciting role’, (52%; n=36). Other common responses were ‘the opportunities offered by Police Now’ (46%; n=32) and ‘a job with variety’ (33%; n=23). Encouragingly, ‘the opportunity to interact with the public and raise their confidence in the police’ and ‘the opportunity to support victims’ were also often mentioned. A sample of comparison officers32 was also asked why they chose a career in policing. Their most common reason was ‘good career progression opportunities’ (44%; n=33); this reason was not identified as important for the Police Now officers. 31 Wards were selected based on analysis of levels of deprivation across London. Participants were typically placed in neighbourhood teams in areas with the highest rates of deprivation. Where possible, Police Now ensured that participants were posted in boroughs with a group of at least five other participants to develop a sense of bond across the cohort and maximise opportunities for collaborative working. 32 From the general MPS 2015/2016 intakes and a sample of DWOs (n=75); see methodology. 20 There was some variation in Police Now participants’ aspirations with regard to their future career as a police officer: 29% (n=20) stated they wanted to hold a detective role; 28% (n=19) a specialist role (e.g., firearms, protection, dog handler); and 25% (n=17) a senior leadership role (Superintendent or above). While the Police Now officers’ aspirations were varied, the majority of comparison officers aspired to hold a specialist role (51%; n=38). Interestingly, only 6% (n=4) of participants indicated it was their intention to move on to a career outside of policing at the end of the two year Police Now programme. Expectations of being a police officer in the MPS Participants’ expectations of the MPS as an employer were very high (in TS1). For example, 90% (n=62) expected the organisation to provide the training needed to do the job and 91% (n=63) expected the MPS to provide the necessary welfare and support. Furthermore, 91% (n=63) of participants expected to be treated fairly by the MPS, 87% (n=60) thought the organisation would be open and honest with them, 88% (n=61) expected the MPS to recognise their strengths and help them develop, and 87% (n=60) expected regular feedback on their performance. Almost three quarters (73%; n=50) of participants felt the MPS would encourage them to develop innovative solutions to problems; however, 25% (n=17) reported mixed views. Respondents were asked to provide a free text answer to the question, ‘What do you think are the main challenges that you will face as a new police officer in the MPS?’ The answers to this varied; however, a large proportion cited the challenges they would face working with communities with low confidence in the police and others remarked about the cultural resistance they might face from other officers as a result of entering the MPS on a graduate programme. Adapting to new situations, learning new policies and procedures, and applying the training were also commonly mentioned. When considering how being a police officer would impact on other areas of their lives, 99% (n=68) felt it would positively influence their confidence and assertiveness. Likewise, 99% (n=68) thought their communication skills would be positively impacted. Of note, 77% (n=53) felt that being a police officer would have a negative impact on their work-life balance and 78% (n=54) felt their social life would suffer as a result of being a police officer. Gaining the trust of groups of people in my ward who have perhaps not historically had confidence in or much interaction with the police [Police Now participant] 21 The Delivery of the Police Now Graduate Leadership Development Programme The next section explores the delivery of the first Police Now programme. Participants were asked to share their experiences of the Summer Academy training, the 28-day DWO immersion period and how prepared they felt to begin their roles as police officers. Syndicate leads also shared their experiences of the programme and suggestions for improvements and participants’ line managers reflected on how they thought the training had prepared their officers for the job. Delivery of the Summer Academy training Overall, participants were happy and satisfied with the level and quality of training they received at the Summer Academy. Eighty-nine percent of respondents to Tracker Survey 2 (TS2; n=40) reported being very or fairly satisfied with the quality of training received and all participants (n=45) agreed the Summer Academy was a positive learning environment. Overall, how satisfied were you with the quality of training you received during the Summer Academy? Very dissatisfied Neutral Very satisfied 0% 2% 9% 62% 27% 2% 89% Similarly, the interviews conducted with Police Now participants (n=22) demonstrated very positive views about the quality of the training delivered at the Summer Academy. Syndicate leads and the Police Now operations team were also positive about the training delivered at the Summer Academy. In particular, they commented that the training Police Now offers is far superior to the training they received as new recruits. Key Findings The first cohort of 67 Police Now participants included a relatively diverse group of individuals, with 45% female and 12% BAME. These proportions are larger than previous graduate entry programmes. Furthermore, the proportion of female and BAME participants increased in the 2016 cohort of the Police Now programme (54% female and 20% BAME) and overall BAME applications have increased by 34% since the first year of the programme. Participants were posted in some of the most disadvantaged communities across London, with the majority posted in wards within the top 20% of vulnerable localities. Participants had little prior experience of working with the police. The main reasons they chose to pursue a career as a police officer were because they thought it would be an exciting role and because of the opportunities offered by Police Now. Participants’ expectations of the MPS were high – they believed it would be an organisation that would treat them fairly and provide the necessary welfare and support. Participants were realistic about the challenges they expected to face as a new police officer in the MPS, including dealing with communities with low confidence in the police and the cultural resistance they might face from other officers. Participants felt becoming a police officer would have both positive (e.g., increased communication skills) and negative (e.g., work-life balance) impacts on their life. 22 What worked well? The syndicate structure (Figure 4) emerged as an important feature of the Summer Academy (see Figure 8 for a word cloud of the most commonly reported words). Participants thought the syndicate structure gave them the opportunity to bond with their fellow cohort members as well as creating a more relaxed and encouraging learning environment. Syndicate leads were cited as being instrumental to the success of the syndicate model, allowing participants to learn in a safe and supported environment. The positive learning environment was mentioned by many participants. They reported enjoying being surrounded by enthusiastic, interesting, and dynamic frontline officers with expertise in different areas. They also commented on the quality of the visiting fellows. Participants also reported enjoying the intensity of learning more quickly via the six-week Summer Academy and the field- training aspect; however, as discussed in the next section, many participants felt they could have benefited from more field-training prior to starting their roles as DWOs. Finally, participants were particularly positive about the DWO-specific training provided, such as soft skills and problem-solving. Figure 8. Words commonly used by Police Now participants when asked to describe what worked well about the Summer Academy Absolutely fantastic. I didn’t know what to expect beforehand but I previously received training from a graduate employer and I was amazed at how much better the Police Now training was [Police Now participant] Police Now training is inspiring, all the things that training experienced in my career isn’t. It is dynamic, creative and delivered by practitioners [Syndicate Lead] What made it exceptional was having officers who are currently serving being involved; it meant that the training was much more relevant [Police Now participant] 23 Consistent with participants’ perceptions of what worked well, interviews with the syndicate leads and operations team suggested the syndicate structure worked particularly well, along with the combination of trainers (e.g., syndicate leads and visiting fellows with expertise in different subject matters). The combination of practical learning and classroom based exercises was another aspect of the Summer Academy syndicate leads thought worked well. What could be improved? The six-week time period was considered too intense for some participants, especially in relation to: not having enough time to revise and review information and knowledge; long days causing tiredness and exhaustion; and the ‘rushed’ structure of training days. The majority of the cohort (77%; n=35) thought the Summer Academy was too short and 93% (n=42) thought the Summer Academy was exhausting. In addition, many participants thought the Summer Academy could have had a larger focus on the practical aspects of the job; one quarter of participants (n=10) were dissatisfied with the balance between classroom and practical learning. The Summer Academy was too short Strongly disagree Neutral Strongly agree 5% 7% 11% 46% 32% 12% 77% Another suggestion for improvement by the cohort was the field training period, including its timing, structure and organisation. Many participants thought the three week field training period was too short to adequately prepare them for the front line. They also thought the field training shifts could have been longer (i.e., full-day) and more varied (i.e., different shift patterns, including nights). In terms of the organisation of the field training, many participants reported their teams apparently knew nothing about the Police Now programme and were unaware they were supposed to be providing training for recruits. Furthermore, participants reported inconsistencies in the extent to which different team members were willing to help them with their development and training. Some participants also felt the field training could have had a wider emphasis on the processes and procedures of overall policing, rather than specifically focused on the DWO role. A number of participants reported Police Now trainers were proper street cops – they lived and breathed policing [Syndicate Lead] There was little or no time to consolidate knowledge or go over what I’d learnt. The hours were long; we started at 8am and finished at 7pm, plus travel time, so there was no time to go over anything in the evenings either [Police Now participant] I think there needs to be further field training. There simply wasn’t enough time to practice key policing skills [Police Now participant] I think the training was good in that it prepared us for the role that was advertised, however I think in reality I have spent a lot of time doing things I wasn’t trained for [Police Now participant] 24 they had been drawn into police work outside of the DWO role because of the heavy workloads of their team. Syndicate leads agreed with the cohort that for practicality purposes the days should not be split between classroom and field-training and that a full day should be spent out on borough. They also thought more time could be spent on certain exercises (e.g., role plays) in order to boost confidence. Some syndicate leads felt they did not have enough time themselves to prepare for the training they delivered, although recognised part of this problem might be exclusive to the first year rollout. Syndicate leads reported feeling exhausted towards the end of the Summer Academy, finding it difficult to keep morale up. On the back of this feedback (which was published in the Year 1 Report) a number of improvements were made to the 2016 Summer Academy, including altering the structure of the training to allow for full days on borough as opposed to splitting the days between classroom and field training, and introducing an additional seventh week to the programme. Additionally, the total number of field training hours in the 2016 Summer Academy was 20% higher than in the 2015 timetable. The 28-Day DWO Immersion Period Overall, most participants reported they learned a lot from the 28 day DWO immersion period; however, many reported this being a challenging time (see Figure 9 for a word cloud of the most commonly reported words). The quality and availability of their DWO mentor appeared to be a major factor in how participants found the immersion period, as well as the level of support offered by their team. Figure 9. Words commonly used by Police Now participants when asked how they found the 28-day DWO immersion period Review the 6 weeks duration, as syndicate leads were exhausted afterwards. Does there need to be more thinking and reflection time, and perhaps more recovery time? [Syndicate Lead] 25 Participants who were positive about the 28 day DWO immersion period reported: a high level of support from their team; having a mentor with identical shift patterns; having a mentor who was experienced and knowledgeable, and who was a fellow DWO. Participants who were less positive about the immersion period reported: having an absent or unavailable mentor or a mentor with little experience; having a mentor who was not interested in mentoring; the team having little knowledge of Police Now or what the immersion period was meant to achieve; a lack of motivation and low morale within the team; and a lack of planning by the borough. For the second year rollout of the programme, Police Now introduced Leadership Development Officers (LDOs) whose primary role was to coach and support the Police Now participants throughout the two-year programme. Furthermore, more effort was made in 2016 to engage with and motivate participants’ mentors and line managers. Confidence and preparedness Many participants reported not feeling prepared to start their roles as police officers. Less than half of the cohort (43%; n=19) thought the Summer Academy equipped them with the necessary skills to embark on their career as a police officer and only 36% (n=16) felt they had the necessary confidence. Confidence appeared to be a particular problem for female participants, with significantly fewer females (20%; n=4) reporting feeling confident compared to males (50%; n=12).33 The Summer Academy equipped me with the necessary skills to embark on my career as a police officer Strongly disagree Neutral Strongly agree 0% 18% 39% 29% 14% 18% 43% At the end of the Summer Academy I had the necessary confidence to start my career as a police officer 7% 25% 32% 27% 9% 32% 36% Developing the confidence and knowledge to deal with the varying aspects of being a DWO was a challenge identified by participants. Some put this down to the inevitable nature of starting a new job, while others felt unprepared and lacked knowledge about certain aspects of their role. A sample of comparison officers (n=13) who joined the MPS through the traditional entry route at the same time as the Police Now cohort were interviewed. The majority of these officers 33 According to an independent samples t-test, t(42) = 2.82, p = .007, 95% CI [.25, 1.47], d = .87 Brilliant. My team made me feel welcome and have provided me with support. Having a mentor with identical shift patterns really helped [Police Now participant] I didn’t really have an immersion period. I was not given one to one mentoring as my DWO ‘partner’ is frequently absent and had no interest in mentoring me [Police Now participant] 26 similarly did not feel confident starting their role as police officers and that their training left them feeling unprepared. Lack of confidence is clearly not unique for the Police Now participants, pointing to an important wider issue for police training. Just over half (54%; n=34) of Police Now participants’ line managers thought their officers were well prepared for the role.34 Most line managers thought their officers’ abilities were either better than other officers (47%; n=30) or just the same as other officers with similar levels of service (40%; n=25). Line managers who did not think the training academy had prepared their officers for the role thought the training period was too short to convey proper knowledge about being a DWO and thought there were discrepancies between the real-life DWO role and the one the programme promotes to officers. Furthermore, a lack of general policing experience was identified by line managers as a major challenge. Positively, the majority of line managers (19 of the 25 line managers interviewed) said their Police Now officers were meeting or exceeding their expectations. The minority of line managers who did not feel their officers were meeting their expectations stated they were anticipating someone with more experience. Overall positive feedback about the delivery of the Police Now programme Overall, participants were very satisfied with the Police Now programme. The overwhelming majority of participants agreed they: would encourage other graduates to join the programme (96%; n=43); talked positively about Police Now to other people (83%; n=37); felt a sense of loyalty to Police Now (83%; n=37); and felt a strong bond with their fellow Police Now colleagues (89%; n=40). Furthermore, the majority of participants agreed they understood the goals of Police Now (87%; n=39) and all (n=45) participants agreed the Police Now team are working hard to make the programme a success.35 These findings are in contrast to the MPS 2013 Graduate Entry Programme. Of these recruits, half (n=39) were dissatisfied with the quality of training received and only 59% (n=46) would encourage other graduates to join the programme. A lack of communication, poor information provision and an alleged failure to meet expectations were commonly cited concerns of the 2013 Graduate Entry cohort. 34 Based on interviews conducted with 25 line managers and a survey conducted by the Police Now leadership team with 38 line managers. 35 Based on survey data from Tracker Survey 2; there were some reductions over time but the majority continued to express positive views toward the Police Now scheme. A lot of the programme is about problem- solving and DWOs. The only thing for me was, as well as doing that side of the job they also have to learn how to be a police officer…they have to understand the basics of what doing our job is about [Police Now line manager] 27 Life in the Met Police The next section explores participants’ experiences of being a police officer in the MPS, including any stigma they encountered upon joining their team and their experiences of life on the job and the DWO role. Line managers and syndicate leads also shared their experiences of participants’ transition to frontline policing. Stigma and acceptance After a few months on the job, the majority of respondents to TS2 (73%; n=33) felt they had been accepted by colleagues on their borough and that they were treated with respect (71%; n=32).36 At the same time, 54% (n=26) of respondents to TS2 reported feeling a stigma attached to being part of the Police Now programme and 49% (n=22) reported feeling uncomfortable disclosing being part of the programme when meeting new colleagues. When participants were asked whether the way they are treated differs to officers not on the Police Now programme, only 5 of the 22 officers interviewed agreed. They reported that managers and superiors can have higher expectations of them and sometimes offer them more opportunities and responsibility than others. 36 This is in contrast to 47% of all MPS staff who think they are treated with fairness and respect in the Met (MPS Staff Survey, 2015) I think it’s because people aren’t fully aware of what it is; they think it’s a high potential scheme and we’ll be promoted in six months [Police Now participant] Key Findings The first cohort of Police Now participants were highly satisfied with the level and quality of training they received as part of the Summer Academy. The syndicate structure emerged as an important feature, along with the positive learning environment and the quality of visiting fellows. The majority of participants thought the Summer Academy was too short and both participants and syndicate leads thought the training could have had a larger focus on the practical aspects of the job and longer, more varied shifts on field training. Syndicate leads also thought more time could have been allocated for them to prepare for the Summer Academy. There was mixed feedback from the cohort about the 28-day DWO immersion period and the level of support offered by DWO mentors and line managers varied significantly. Some participants reported their DWO mentor was unavailable or disinterested in mentoring them, which negatively affected their ability to learn during this period. Many participants reported not feeling prepared to start their roles as police officers. However, confidence and preparedness also appeared to be an issue for recruits who joined via the traditional MPS entry route, suggesting confidence is not a problem unique to the Police Now officers but a wider issue for policing training. Feedback from participants’ line managers showed that almost half thought their officer was not sufficiently prepared for the role and many thought they could have benefitted from more training in wider policing skills. Despite this feedback, the majority of line managers thought their Police Now officers were either meeting or exceeding their expectations. Overall, satisfaction with the delivery of the Police Now programme was high and the overwhelming majority of participants would encourage other graduates to join the programme. 28 There is a stigma attached to being part of the Police Now programme Strongly disagree Neutral Strongly agree 0% 31% 13% 33% 22% 31% 55% When meeting new colleagues I feel comfortable disclosing that I am part of the Police Now programme 7% 20% 26% 33% 15% 27% 48% Interviews with Police Now participants indicated they thought a lack of awareness and incorrect assumptions about the Police Now programme were the causes of the stigma they had experienced, together with more general prejudices towards graduate programmes within policing. Those who had not felt stigma suggested either their team knew nothing about the Police Now programme, which meant they had been treated in the same way as other probationers, or their colleagues were more open-minded and accepting of them. Fewer Police Now participants reported feeling stigma compared to the 2013 Graduate Entry Programme cohort: 73% (n=57) of this cohort reported feeling stigma attached to being part of a graduate programme (see Figure 10). According to the Police Now leadership team, the syndicate structure was designed to increase resilience toward stigmatisation and enable participants to better cope with the pressures of their new role. Interviews with syndicate leads indicated they agreed there is often a burden of expectation and resentment toward those on graduate programmes like Police Now. Syndicate leads believed a lot of the stigma Police Now participants face will be as a result of misconceptions about the programme. Line managers described mixed reactions from within their teams to a Police Now recruit joining them. They described some initial scepticism due to: a lack of awareness of the difference between Police Now and other similar programmes; current and established DWOs either losing their role to Police Now officers or being required to share a previous standalone job; a perception of too many probationers being introduced into the MPS at the expense of more experienced staff; and doubt as to whether having a university degree would make a difference to someone’s policing skills. There’s a lot of expectation; a lot of people want the scheme to succeed and are watching us closely just in case it doesn’t [Police Now participant] Stigma is unavoidable, they will get labelled… If they excel and show people around them they are competent, then that will be it [Syndicate Lead] 29 Positively, many of the initial negative reactions from existing team members were quickly overcome once the Police Now officers arrived on their wards. Almost all line managers felt that their Police Now officer integrated well into their team, describing them as “popular” and “well- liked”. Furthermore, line managers’ experiences of line managing their Police Now officer were overwhelmingly positive, with the majority commenting on how impressed they were with their: enthusiasm, keenness to learn, proactive attitude, innovative thinking, intelligence and commitment. They variously described their experience as “thoroughly enjoyable”, “an amazing experience”, and “an absolute delight”. Although several line managers did not perceive any differences in the way Police Now officers are treated compared to others within the MPS, some felt that management has higher expectations of them, and that they can have a difficult time living up to these. Job satisfaction and motivation In TS2 administered after a few months on the job, the majority of participants reported being satisfied with their job as a DWO (78%; n=35) and being motivated at work 82% (n=37). Participants mentioned many positive aspects to their jobs, including: getting to know the area and its residents in depth; having the freedom and independence to be responsible for planning their own day; the varied and diverse nature of the role; being able to make positive long-term changes and improve people’s lives; and being able to see projects through to conclusion. I am satisfied with my current job Strongly disagree Neutral Strongly agree 0% 7% 16% 62% 16% 7% 78% Participants also spoke of the challenging aspects to their jobs, including: prioritising busy workloads; a lack of motivation and supervision from their team; the logistical frustrations of trying to solve problems and implement new strategies; complications caused by cumbersome MPS processes and systems; and managing public expectations. Almost one third of participants (n=14) felt that serving as a police officer was not as exciting as they anticipated and the same proportion (31%; n=14) answered neutral. In their roles, 82% of participants (n=37) felt they were able to positively influence confidence in the police and 72% (n=32) felt they were encouraged to develop innovative solutions to problems. However, only 53% of participants (n=24) felt they were able to influence local crime patterns in their role. The general feeling was ‘oh here we go; another of these schemes’. That was definitely the most general view, people thinking they’re going to be leap-frogged. I’ve had to point out to people that it isn’t a fast-track scheme [Police Now line manager] 30 Over the course of the two year Police Now programme—consistent with previous research into new police recruits—participants’ job satisfaction and motivation at work declined slightly (see Figure 11 and Appendix E for a table comparing responses across all four tracker surveys).37 Furthermore, over time, participants became less satisfied that their chosen career as a police officer was right for them (from 93%; n=42 in TS2 to 78%; n=18 in TS4).38 Figure 11. Participants’ ratings of job satisfaction and motivation at work declined over time Life on the job In TS2, participants were asked about different aspects of ‘life on the job’. Positively, the majority of participants: felt encouraged to use their discretion when making decisions (76%; n=34); felt comfortable expressing their opinion to people more senior than them (76%; n=34); and thought the amount of work they were expected to do was fair (76%; n=34). However, only 37% (n=17) of participants felt they were given the resources and equipment needed to do their job, fewer than half (46%; n=21) felt they were rewarded fairly for the work they do, and just over half (57%; n=26) felt there was a positive atmosphere among the people they work with. Positively, over the two year programme, more participants felt encouraged to use their discretion (from 76%; n=34 in TS2 to 82%; n=19 in TS4) and a similar proportion reported feeling comfortable expressing their opinion to senior colleagues (from 76%; n=34 in TS2 to 77%; n=18 in TS4). Furthermore, over time, more participants felt accepted by their colleagues on borough (from 74%; n=33 in TS2 to 86%; n=20 in TS4). However, over time, fewer participants felt the work they were expected to do was fair (from 76%; n=34 in TS2 to 50%; n=12 in TS4). In addition, fewer believed they were given the resources and equipment needed to do their job (from 37%; n=17 in TS2 to 18%; n=4 in TS4); fewer felt rewarded fairly for the work they do (from 46%; n=21 in TS2 to 18%; n=4 in TS4); 37 These results should be interpreted with caution given the small sample size. In addition, not all participants who responded to TS1 or TS2 responded to TS4, therefore the sample is slightly different at each time point. 38 Interestingly, significantly fewer female participants (70%; n=14) in TS2 reported job satisfaction compared to male participants (84%; n=21), t(43) = 2.43, p = .019, 95% CI [.10, 1.06], d = .74. 50 55 60 65 70 75 80 85 Tracker Survey 2 (n=45) Tracker Survey 3 (n=38) Tracker Survey 4 (n=23) % agree/strongly agree Job satisfaction Motivation 31 and fewer felt a positive atmosphere among the people they work with (from 57%; n=26 in TS2 to 50%; n=12 in TS4). Only half of participants (50%; n=23) in TS2 felt they were achieving a good work-life balance, which remained relatively consistent over time (46%; n=11 in TS4). These are not Police Now related issues per se but strike a wider concern within policing in general; however, understanding such findings and attempting to counter them should be one of the goals of Police Now in order to ensure the initial positive views of the cohort are maintained. Participants’ ratings of life in the Met police were reasonably consistent with other MPS staff with similar lengths of service, again suggesting this is a wider issue for the MPS.39 The DWO role In terms of the DWO role, interviews with syndicate leads indicated they thought the role was the right choice for the Police Now programme. They believed the role gives the cohort the opportunity to make a difference and allows them to see any impact they are having directly on their community. They also thought the role gives the cohort a good grounding in the skills needed to be a leader. There were mixed views amongst the cohort. Some participants were concerned they are not able to experience ‘wider’ policing while working as a DWO, which they felt was negatively impacting on their overall development as a police officer. Other participants were positive about the DWO role. Some mentioned the greater levels of responsibility that the DWO role brings, along with the ability to work on a more varied range of tasks and to carry out more community engagement. Most line managers interviewed were positive about the DWO role; however, many thought participants should have experience in the wider policing role as well. 39 MPS Staff Survey (2015) Gives leadership and challenge that DWO role bring instead of getting 'lost in the numbers'. Dedicated to that ward gives you accountability, and you set your own goals [Syndicate Lead] I feel like I’m not learning as much as I should; I feel like after the two years I won’t know as much as other police officers because I won’t have experienced these confrontational situations and the DWO role is much more slower paced [Police Now participant] 32 Participants’ Attitudes Toward the MPS and Serving the Public This section explores Police Now participants’ attitudes and how these change over time, including their attitudes toward the MPS as an organisation and the leadership and support provided to them. Participants’ orientation and views toward procedural justice and serving the public are also explored, and comparisons are made to police officers not on the programme. Attitudes toward the MPS After a few months on the job, participants were asked for their thoughts on the MPS. Participants were mostly positive: 76% (n=34) agreed the MPS is a good organisation to work for; 80% (n=36) felt a sense of loyalty to the MPS; 93% (n=42) stated they were proud to work for the MPS; and 65% (n=29) felt clear on the organisational goals of the MPS. However, participants’ attitudes toward the MPS declined over time. For example, fewer participants believed the MPS was a good organisation to work for (from 75%; n=34 in TS2 to 52%; n=12 in TS4) and fewer were clear on its organisation goals (from 65%; n=29 in TS2 to 48%; n=11 in TS4). Key Findings After a few months on the job, most participants felt accepted by their colleagues on borough; however, around half of the cohort reported a stigma associated with being part of the Police Now programme and a similar proportion felt uncomfortable disclosing they are part of the programme when meeting new colleagues. Participants and syndicate leads thought part of this was attributable to a lack of awareness and misconceptions about the programme. The Police Now cohort reported feeling less stigma about being attached to a graduate programme than officers in the MPS 2013 Graduate Entry Programme cohort, which could be due to the greater level of support offered to the cohort. It is also possible this stigma will reduce in later cohorts once Police Now has had time to embed within the organisation. After a few months on the job, participants’ satisfaction with their role as a DWO was high and the majority reported being motivated at work. However, consistent with previous research into new police recruits, the Police Now officers’ job satisfaction and motivation at work declined over time. There were mixed feelings about different aspects of ‘life on the job’. Positively, the majority of participants felt encouraged to use their discretion when making decisions and felt comfortable expressing their opinion to senior colleagues. Of concern, the majority of participants felt they were not given the resources needed to do their job and approximately half of participants felt they were not achieving a good work-life balance. These areas deteriorated further over time; for example, in the final survey only 18% of participants felt rewarded fairly for their work. These are not Police Now related issues per se but strike a wider concern within policing in general. Participants’ ratings were also reasonably consistent with other MPS staff with similar lengths of service. In their roles as DWOs, the cohort felt they were most able to influence confidence in the police and least able to influence local crime patterns. Most people interviewed thought the DWO role was the right choice for the Police Now scheme; however, more experience in wider policing duties was thought necessary to assist participants in their overall development as a police officer. 33 The MPS is a good organisation to work for Strongly disagree Neutral Strongly agree 0% 9% 15% 63% 13% 9% 76% In the induction survey administered during the Summer Academy (TS1), the majority of participants were confident they would receive a good service if they contacted the MPS as a member of the public (75%; n=52). However, in Tracker Survey 4 (TS4), only 35% (n=8) reported feeling confident. This is a sizable drop off but still higher than MPS staff overall (28% of MPS staff in the 2015 survey thought they would receive a good service). In TS1, 65% of participants (n=45) saw themselves as spending their whole career in the MPS. Over time, slightly fewer participants thought they would spend their whole career in the MPS (52%; n=12 in TS4). A smaller proportion of Police Now officers intended to spend their whole career in the MPS compared to the comparison group (76%; n=57). Participants’ confidence in the leadership provided by senior leaders in the MPS was positive early on in their careers (74% in TS2; n=33), but dramatically declined over time (to 36%; n=8 in TS4). Furthermore, over time, more participants believed senior managers were out of touch with what is happening on the front line (from 24%; n=11 in TS2 to 46%; n=11 in TS4). These findings are completely in line with findings from the 2015 MPS Staff Survey. Over time, fewer MPS staff were confident in the senior leadership (63% for staff with less than 6 months service vs. 30% for staff with 1-2 years service). The cohort’s confidence in the leadership provided by their local senior leadership team also declined (from 65%; n=29 in TS2 to 55%; n=13 in TS4) and, over time, fewer participants believed the values of the MPS were reflected in their local senior leadership team (from 67%; n=30 in TS2 to 55%; n=13 in TS4). Again, these declines are consistent with findings from the MPS Staff Survey. In terms of their professional development, over time, fewer participants reported being satisfied with the level of support they had received from the MPS (from 67%; n=30 in TS2 to 55%; n=13 in TS4). Positively, satisfaction with the level of leadership and support they received from their line manager increased over the course of the programme (from 74%; n=33 in TS2 to 91%; n=21 in TS4). Furthermore, over time, more participants agreed their line manager gave them regular and constructive feedback on their performance (from 44%; n=20 in TS2 to 68%; n=16 in TS4) and more felt respected by their line manager (from 76%; n=34 in TS2 to 91%; n=21 in TS4). Positive attitudes toward serving the public Encouragingly, findings from the induction survey (TS1) indicated that participants had a very strong orientation toward the ideas associated with neighbourhood policing and procedural justice. Illustrating this, in the first tracker survey, 96% (n=66) agreed that neighbourhood policing was ‘real’ policing and 91% (n=63) recognised that ‘customer service’ is central to a police officer’s job. 34 Being concerned about customer service should be a big part of a police officer’s job Strongly disagree Neutral Strongly agree 4% 3% 1% 27% 64% 7% 91% Furthermore, 94% (n=65) agreed officers should treat everyone with the same level of respect, regardless of how they behave toward the police, and 88% (n=61) agreed it is important the MPS workforce ‘looks and feels’ like the diverse communities it serves. When participants were asked what factors motivate them in their role, the knowledge that they are improving people’s lives and ‘making a difference’, the varied and interesting nature of the DWO role, the dedication of their colleagues, and the daily interaction with diverse people were most often mentioned. Attitudes relating to procedural justice and serving the public saw small declines over time. For example, fewer participants agreed that customer service should be a big part of a police officer’s job (from 91%; n=63 in TS1 to 73%; n=50 in TS4); fewer agreed officers should treat everyone with the same level of respect, regardless of how they behave toward the police (from 94%; n=65 in TS1 to 77%; n=18 in TS4); fewer agreed it is important the MPS workforce ‘looks and feels’ like the diverse communities that it serves (from 88%; n=61 in TS1 to 64%; n=15 in TS4); fewer agreed that good police-community relations have an effect on the crime rate (from 94%; n=65 in TS1 to 77%; n=18 in TS4) and fewer agreed that all victims of crime are deserving of the same good service (from 88%; n=61 in TS1 to 77%; n=18 in TS4). Participants’ attitudes toward neighbourhood policing remained the same over time: the majority recognised that neighbourhood policing is ‘real’ policing (96%; n=66 in TS1 and 96%; n=22 in TS4). Positively, and as we might expect, more participants agreed they understood the needs of the community they serve (from 78%; n=35 in TS2 to 91%; n=21 in TS4). There were some differences between the Police Now and comparison officers in their attitudes toward serving the public, with the Police Now officers holding more favourable views. Police Now participants were more likely than comparison officers to agree that officers should treat everyone with the same level of respect regardless of how they behave toward the police (74%; n=28 in TS3 vs. 63%; n=39). Furthermore, Police Now officers were more likely to agree that all victims of crime are deserving of a good service (92%; n=35 vs. 84%; n=52) and less likely to agree that there are certain communities who do little to deserve the respect to the police (3%; n=1 vs. 16%; n=5). Police Now participants were also more likely to agree they understood the needs of the community they serve (95%; n=36 vs. 79%; n=26). Other attitudes were comparable between the two groups; for example, a similar proportion of Police Now (76%; n=29) and comparison officers (76%; n=46) agreed it is important the MPS workforce ‘looks and feels’ like the diverse communities it serves. 35 Evidence Based Policing and Police Now In all of the conversations around police transformation, one of the strongest threads is the growth of EBP. This is not a new concept,40 yet it is only relatively recently that the term has become part of the lexicon of policing. There has been a huge amount written about EBP, be it definition discussions, what is good enough ‘evidence’ for policing (e.g., randomised controlled trials versus other evidence), police culture and receptivity to evidence, and how to embed EBP into practice.41 Within Police Now, officers receive training around EBP and are expected to incorporate this routinely within their craft. This section of the report covers the officers’ perceptions toward EBP and how well they have embedded these skills within their role (i.e., as demonstrated via their 100 day impact assessments). 40 Sherman (1998); Weiss & Bucuvalas (1980). 41 See for example Lum et al. (2012) or Stanko & Dawson (2016) Key Findings Early on in their careers as police officers, participants’ attitudes toward the MPS as an organisation were mostly positive. They were proud to work for the MPS and felt it was a good organisation to work for. However, over time, consistent with the wider MPS workforce, these attitudes declined. For example, after almost two years working for the MPS, only around a third of participants were confident of receiving a good service if they contacted the MPS as a member of the public and fewer than half were clear on what the organisational goals of the MPS are. Although there were sizeable reductions, the Police Now cohort still expressed more positive views toward the MPS than all MPS staff combined. Police Now participants were less likely to want to spend their whole career in the MPS compared to the comparison officers surveyed. This is consistent with the recruitment model of Police Now; participants are encouraged to use the programme as a platform for developing their leadership career. After the two year programme, participants can choose to pursue a career outside policing. Indeed, two-thirds of the 2015 cohort chose to remain in policing at the end of the two-year programme, with the remainder leaving to pursue a career in other sectors. Similar to participants’ attitudes toward the MPS as an organisation, participants’ confidence in senior leaders of the MPS started off high but deteriorated over time. Towards the end of the two year programme, only a minority of participants (36%) were confident in the leadership provided. Again, these declines are entirely consistent with other MPS staff with similar service lengths. For MPS staff with 1-2 years’ service, only 30% reported confidence with the organisation’s senior leadership. Positively, the level of leadership and support received from participants’ line managers was rated more positively over time. Encouragingly, Police Now participants demonstrated a strong orientation toward the principles of procedural justice and neighbourhood policing. Nearly all participants agreed that customer service was an important part of a police officer’s role and that all members of the public should be treated with the same level of respect. Although some of these attitudes declined over time, the majority still expressed favourable views toward serving the public. A comparison of the Police Now cohort with other police officers showed that Police Now participants held more favourable views toward procedural justice and serving the public. For example, compared to other police officers, Police Now participants were more likely to agree that officers should treat everyone with the same level of respect regardless of how they behave toward the police. The Police Now cohort was also more likely to agree they understood the needs of the community they serve. 36 Understanding evidence based policing Positively, the majority of participants (72%; n=27) agreed they understood the principles of EBP. This is a similar proportion to previous iterations of a similar survey conducted with senior MPS officers (Superintendent rank and above) in 2012.42 In defining evidence based policing, the most common words mentioned by the cohort were ‘data’, ‘research’, ‘information’, ‘decisions’ and ‘outcomes’ (see Figure 12). Figure 12. Words commonly used by Police Now participants when asked to define the term ‘evidence based policing’. However, findings were more mixed around whether they thought their training taught them how to: locate different kinds of evidence (42%; n=16); apply evidence to a project (42%; n=16); or assess the impact of a project (36%; n=14). Only one quarter (n=10) of participants thought their training taught them how to integrate different kinds of evidence and only 28% (n=11) thought their training taught them how to critically appraise evidence. Participants were asked how willing they would be to take a particular action to test whether a tactic the police are currently using is effective (see Table 1). Participants were most willing to stop a tactic in one small area and compare what happens in another area (72% very or quite willing; n=27) and least willing to stop a tactic on the basis that a researcher told them there was research showing it was ineffective (43% very or quite willing; n=16). Twenty percent (n=8) would not stop a tactic on the basis that a researcher told them there was research showing it was ineffective. This is a similar proportion to the previous iteration of the survey conducted with senior MPS officers: 18% of these officers would not stop a tactic if research showed it was ineffective. When comparing these responses to the comparison officers, the Police Now cohort was more in favour of randomised control trials than the comparison officers (50% vs. 14% very or quite 42 Stanko & Dawson (2016) 37 willing), and slightly more willing to stop a tactic on the basis that a researcher told them research showed it to be ineffective (43% vs. 38% very or quite willing). Table 1. How willing would you be to take the following actions to test whether a particular tactic the police are using is effective? Very/Quite Willing Somewhat Willing Not Willing Police Now Comparison Police Now Comparison Police Now Comparison Stop a tactic to see if problem gets worse 44% 43% 33% 29% 22% 29% Stop a tactic in one small area and compare what happens in another area 72% 67% 22% 33% 6% 0% Find the top 20 areas where problem exists and toss a coin to assign 10 areas to have a tactic and 10 areas not to and compare 50% 14% 28% 38% 22% 48% Stop a tactic on the basis that a researcher told you there was research showing it was ineffective 43% 38% 37% 48% 20% 14% Use of scientific research versus professional expertise Participants were asked the question, “In day to day decision making, what do you think the balance should be between the use of scientific research/knowledge and professional expertise?”. The majority of respondents (64%; n=24) thought scientific knowledge and professional expertise should make an equal contribution to day-to-day decision making. Just over a quarter (28%; n=11) thought experience should be more important but scientific knowledge should make some contribution. Very few participants (6%; n=2) thought scientific knowledge should make more of a contribution than experience. In contrast to the Police Now cohort, the majority of comparison officers surveyed thought professional expertise should make more of a contribution than scientific research (61%; n=13). Overall, the findings so far present a relatively positive narrative in terms of participants’ understanding of and receptivity toward EBP. The most important aspect, however, is whether they have practically embedded such working within their role. A lack of robust evidence within the 100 day impact assessments The 100 day impact assessments are an opportunity for the cohort to describe innovations they had made over the preceding 100 days to decrease crime or increase public confidence in their ward. Participants are encouraged to focus on measurable impact. The assessments are a compulsory requirement and all officers have to generate a product showing they have identified and tackled a problem and present evidence of success (or not). Outside of Police Now, as far as we know, no other neighbourhood police officers are held to account in this way. 38 To explore whether participants are embedding evidence based thinking within their work, a random selection of 70 assessment presentations (65 posters and 5 verbal presentations) were assessed. A set of criteria was developed that covered many essential practices within EBP (i.e., is there evidence of a literature review, is there a clear rationale for their project, are there clear outcomes, along with issues around methodology, accurate interpretation and inclusion of caveats; see Appendix D for the coding criteria). The coding framework was designed to cover all stages of EBP ranging from problem identification through to integrity in reporting findings. This should be caveated in that such evidence can only be captured if the posters and presentations include the information. For example, it is possible that the officers are using evidence from the literature but not including evidence of this in their poster or presentation; the coder can only rate what is documented in the presentations. Given the relatively small number of assessments, and the other caveats outlined earlier, these findings should be seen as indicative only; however, they still provide important insights into the practical use of EBP by the officers. Overall, there was little indication of effective evidence based workings documented within the impact assessments (see Table 2). To illustrate, only 6 officers (8%) showed evidence of literature (of any kind) within their poster or presentation. Only a handful of participants defined success for their project clearly (5%; n=4). In the majority of assessments (87%; n=61), it was not clear whether data to measure outcomes was captured. On a more stringent note, in the majority of cases (99%; n=72) it was not stated whether the initiative was actually delivered as expected. One third (37%; n=26) of assessments actively sought to empirically demonstrate success, even though they were all seeking to make statements about their initiative’s impact. Many participants used anecdotal data or activities to illustrate impact (i.e., informal conversations with residents; see quotes below). Where impact was examined, this was predominantly through a before and after methodology (n=18; i.e., no comparison group) - a weak design academically and one that does not enable confident statements around impact. Only four impact assessments used any type of comparison group, comprising crime data from other wards and borough or Met-wide statistics (i.e., there was no statistically matched or robust comparison groups). These types of designs can provide valuable insights and learning if reported and interpreted well; however, only 7 cases (10%) from the total number included any research caveats or limitations to their projects (e.g., small sample size, lack of comparison group, short analytical timeframes, data quality), despite most projects having numerous caveats. No officers conducted any statistical analysis to measure impact. The response that we received from students and staff was very positive, and this is something that will be built on in the future [Police Now participant impact poster] Visits afterwards confirmed that neighbours were happy that their issues were being taken seriously [Police Now participant impact poster] 39 Related to this, when asked about EBP, some officers honestly noted they have not used evidence to inform their operational decision-making. Participants also spoke of the challenges they have faced implementing EBP in their ward. These findings paint a rather less positive picture of the cohort’s incorporation of EBP within their work and how they perceive and seek to evidence impact. Some may say it is unfair to expect police officers to be learning from the literature, clearly defining project success, robustly measuring outcomes, understanding the limitations or ensuring accurate conclusions. However, these are essential aspects of EBP and Police Now should continue to support and encourage their officers to embed more evidence based working within their assessments. Indeed, Police Now has improved their training and support for EBP throughout the lifetime of the evaluation, including developing EBP masterclass days and developing more guidance for participants to evidence the impact of their work. Future independent evaluations should seek to measure the success of these developments. Table 2. Aspects of evidence based policing found in participants’ 100 day impact assessments Aspect of EBP Extent to which aspect of EBP shown Yes Partly/to some extent No Not clear Literature review carried out 1 (1%) 5 (7%) 63 (90%) 1 (1%) Lessons incorporated into project design 1 (1%) 1 (1%) 67 (96%) 1 (1%) Clear problem verified by data 2 (3%) 25 (36%) 12 (17%) 31 (44%) Project success clearly defined 1 (1%) 3 (4%) 62 (89%) 4 (6%) Project delivered as expected 0 (0%) 1 (1%) 0 (0%) 69 (99%) Data captured throughout 0 (0%) 2 (3%) 7 (10%) 61 (87%) Outcomes clearly measured 0 (0%) 26 (37%) 37 (53%) 7 (10%) Caveats outlined 1 (1%) 6 (9%) 58 (83%) 5 (7%) There has not been significant scope for evidence-based policing to challenge previous ways of policing the ward… [Police Now participant] It’s very difficult to get any borough level support for EBP, so largely EBP rests on stats that are easily available to me…I would like to be able to do more EBP but it’s very difficult currently to realistically integrate. [Police Now participant] 40 Police Now’s Impact on Public Confidence and Perceptions of the Police and Crime Finally, an objective measure of Police Now’s impact was assessed by examining ward-level public confidence and perceptions of the police and crime rates. Police Now’s impact on public confidence and perceptions was explored using the MOPAC Public Attitude Survey (see Methodology for a description of the methods used to select the comparison group). This analysis incorporated 2,732 respondents from Police Now wards and 4,218 respondents within statistically matched comparison wards. Survey data was weighted to adjust for under- representation of the population from which the sample was drawn. Police Now’s impact on crime was assessed using ward-level crime data (total notifiable offences plus specific crime types). The year (12 months) prior to the introduction of the Police Now programme (October 2014- September 2015) was compared to the year (12 months) subsequent (October 2015 – September 2016). Public confidence and perceptions of policing Overall, across the total number of respondents to the PAS, there was no compelling evidence of improvements within public confidence and perceptions for the Police Now wards (see Appendix F for a table comparing public perceptions for the Police Now and comparison wards and Figure 13 for change in the main measures). There was a 2% increase in ‘how good a job’ people thought the police were doing (i.e., confidence) in Police Now wards compared to a significant 3% increase in comparison wards. Across all questions explored (’how good a job’ and a number of key drivers of public confidence and other questions related to the public’s perceptions of Key Findings Police Now participants appeared to have a good understanding of the principles of EBP. However, many participants thought their training did not effectively teach them how to use and apply EBP in their work and a number of participants spoke of the challenges they have faced implementing EBP in their ward. Compared to police officers who joined the force through other routes, the Police Now cohort seemed to be more receptive to using scientific research in their day-to-day decision making. Although the Police Now 100 day impact assessments represent an innovative process through which the officers seek to evidence their success, an inspection of the Police Now Impact Library showed little indication of evidence based workings. Only a minority of participants: incorporated previous learning into the design of their initiative; clearly defined the problem through the use of data; attempted to measure outcomes or reported caveats or limitations within the work. Those who did seek to measure outcomes used weak methodological approaches. When speaking to the officers, some honestly noted they have not used evidence to inform their work and a number spoke of the challenges they have faced implementing EBP. These findings have implications for Police Now in ensuring participants have the knowledge and training to apply EBP in their work, but also for the police organisation as a whole in supporting and encouraging their officers to practically carry out EBP. Without this organisational support, training officers in EBP will in all likelihood not lead to tangible results. Police Now has improved their training and support for EBP throughout the lifetime of the evaluation, including developing EBP masterclass days and developing more guidance for participants to evidence the impact of their work. Future independent evaluations should seek to measure the success of these developments. 41 neighbourhood policing), on average, the Police Now wards showed no change in public perceptions in the year subsequent to the programme (an average change of -0.2% across the 12 questions explored).43 Furthermore, generally, the trends within the comparison and Police Now wards paralleled one another. On average, comparison wards showed a slight improvement in public perceptions across all questions explored (an average change of 2.3%). Across the total number of respondents, two questions showed significant improvements: more respondents in Police Now wards thought the police were ‘dealing with things that matter’ (from 69% to 73%) and more reported not being worried about crime in their area (from 58% to 62%). However, these differences were small and replicated across the comparison wards. There were other questions which showed significant declines in the year subsequent to Police Now (see Appendix F). For example, significantly fewer respondents reported knowing how to contact their local policing team (from 48% to 37%); this question also fell for the comparison wards (a 3% decline in comparison wards compared to an 8% decline in Police Now wards). Figure 13. Change in public perceptions for the Police Now and comparison wards *Change is significant at the p<.05 level Impact on public confidence and perceptions of the police was also explored by age group (see Figure 14 and Appendix F); however, the small number of young respondents means these specific results should be interpreted with caution and precludes firm ‘cause and effect’ statements on impact.44 We can see that, for the most part, across Police Now and comparison wards young people (16-24 year olds) showed more improvements than older people (25+ year olds). There were some differences between Police Now and comparison wards. There were larger improvements in ‘how good a job’ young people thought the police were doing (i.e., confidence) in Police Now wards (a significant 15% increase vs. a non-significant 4% increase in comparison 43 This figure is the average amount of change across the 12 questions explored and is presented here purely for descriptive purposes. 44 There were 206 16-24 year old respondents in Police Now wards in the year prior to the scheme and 202 in the year after. There were 276 16-24 year old respondents in comparison wards in the year prior to the scheme and 237 in the year after. 2% -8%* 0% 2% 3% 4%* 2% 3%* -3%* 0% 4%* 6%* 5%* 8%* -10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% Good job Contact local policing team Informed Relied on Treat fairly Dealing with things that matter Listen to concerns Police Now Comparison 42 wards).45 Conversely, there were larger improvements in comparison wards in the proportion of young people who thought the police were ‘dealing with the things that matter’ (a significant 11% increase vs. a non-significant 6% increase in Police Now wards), and that the police ‘listened to the concerns of local people‘(a significant 11% increase in comparison wards vs. a non-significant 6% increase in Police Now wards). Figure 14. Change in public perceptions for young respondents (16-24 year olds) versus older respondents (25+ year olds) for the Police Now and comparison wards *Change is significant at the p<.05 level Crime Police Now’s impact on crime was also explored. In the year prior to the Police Now programme (October 2014-September 2015), there was a significant difference between the Police Now and comparison wards, with a greater number of offences occurring in the Police Now wards. This pattern was consistent for total notifiable offences (see Figure 15) and across all specific crime types, including: total sexual offences; personal property offences; total theft offences; total motor vehicle offences; total criminal damage offences; serious youth violence; non domestic abuse violence with injury; knife crime offences and knife crime with injury offences. The pre- existing differences in crime rates between the two groups of wards potentially indicates a limitation to the matching procedure which was based on vulnerable locality score. There was a slight but non-significant increase in crime in the year subsequent to the Police Now programme (October 2015-September 2016) for both the Police Now and comparison wards. For example, for the Police Now wards, the average monthly total notifiable offences rose from 182 to 191; in the comparison wards, total notifiable offences rose from 119 to 127. There was no significant difference between the two sets of wards in crime trends over time. Therefore, 45 In their Impact Report, Police Now documented a 17% increase in the same question for young people in Police Now wards compared to a 3% increase in comparison wards. This analysis compared the 12 months from October 2014 to the 15 months from October 2015 which explains the slightly different figures here. 15%* -11%* 10%* 6% 6% 4% -1% 10%* 11%* 11%* -1% -10%* 1% 3% 1% 3%* -4%* 5%* 3%* 7%* -15% -10% -5% 0% 5% 10% 15% 20% Good job Contact local policing team Treat fairly Dealing with things that matter Listen to concerns Police Now 16-24 yrs Comparison 16-24 yrs Police Now 25+ yrs Comparison 25+ yrs 43 although the Police Now wards had a higher volume of crime overall, the two sets of wards showed the same trend in how crime volume changed over time. Figure 15. Average monthly total notifiable offences for the year prior (Oct 2014-Aug 2015) and subsequent (Oct 2015-Aug 2016) to Police Now. 80 100 120 140 160 180 200 2014-2015 2015-2016 Total Notifiable Offences Police Now Comparison Key Findings Across the total number of respondents to the PAS, there was no compelling change in public confidence or perceptions of the police for the Police Now wards in the first year of the programme. Generally, the trends within Police Now and comparison wards paralleled one another. There were larger improvement in how good a job young people thought the police were doing (i.e., confidence) in Police Now wards (a significant 15% increase vs. a non-significant 4% increase in comparison wards). However, there were other drivers of public confidence that showed the opposite pattern; for example, there was a significant 11% increase in comparison wards in the proportion of young people who thought the police were dealing with the things that matter compared to a non-significant 6% increase in Police Now wards. The small number of young respondents means these specific results should be interpreted with caution and precludes firm ‘cause and effect’ statements on impact. There was no evidence of impact when exploring ward level crime data. Both sets of wards showed comparable increases in crime subsequent to the introduction of the Police Now. The Police Now wards had higher crime rates in the year prior to Police Now, which may reflect limitations to the matching procedure. As a further point, the crime data we used does not capture the large number of crimes that go unreported. Future evaluations of Police Now could incorporate unofficial crime data, including information from public surveys (e.g., the Crime Survey of England and Wales) and data from the voluntary sector. These findings do not indicate that Police Now does not work or cannot eventually reach such outcomes. As Police Now grows, independent analysis should continue to explore these measures and the longer-term impact of Police Now on public confidence and crime. 44 4. Discussion This report presents findings from a large-scale two-year evaluation of the first cohort through the Police Now Graduate Leadership Development programme. The evaluation sought to explore the programme in detail, including the perceptions of the officers, experiences and delivery of the programme, evidence based working and potential impact. Sitting under these main themes is learning that not only applies to Police Now but also to the wider police organisation, culture and training moving forward. Delivery of the First Police Now Graduate Leadership Development Programme The experiences of the first Police Now cohort showed that the Summer Academy training was well-received. The enthusiasm and energy of the trainers, the quality of the material presented, and the strong bond created among participants were common themes throughout the surveys and interviews. Observations of the Summer Academy revealed a stimulating and energetic environment where participants, syndicate leads and those delivering the training were fully engaged in the sessions. Despite this positive feedback, there were a number of suggestions for improvement to the Summer Academy. Encouragingly, many of these suggestions (which were initially published in the Year 1 interim report - April 2016) have been incorporated by Police Now into the second and third year roll-outs of the programme. For example, although the intensity of the Summer Academy was praised by participants and syndicate leads, the length and timing of the training was also mentioned as an area for improvement. Specifically, participants thought that more time could be spent on field training and that the Summer Academy could be lengthened to make time for this training. As a result of this feedback, the structure of the 2016 Summer Academy was altered to allow for full days on borough as opposed to splitting the days between classroom and field training, and introduced an additional seventh week to the programme. Additionally, the total number of field training hours in the 2016 Summer Academy was 20% higher than in the 2015 timetable. Interviews with syndicate leads and comments from participants also suggested a review of certain topics could be undertaken. For example, syndicate leads suggested more time could be spent on role-play exercises to boost confidence. Participants also reported the training could have spent more time on practical skills/role-plays and less time on other areas, such as soft skills. As a result of this feedback, the 2016 Summer Academy increased the number of role-plays and other practical, hands-on learning exercises. The Police Now team should continue to evaluate future rollouts of the programme and continue to refine and improve the Summer Academy training. Transition to Front-line Policing The 28-day DWO immersion period was designed to ease the transition from the Summer Academy training to front-line policing. However, participants’ experience of this period was inconsistent and the level of support offered by DWO mentors and line managers varied 45 significantly. Some officers were positive about the immersion period and felt it was a great learning opportunity, whereas others reported it being an extremely challenging time. The main criticism of the immersion period was the lack of interest and availability of some of the cohort’s DWO mentors. Despite Police Now’s efforts to engage line managers, a number were unavailable for their Police Now officer or were disinterested in mentoring them. Ensuring consistency in the cohort’s transition to front-line policing is vital for the success of the programme and for the participants’ development as police officers. Police Now has since introduced Leadership Development Officers (LDOs) for the first and subsequent cohorts whose primary role is to coach and support the Police Now participants throughout the two-year programme. Furthermore, more effort has been made to engage with and motivate participants’ mentors and line managers. The success of these measures should be explored in future evaluations of the programme. Almost half of participants’ line managers thought their officer was not sufficiently prepared for the role and that they could have benefited from more training in wider policing skills prior to starting their roles as DWOs. Many participants reported having to get involved in wider policing duties because of heavy workloads within their team, despite it not being part of the DWO role. This finding suggests the need to ensure the officers are adequately trained to respond to the varied tasks they may be expected to do as part of a team. Although survey findings showed that many participants felt unprepared to begin their roles as police officers and lacked confidence (particularly female officers), a comparison with other police officers who joined the MPS via the traditional entry route showed that they too felt unprepared when they started out as new recruits. These findings suggest confidence is not unique for the Police Now participants but a wider issue for police training. Research should be conducted to explore why confidence was lower for female participants. Experience of the MPS as an Organisation Participants’ experiences of the MPS as an organisation revealed a picture consistent with previous research into new recruits. In a relatively short period of time, we observed a decline in attitudes toward the MPS and its senior leadership. Of particular concern, after less than two years working for the MPS, only a third of participants were confident of receiving a good service if they contacted the MPS as a member of the public compared to 75% at the start of the programme. This decline is not unique to Police Now participants and is consistent with findings from the MPS Staff Survey where satisfaction with the organisation generally decreases as service length increases. There is clearly a challenge here not only in recruiting exceptional graduates but sustaining their motivation and commitment to the organisation. Much has been written around policing morale in recent years, particularly given the climate of budget cuts and austerity and the various challenges facing the police organisation. The decline in attitudes toward the MPS shown by the Police Now officers—and other new officers within the MPS—may well be indicative of a wider attitude malaise within the organisation. Either way, if not addressed, there is a risk these new officers will become further disengaged from the organisation with potentially damaging implications. 46 Positive Attitudes Toward Serving the Public Encouragingly, the Police Now participants demonstrated a strong orientation toward the principles of procedural justice and neighbourhood policing. Although some of these attitudes declined over time—something that should be explored in more detail going forward, and possibly reflective of exposure to the wider policing culture—the majority still expressed favourable views toward serving the public. Furthermore, a comparison of the Police Now officers with other police officers in the MPS showed the Police Now cohort held more favourable views toward procedural justice and neighbourhood policing. It was anticipated that programmes such as Police Now—through recruiting graduates from ‘a different audience’ who might not have considered a career in policing before—might contribute toward diluting traditional understandings of police culture and promote support for more service based policing philosophies. 46 Although this was not measured in the current evaluation, it remains a solid aspiration. The Police Now cohort’s orientation toward service based policing suggests they have the potential to transform the current culture of the police service, but only if the organisation fosters and sustains their motivation. This could be supported by recruiting more Police Now officers, but such change will take time. Engagement with Evidence Based Policing Practices The Police Now cohort had a good understanding of evidence based policing (EBP) and their attitude towards it was favourable; however, many participants thought their training did not effectively teach them how to use and apply evidence in their work. Furthermore, a number of participants spoke of the challenges they have faced implementing evidence based policing in their ward, with some citing the difficulties in gaining support from their borough to use EBP. This finding is concerning but further signals the difficult occupational terrain participants have to navigate upon joining the force. The Police Now 100 day impact assessments are a process through which participants feed back about the work they have done to reduce crime and increase confidence in the police in their ward. As far as we are aware, there is no other comparison in UK policing currently where neighbourhood officers are required to showcase their work in this way. The impact assessments can therefore be seen as an attempt by Police Now to ‘raise the bar’ through which officers seek to evidence their impact and share innovation. The Police Now Impact Events that the authors went to were well-attended and engaging events and clearly showed the enthusiasm and passion of the Police Now cohort and those involved in running the organisation. One of the most important outcomes of this work, however, should be the extent to which the Police Now cohort is implementing (effective) EBP in their wards. An inspection of participants’ impact assessments showed little indication of effective evidence based workings. The analysis covered the entire journey around EBP from drawing upon the wider academic literature to the methodologies used and the accurate interpretation of findings. At present, these findings set a rather less positive picture of how the Police Now participants seek to evidence their work. Of 46 Police Now Case for Change (2014) 47 course the caveats should be acknowledged again that we only assessed a small number of posters and presentations and we were only able to assess what was written down by participants. It is entirely possible more evidence based workings are going on that we did not capture in our assessment, and it is also possible the workings of Police Now participants we have seen here are better than regular DWOs. The findings here illustrate that more work is required to support officers to practically embed EBP within their roles. Over the lifetime of this evaluation, Police Now has improved their training and support for EBP, including developing EBP masterclass days and developing more guidance for participants to effectively evidence the impact of their work. These are positive steps and ultimately, one of the measures of success for Police Now should be the extent to which their officers are effectively engaging in EBP (i.e., drawing on up-to-date literature, identifying clear problems and clear success criteria, delivering the project as expected, carrying out robust analysis and making accurate interpretations). Future independent evaluations should seek to measure this. Within this conversation is the wider challenge of enabling the use of evidence within the police service in general. This is clearly not only a Police Now challenge. Much has been written about how to promote EBP within policing and many hurdles have been outlined, some of which are complex (e.g., receptivity to evidence, risk averse cultures, pressures to find and report positive results), whereas others are seemingly easier to address (e.g., training, access to academic journals, time to read).47 It is to the credit of Police Now that this dialogue is happening with their officers; however, more still needs to be done, and with this in mind, both the expansion of Police Now and the upcoming PEQF may have important roles to play in seeking to cultivate professional change. Measuring Police Now’s Impact Police Now’s stated aim is to “transform communities, reduce crime and increase the public’s confidence in policing, by recruiting and developing outstanding and diverse individuals to be leaders in society and on the policing frontline”. The current evaluation explored whether Police Now had any measurable impact on these outcomes using a robust methodology. Data from the MOPAC PAS showed that, overall, when looking at all respondents (i.e., 2,732 respondents from Police Now wards and 4,218 respondents from comparison wards), there was no change in public confidence or perceptions of the police within the first year of the programme for Police Now wards compared to a set of statistically matched wards. Looking deeper, there were larger improvements in ‘how good a job’ young people thought the police were doing in Police Now wards; but, in contrast, there were other questions that showed the opposite pattern. For example, comparison wards made more improvements in the proportion of young people who thought the police were ‘dealing with the things that matter’. These specific findings relating to young people are based on relatively small numbers of respondents and preclude firm ‘cause and effect’ statements on impact. 47 Lum et al. (2012); Weatheritt (1986) 48 When looking at ward-level crime data, Police Now wards had higher crime volumes than comparison wards in the year prior to the programme. Although the two sets of wards were statistically equivalent on the vulnerability locality index (which includes burglary and criminal damage crime rates), the non-equivalence in total notifiable offences and other specific crime rates suggests some limitations with the comparison group. Finding a well-matched comparison group will always be a challenge for Police Now given they specifically select the most vulnerable and high crime areas. Despite possible limitations to the matching, there were no differences between the groups in how crime changed subsequent to Police Now; both sets of wards showed comparable increases in crime. These findings do not indicate that Police Now does not work or cannot eventually reach such outcomes. Indeed, innovation can take time to bed in and it is worth noting that the research was undertaken with only the first cohort of Police Now. As Police Now grows, independent analysis should continue to evaluate these outcomes. As a first step, it would worthwhile to repeat the analysis in this report with a longer time period (i.e., 18 months before the programme vs. 18 months after) and to combine the first cohort with later cohorts. However, on a wider note, it should be stated that looking for ward level impacts upon such measures as public confidence and crime—measures that can both be stubborn to change—will be difficult, let alone ensuring the research is able to confidently assign any impact to Police Now. This makes it all the more important for Police Now officers to evidence their individual impact through their impact assessments. Conclusion This report presented findings from a large-scale two-year evaluation of the first Police Now programme. It includes learning both for Police Now and the wider police organisation. Findings were generally positive about the training delivered by Police Now and the majority of the first cohort of participants would recommend the programme to other graduates. However, there were some criticisms around the length and timing of the training and the training given on the practical aspects of the job. Addressing this balance between keeping the intensity of the training and at the same time incorporating more practical and in-depth learning (around areas such as EBP) will be a key challenge for Police Now going forward. Other key challenges moving forward will be around strengthening their participants’ practical use of EBP, making sure the officers continue to engage with Police Now’s aim despite the challenging occupational terrain, and continuing to explore ways in which the programme can robustly evidence its success. 49 Appendix A Police Now DWO Job Description Key Objective To police one of the most challenged Wards in the United Kingdom (by deprivation, crime levels and levels of confidence in policing) and contribute to the transformation of that community as a key local community leader. Key Measures of Success a) Contribute to the delivery of measurable reductions in crime and increases in the public’s confidence in policing within that ward; b) Develop the partnerships necessary to deliver effective long-term problem solving leading to the reduction of crime and anti-social behaviour; c) Develop relationships with members of the local community and effectively respond to and anticipate their concerns leading to an increase in the public’s confidence in policing; d) Contribute to the development of evidence based practice in policing. Key Duties and Tasks 1. UNDERSTAND - Develop an understanding of the Ward and Borough and build productive relationships ▪ Analyse and develop an understanding of the Ward and Borough environment, community and socioeconomic factors ▪ Analyse and develop an understanding of the crime drivers, crime patterns and drivers of public confidence in policing both generally and specific to the Ward and Borough ▪ Speak with residents, other agency partners and local business people to understand their context, goals and views ▪ Develop an understanding of the history of policing in the Ward and Borough – what has happened in the past and what is happening now; what has worked and what hasn’t worked? ▪ Develop an understanding of the evidence base of what works in reducing crime and increasing the public’s confidence in policing – which interventions are likely to be most successful? ▪ Identify potential sources of local intelligence and act appropriately on information received 2. CONSULT & PLAN - Develop a Ward Policing Plan, in line with Borough Strategic Objectives, which effectively targets the key drivers of crime (Victims, Offenders, Locations & Times) ▪ Develop with local partners and the community an effective 100 day Ward Policing ▪ Plan based on evidence based practice - targeting interventions which will have the most disproportionate impact 50 ▪ Identify the key outcome measures of success ▪ Through a range of means consult with and present the plan to stakeholders, including colleagues, partners, local business people and residents ▪ Liaise effectively with MPS colleagues to deliver all aspects of the Policing Plan – including presentation to senior officers and colleagues ▪ Be a positive agent for change in the community 3. DELIVER – Deliver the Ward Policing Plan and conduct the duties and tasks of a Police Constable ▪ Lead locally based policing operations and problem solving approaches (for example local regeneration projects, youth crime diversion programmes etc.) in accordance with the Ward Policing Plan ▪ Deliver neighborhood promises and priorities on behalf of, and in partnership with, the MPS partner agencies and the local community ▪ Develop, engage with and support community partnerships ▪ Patrol the Ward providing effective visibility in accordance with the Ward Policing Plan ▪ Provide an effective initial response to crime, anti-social behaviour and other incidents – including the preservation of life, interviewing and arresting suspects, interviewing witnesses and victims and securing evidence conduct primary & secondary investigations – including progressing cases through the criminal justice process ▪ Incorporate outstanding and empathetic victim care in all policing activities ▪ Identify and police offenders within the community ▪ Support the response to critical incidents 4: MONITOR & TAKE FEEDBACK – Review the Ward Policing Plan and evaluate what works and what doesn’t ▪ Evaluate the impact of operations and problem solving approaches ▪ Using operations and problem solving work contribute to the development of evidence based policing practice including completing publications where possible ▪ Represent the MPS at meetings with stakeholders, residents and partners – take regular feedback and act proactively on it ▪ Be a role model within the local community ▪ Use social media and other innovative means of effectively communicating with the public and stakeholders - in particular enabling the communication with hard to reach groups and young people to understand if the plan is working or not and how it is being received. 51 Appendix B 2015 Police Now Programme Assessment Process 52 Appendix C Public Attitude Survey questions relating to public confidence and neighbourhood policing 1. Taking everything into account, how good a job do you think the police in this area are doing? {1=Excellent, 2=Good, 3=Fair, 4=Poor, 5=Very poor} 2. Do you know how to contact your local policing team? {Yes/No} 3. How well informed do you feel about what the police in this area have been doing over the last 12 months? {1=Very well informed, 2=Fairly well informed, 3=Not at all informed} 4. To what extent do you agree with these statements about the police in this area? {1=Strongly agree, 2=Agree, 3=Neither agree nor disagree, 4=Disagree, 5=Strongly Disagree} a. They can be relied on to be there when you need them b. The police in this area treat everyone fairly regardless of who they are c. They are dealing with the things that matter to people in this community 5. The police in this area listen to the concerns of local people {1=Strongly agree, 2=Agree, 3=Neither agree nor disagree, 4=Disagree, 5=Strongly Disagree} 6. How much of a problem are…? {1=Very big problem, 2=Fairly big problem, 3=Not a very big problem, 4=Not a problem at all} a. Vandalism, graffiti and other deliberate damage to property or vehicles b. People using or dealing drugs c. People being drunk or rowdy in public places 7. To what extent are you worried about crime in this area? {1=Very worried, 2=Fairly worried, 3=Not very worried, 4=Not at all worried} 8. How safe do you feel walking alone in this area after dark? {1=Very safe, 2=Fairly safe, 3=Fairly unsafe, 4=Very unsafe} 53 Appendix D Evidence Based Policing Rating Criteria for 100 Day Impact Assessments Using Evidence 1. Was a review of relevant literature carried out? 2. Were lessons from the literature incorporated into project design? Rationale 3. Is there a clear problem that has been verified by data? 4. Is project success clearly defined? Delivery 5. Was the project delivered as expected? 6. Was data captured throughout? Measuring Outcomes 7. Have outcomes been clearly measured? 8. What methodology was used to measure outcomes? 8a. Process: What data collection methods were used (e.g., interviews, surveys, focus groups, observations)? 8b. Process: How many participants were included? 8c. Process: What type of participants were included (e.g., residents, business owners, Local Authority staff)? 8d. Process: How was data analysed (e.g., thematic analysis, no evidence of thematic analysis)? 8e. Impact: What type of impact methodology was used? 8f. Impact: How many participants/cases were included? 8g. Impact: Timeframes of monitoring (e.g., 6 months before vs. 6 months after)? 8h. Impact: How was data analysed (e.g., basic numerical analysis, basic statistical tests, advanced statistical tests, no evidence of analysis)? Interpretation 9. Are caveats outlined (e.g., sample size, lack of comparison, timeframes)? 10. Are appropriate interpretations made? 54 Appendix E Question Tracker Survey 1 July 2015 (n=69) Tracker Survey 2 Oct 2015 (n=45) Tracker Survey 3 July 2016 (n=38) Tracker Survey 4 Feb 2017 (n=23) % change At this stage, how satisfied are you that your chosen career as a police officer is right for you? (% very/fairly satisfied) 92.7% 76.1% 76.3% 78.3% -14.4% I am satisfied with my current job n/a 78.2% 73.7% 69.6% -8.6% I feel motivated at work n/a 82.6% 76.4% 65.2% -17.4% I see myself spending my whole career in the MPS 65.0% 32.6% 31.6% 52.2% -12.8% The MPS is a good organisation to work for n/a 76.0% 57.9% 52.2% -23.8% I feel a sense of loyalty to the MPS n/a 78.3% 73.7% 73.9% -4.4% It is clear to me what our organisational goals are n/a 65.2% 44.7% 47.8% -17.4% If I contacted the MPS as a member of the public I would be confident of receiving a good service 75.4% 63.0% 50% 34.8% -40.6% I am proud to work for the MPS n/a 93.5% 86.9% 87.0% -6.5% Being concerned about customer service should not be a big part of a police officer’s job 7.2% 17.7% 21.0% 18.2% 11.0% Officers should treat everyone with the same level of respect regardless of how they behave towards the police 94.2% 86.6% 73.7% 77.3% -16.9% Good police-community relations has no effect on the crime rate 2.9% 4.4% 10.5% 4.6% 1.7% Neighbourhood policing is not ‘real’ policing 2.8% 0% 2.6% 0% -2.8% Some victims of crime are more deserving of a good service than others 2.9% 2.2% 7.9% 13.6% 10.7% There are certain communities that do little to deserve the respect of the police n/a 4.3% 2.6% 4.6% 0.3% At this stage, to what extent do you agree that you understand the needs of the community you serve? n/a 78.3% 94.7% 90.9% 12.6% I have confidence in the leadership provided by my local senior leadership team n/a 65.2% 44.8% 54.6% -10.6% I have confidence in the leadership provided by the senior leaders in the MPS n/a 73.9% 36.8% 36.4% -37.5% Senior managers are out of touch with what is happening on the front-line n/a 23.9% 47.4% 45.5% 21.6% The values of the MPS are reflected in the behaviour of my local senior leadership team n/a 67.4% 55.3% 54.6% -12.8% Managers on my borough treat all staff fairly, regardless of their background n/a 76.1% 63.2% 72.7% -3.4% At this stage, how satisfied are you with the level of leadership and support you have received from your line manager? (% very/fairly satisfied) n/a 73.9% 78.9% 90.9% 17.0% My line manager gives regular and constructive feedback on my performance n/a 43.5% 57.9% 68.2% 24.7% I feel respected by my line manager n/a 76.0% 86.8% 90.9% 14.9% I feel comfortable expressing my opinion to people more senior than me n/a 76.1% 76.3% 77.3% 1.2% 55 Note: figures represent % strongly agree/agree unless otherwise stated; n/a=not asked in that wave of the survey I am encouraged to use my discretion when making decisions n/a 76.0% 73.7% 81.8% 5.8% I am treated with respect by my colleagues on borough n/a 71.8% 86.8% 72.7% 0.9% I achieve a good balance between my work life and private life n/a 50.0% 47.4% 45.5% -4.5% I am given the resources and equipment I need to do my job n/a 37.0% 18.4% 18.2% -18.8% I am rewarded fairly for the work I do n/a 45.7% 31.6% 18.2% -27.5% There is a positive atmosphere among the people I work with n/a 56.5% 39.5% 50.0% -6.5% The amount of work I am expected to do is fair n/a 76.1% 36.8% 50.0% -26.1% Serving as a police officer is not as exciting as I anticipated n/a 30.4% 23.7% 31.8% 1.4% In my role I am encouraged to develop innovative solutions to problems n/a 71.7% 73.7% 63.6% -8.1% In my role I am able to positively influence local crime patterns n/a 54.4% 60.6% 59.1% 4.7% In my role I am able to positive influence community confidence in the police n/a 82.6% 81.6% 63.6% -19.0% There is no point in officers trying to convince some communities in London that the police can be trusted 0% 2.2% 2.2% 9.1% 6.9% It is important that the MPS workforce ‘looks and feels’ like the diverse communities that it serves 88.4% 77.7% 76.1% 63.6% -24.8% There is a stigma attached to being part of the Police Now programme n/a 54.3% 52.7% 50.0% -4.3% When meeting new colleagues I feel comfortable disclosing that I am part of the Police Now programme n/a 47.8% 39.5% 50.0% 2.2% I feel a sense of loyalty to Police Now n/a 82.6% 71.1% 63.6% -19.0% I have a strong bond with my fellow Police Now colleagues n/a 89.1% 65.8% 68.2% -20.9% I talk positively about Police Now when I speak to other people n/a 82.6% 76.4% 72.7% -9.9% I would encourage other graduates to join Police Now n/a 95.7% 81.6% 72.7% -23.0% I feel I have been accepted by colleagues on my borough n/a 74.0% 92.1% 86.4% 12.4% It is clear to me what the goals of Police Now are n/a 86.9% 89.5% 86.4% -0.5% The Police Now team are working hard to make the programme a success n/a 100.0% 94.7% 90.9% -9.1% In terms of your professional development, how satisfied are you with the level of support you have received from the MPS since completing the Summer Academy? n/a 66.7% 42.2% 54.6% -12.1% 56 Appendix F Public Attitude Survey Data – Police Now and Comparison Wards Police Now Wards Comparison Wards Oct 14- Sep 15 n=1,450 Oct 15- Sep 16 n=1,282 Change Oct 14- Sep 15 n=2,319 Oct 15- Sep 16 n=1,899 Change Taking everything into account, how good a job do you think the police in this area are doing? (% Excellent/Good) 61.8% 64.1% 2.3% 65.1% 68.5% 3.4%* Do you know how to contact your local policing team? (% Yes) 45.4% 37.4% -8.0%* 39.8% 36.7% -3.1%* How well informed do you feel about what the police in this area have been doing over the last 12 months? (% Very/Fairly well informed) 41.5% 41.2% -0.3% 41.5% 41.1% -0.4% They can be relied on to be there when you need them 74.9% 77.0% 2.1% 75.5% 79.2% 3.7%* The police in this area treat everyone fairly regardless of who they are 71.5% 74.4% 2.9% 72.0% 78.0% 6.0%* They are dealing with the things that matter to people in this community 68.9% 72.7% 3.8%* 70.7% 75.2% 4.5%* The police in this area listen to the concerns of local people 72.8% 75.0% 2.2% 70.5% 78.0% 7.5%* How much of a problem are…? (% Not a very big problem/Not a problem at all) Vandalism, graffiti and other deliberate damage to property of vehicles 82.5% 81.1% -1.4% 84.6% 82.5% -2.1% People using or dealing drugs 66.7% 61.0% -5.7%* 69.7% 72.3% 2.6% People being drunk or rowdy in public places 73.3% 68.5% -4.8%* 78.2% 78.2% 0.0% To what extent are you worried about crime in this area? (% Not very/Not at all worried) 58.3% 62.2% 3.9%* 62.9% 66.0% 3.1%* How safe do you feel walking alone in this area after dark? (% Very/Fairly safe) 77.1% 77.4% 0.3% 77.9% 80.2% 2.3% Note: figures represent % strongly agree/agree unless otherwise stated; *Change is significant at the p<.05 level 57 Public Attitude Survey Data – Age Comparison Police Now Wards Comparison Wards 16-24 year olds 25+ year olds 16-24 year olds 25+ year olds Oct 14- Sep 15 n=206 Oct 15- Sep 16 n=202 Change Oct 14- Sep 15 n=1,238 Oct 15- Sep 16 n=1,078 Change Oct 14- Sep 15 n=276 Oct 15- Sep 16 n=237 Change Oct 14- Sep 15 n=2,029 Oct 15- Sep 16 n=1,658 Change Taking everything into account, how good a job do you think the police in this area are doing? 51.9% 66.4% 14.5%* 64.1% 63.6% -0.5% 66.2% 70.6% 4.4% 64.8% 68.0% 3.2%* Do you know how to contact your local policing team? 43.8% 33.0% -10.8%* 46.0% 38.4% -10.2%* 37.8% 37.2% -0.6% 40.2% 36.6% -3.6%* How well informed do you feel about what the police in this area have been doing over the last 12 months? 30.5% 37.6% 7.1% 44.0% 42.2% -1.8% 37.3% 34.9% -2.4% 42.5% 42.4% -0.1% They can be relied on to be there when you need them 75.0% 79.6% 4.6% 74.9% 76.4% 1.5% 75.5% 79.2% 3.7% 75.7% 79.3% 3.6%* The police in this area treat everyone fairly regardless of who they are 66.7% 76.8% 10.1%* 72.7% 73.8% 1.4% 66.1% 76.2% 10.1%* 73.3% 78.5% 5.2%* They are dealing with the things that matter to people in this community 66.0% 71.7% 5.7% 69.5% 72.9% 3.4% 69.8% 81.2% 11.4%* 70.9% 74.1% 3.2%* The police in this area listen to the concerns of local people 69.9% 75.4% 5.5% 73.7% 74.9% 1.2% 68.7% 80.0% 11.3%* 71.0% 77.6% 6.6%* *Change is significant at the p<.05 level 58 References Build a Better Met: MPS Staff Survey 2015 College of Policing (2016). http://www.college.police.uk/What-we-do/Learning/Policing- Education-Qualifications-Framework/Pages/PEQF-archive.aspx Fenn, L. (April 2014). Expectations MET? Evidence and Insight, MOPAC. Karn, J. (2013). The Police and Crime Reduction: The evidence and its implications for practice. London: Police Foundation. Loftus, B. (2009). Police Culture in a Changing World. Oxford: Oxford University Press. Lum, C., Telep, C., Koper, C., & Grieco, J. (2012). Receptivity to research in policing. Justice, Research and Policy, 14, 1-35. Neyroud, P. (2011). Review of police leadership and training, two volumes. London: Home Office. Paoline, E. A. (2003). Taking Stock: Toward a Richer Understanding of Police Culture. Journal of Criminal Justice, 31, 199-214. Police Now. Impact Report 2017 https://www.policenow.org.uk/wp- content/uploads/2017/08/2017-Police-Now-IMPACT-Report-1.pdf Sherman, L. (1998). Evidence-based policing. Ideas in American policing. Series. Washington, DC: Police Foundation. Spencer, D., Lloyd, M., & Stephens, L. (2014). Police Now: The Case for Change http://www.policenow.org.uk/downloads/case-for-change.pdf Stanko, E., A., & Bradford, B. (2009). Beyond measuring ‘how good a job’ police are doing: The MPS model of confidence in policing. Policing: A Journal of Policy and Practice, 3, 322- 330. Stanko, E. A., & Dawson, P. (2016). Police use of research evidence: Recommendations for improvement. Springer TeachFirst. Our Impact https://www.teachfirst.org.uk/what-we-do/our-impact The UK Graduate Careers Survey 2015. High Fliers Research Ltd: London. Weatheritt, M. (1986). Innovations in Policing. Beckenham: Croom Helm. Weiss, C., & Bucuvalas, M. J. (1980). Social science research and decision-making. New York, NY: Columbia University Press. 1 Police Now Evaluation Year 1 Report April 2016 Julia Yesberg, Liam Fenn & Paul Dawson MOPAC Evidence and Insight 2 Contents Executive Summary .................................................................................................................. 3 1. Introduction ......................................................................................................................... 4 Report Objectives .................................................................................................................... 4 About Police Now .................................................................................................................... 4 Police Now Evaluation ............................................................................................................. 5 The Police Now Cohort ............................................................................................................ 5 The Summer Academy ............................................................................................................. 6 The 28 Day DWO Immersion Period ........................................................................................ 8 2. Methodology ........................................................................................................................ 9 Data Collection and Methods .................................................................................................. 9 3. Results ................................................................................................................................ 10 The Summer Academy Training ............................................................................................. 10 Confidence and Preparedness................................................................................................ 13 The DWO Role ....................................................................................................................... 15 Impact on the Community ..................................................................................................... 18 Stigma and Acceptance ......................................................................................................... 19 Experience of the Organisation.............................................................................................. 21 Job Satisfaction and Career Aspirations ................................................................................. 24 Attitudes toward the MPS and Policing ................................................................................. 25 Overall thoughts on Police Now ............................................................................................ 27 4. Discussion ........................................................................................................................... 29 Summary and Conclusion ....................................................................................................... 29 Next Steps in Evaluation ........................................................................................................ 30 References .............................................................................................................................. 31 3 Executive Summary Police Now is an innovative two year graduate programme putting graduates on the frontline of policing. The programme aims to bring outstanding graduates into policing with the purpose of transforming some of the most disadvantaged communities. The first cohort of Police Now participants joined the Metropolitan Police Service in July 2015 and completed the intensive six week Summer Academy training in August 2015. Since then, participants have been working as Dedicated Ward Officers in some of London’s most challenged communities. An evaluation of Police Now is being conducted by the Evidence and Insight team at the Mayor’s Office for Policing And Crime (MOPAC) supported by funding from the Home Office Police Innovation Fund. The research uses a longitudinal mixed-methods approach to explore how Police Now participants’ experiences of the programme, the police organisation and the job of policing itself, change over time. This second of four interim reports will share emerging findings from the research, including results from the second tracker survey, interviews with syndicate leads and the Police Now operations team, and interviews with Police Now participants and their line managers. The report will also provide an update on the overall progress of the evaluation and next steps. Key Findings  The first cohort of Police Now participants were highly satisfied with the level and quality of training they received as part of the Summer Academy.  The syndicate structure—where participants are broken into syndicates of 8-9 participants led by a Syndicate Lead—emerged as an important feature of the Summer Academy, along with the positive learning environment and the quality of the visiting fellows and guest lecturers.  Participants thought the training could have had a larger focus on the practical aspects of the job and longer shifts on field training. Many participants felt unprepared to begin their roles as police officers and confidence was low.  More than half of participants reported feeling a stigma associated with being part of the Police Now programme and half reported feeling uncomfortable disclosing they are part of the programme when meeting new colleagues.  Interviews with line managers revealed that many felt unprepared to receive their Police Now officer and some teams were unaware of the Police Now scheme. Line managers reported this lack of awareness may have increased stigma and caused some initial scepticism.  Line managers were extremely positive about their Police Now officers – they described them as enthusiastic, proactive and valued members of their team.  Participants’ attitudes toward policing and serving the public were favourable and higher than the overall MPS staff survey results, although there were some declines in attitudes from the beginning of the Summer Academy. For example, fewer participants were confident of receiving a good service if contacting the MPS as a member of the public after only a few months on the job.  Most respondents thought policing would benefit from the Police Now scheme in the long- term, most notably because it is attracting the ‘right’ people to the role. 4 1. Introduction Report Objectives This report is the second of four reports that will be published by Evidence and Insight1 as part of the Police Now evaluation. The purpose of these reports is to provide a detailed account of the design and progress of the evaluation and to present findings and learning from the first cohort of Police Now participants. The first interim report was published in October 2015.i It outlined the aims and proposed design of the evaluation, introduced the first cohort of Police Now participants, and presented findings from the induction survey administered to the cohort during their first week of training with the MPS. The induction survey captured participants’ motivations for joining the police, aspirations for their career as a police officer, their attitudes towards serving the public, and their expectations of the MPS as an employer. The second interim report will share emerging findings to date and provide an update on the overall progress of the evaluation. Findings from the second ‘Tracker Survey’ will be presented, which captures participants’ early career attitudes, their experiences of the Police Now Summer Academy, and how they are settling in to life as a police officer since completing their training. Findings from this survey will be compared to the first tracker survey to explore how participants’ attitudes change over time and to other MPS staff surveys to explore how Police Now participants’ attitudes compare to the attitudes of other police officers. Findings from interviews conducted with Police Now participants and their line managers will further explore themes that emerged from the tracker survey. Findings will also be presented from interviews conducted with syndicate leads and the Police Now leadership and operations team. Recommendations for Police Now will then be discussed based on the findings and an overview of next steps and timescales for the remainder of the evaluation will be outlined at the end of the report. About Police Now Police Now is an innovative, two year graduate programme putting graduates on the frontline of policing. The programme aims to bring graduates into policing with the purpose of transforming some of the most disadvantaged communities. The programme is committed to developing a new generation of inspiring leaders from a diverse range of backgrounds, whilst providing graduates with a unique foundation for a leadership career. 1 Evidence and Insight is an in-house research and performance team based at the Mayor’s Office for Policing And Crime (MOPAC). The first interim report is available upon request. 5 Neighbourhood policing is “at the heart of the Police Now model”.ii Following their training, the first wave of Police Now participants were placed as Dedicated Ward Officers (DWOs) in some of London’s most challenged neighbourhoods for two years.2 In 2016, the Police Now scheme is expanding to work with seven police forces across the UK: the MPS, Cheshire Constabulary, Lancashire Constabulary, Northamptonshire Police, Surrey Police, Thames Valley Police and West Midlands Police. Police Now Evaluation The Evidence and Insight team at MOPAC are conducting an evaluation of Police Now on this first cohort. The research employs a longitudinal mixed-methods design with the aim of examining how the Police Now participants’ experiences of the programme, the police organisation, and the job of policing itself change over time. The participants themselves have centre stage in this evaluation which will cover key areas such as why people join the police, how they integrate into the organisation, their appraisal of training, and their attitudes toward different policing styles.3 The findings from the Police Now evaluation will make a significant contribution to the policing evidence-base, developing knowledge in key areas such as police recruitment, workforce development, police training and police culture. The Police Now Cohort The first cohort of Police Now participants joined the Metropolitan Police Service (MPS) in July 2015 and completed the intensive six week Summer Academy training in September 2015 (see Figure 1 for the ‘Participant Journey’). Almost half of the 67 participants who completed the training were female and 12% (n=8) were from Black and Minority Ethnic (BME) groups. The proportion of females in the Police Now cohort (45%; n=30) is larger than the overall proportion of female police officers in the MPS (25%)iii and the 2013 Police Officer Graduate Entry programme (40%).iv The proportion of participants from BME groups is double the proportion of BME recruits in the 2013 graduate cohort (6%), but similar to the overall proportion of BME police officers in the MPS (11%). 2 The DWO role was introduced in the MPS in 2013 as part of the development of the Local Policing Model. The Met’s vision for the role is for a single police officer to take responsibility for their geographical patch in each of London’s 624 wards. 3 See Report 1 for a more detailed description of the Police Now Evaluation. Police Now’s stated vision: “transform the most challenged communities by making them safer, reducing crime and increasing confidence in the police by recruiting and developing outstanding individuals to be leaders in their communities and on the policing frontline” 6 Police Now graduates4 45% female from a BME background After completing the Summer Academy in September 2015, participants were posted into their community as DWOs and began the 28-day DWO immersion period (outlined below). Over the next two years, participants will continue to develop their skills as a police officer. Participants will also be given the opportunity to apply for a three-week internship in policing or with one of Police Now’s ‘graduate partner’ organisations.5 At the end of the two years, participants can choose to move on to a career outside of policing or to continue as a police officer. The second cohort of Police Now participants is due to begin the Summer Academy training in July 2016. Figure 1. Police Now ‘Participant Journey’ The Summer Academy Previous research has highlighted the shortcomings of police officer training in England and Wales.v The intensive six week long Summer Academy represents an innovative, forward thinking approach to police officer training and leadership development and aims to build on and enhance traditional police training. The Police Now training is based on National Standards (e.g., College of Policing Learning Objectives) and includes all mandatory training (e.g., OST and ELS training) plus additional Police Now training (see Figure 2 for a breakdown of the training). The four objectives of the Summer Academy are summarised below: - To enable participants to excel at delivering the Police Now mission of transformation in communities. - To enable participants to be exceptional police officers. - To enable participants to be leaders on the policing frontline and in wider society. - To deliver the gold standard in policing and leadership training and development. 4 69 participants started the programme but 2 left during the Summer Academy. 5 Partners include Accenture, Civil Service Fast-Stream, Covingtons, Clifford Chance, KPMG, PA Consulting. 6712% March '15 Spring Induction March-July '15 Pre-Learn July-Aug '15 Summer Academy Sept-Oct '15 28 Day DWO Immersion Period Oct '15-July '17 DWO Posting Oct '15- July '17 Ongoing Development Seminars July ‘16-July '17 3 Week Internship July '17 Programme Graduation 7 Figure 2. Hours breakdown of the Summer Academy The Summer Academy consists of a mixture of classroom and practical training and is highly intensive; participants attend 11 hour days over 6 weeks consisting of approximately 300 hours of learning (see Figure 3 for the Summer Academy timetable). Throughout the Summer Academy, visiting fellows drawn from inside and outside of policing attend to deliver lectures, seminars and practical training on their areas of expertise. In the last three weeks of the Summer Academy, participants take part in Borough Field Training to develop their skills and knowledge in a real life environment. The field training consists of 15 shifts (5 ERPT shifts, 5 SNT shifts and 5 CID/CPU shifts) under the 1:1 supervision of an experienced officer. Throughout the Summer Academy, participants undergo regular theoretical and practical assessments which meet and exceed current MPS foundation course assessments. During their 2 year postings, participants also complete 100 Day Impact Assessments. In these assessments, participants must answer the question: What is the most impactful thing you have done on your Ward to reduce crime and increase the public’s confidence in policing during the last 100 days? Participants are encouraged to focus on measurable impact and describe the innovations they have made over the preceding 100 days. Regular reflection and support is also a feature of the Police Now programme, with coaching, mentoring and buddying provided throughout. Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 0800-0850 Written Assessment, Practical Assessment or Reflection Session 0900-1300 Classroom and Practical Exercises 1300-1400 Lunch and/or Travel 1400-1900 Mandatory OST/ELS and Practical Exercises Practical Exercises and Borough Field Training Sessions Figure 3. Police Now Summer Academy Timetable 36 27 13 17 7 0 5 10 15 20 25 30 35 40 45 50 Assessment & Immersive Learning Traditional Police Training Police Now Additional Training Mandatory Training Administration % of Summer Academy 8 The syndicate structure is an integral part of the Summer Academy and the Police Now programme. Police Now participants are broken into 8 syndicates of 8-9 participants led by a Syndicate Lead who is a substantive police officer with recent frontline policing experience. The syndicate lead’s role is to inspire and lead participants through the Summer Academy while also preparing and delivering training content. They are responsible for monitoring and evaluating participant performance, providing advice, and identifying further development needs. Syndicate leads are also responsible for identifying and responding to individual welfare needs and resolving problems when they arise. Figure 4. Police Now Summer Academy Staff Structure The 28 Day DWO Immersion Period Following the Summer Academy, the 28 day DWO immersion period is designed to introduce participants to the DWO role under the 1:1 supervision of an experienced DWO mentor based on a nearby or neighbouring borough. As part of the immersion period, participants are required to complete the Student Officer Record of Competence (SOROC) assessments and are also expected to commence their first 100 day plan for their community.6 6 See https://policenow.wordpress.com/2016/02/29/100-day-impact-event/ for a video of the first 100 day impact event held on 22 January 2016. Gold: Programme Lead Silver: Course Director Bronze: 8 Syndicate Leads 8 Syndicates: Each with 8-9 PN Participants 9 2. Methodology Tracker Survey The second ‘Tracker Survey’ was administered to participants in October 2015 at the end of the 28 day DWO immersion period.7 The survey was designed to capture participants’ early career attitudes, their experiences of the Police Now Summer Academy and 28 day DWO immersion period, and their aspirations for their future career. Each participant was emailed a link to an online survey; 45 people completed the survey, indicating a response rate of 67%. Participants were asked to provide their warrant number to match their responses to the first tracker survey. Interviews Interviews were conducted with syndicate leads and the Police Now leadership team in October-November 2015. The interviews were designed to capture their experiences of the Summer Academy (including what worked and what could be improved upon), their insight into their own career and how Police Now differs to the training they received as new recruits, and the current state of policing and the MPS. Face-to-face interviews were completed with 18 people lasting approximately 45 minutes. Interviews were recorded, transcribed and analysed. Telephone interviews were also conducted with a sample of Police Now participants (n=22) and a sample of participants’ line managers (n=25) to further explore themes uncovered in the tracker survey.8 Interviews were conducted in February-March 2016 and lasted approximately 40 minutes. Interviewees were recruited via email or in person at the first 100 Day Impact Event and were asked to sign a consent form outlining confidentiality and anonymity. Reflection Journals A sample of Police Now participants (n=16) submitted reflection journals for thematic analysis.9 These journals were completed by participants at two time points: during the Summer Academy and the 28 day DWO immersion period. Participants were asked in these journals to set goals, list their strengths and areas for improvement, and document their experiences of the training. Other Surveys This report also draws on data collected from surveys administered to participants by the Police Now operational team at fortnightly intervals throughout the Summer Academy and a short survey administered to the line managers of participants following the Summer Academy. Comparisons to other surveys (e.g., surveys conducted with recruits from the Police Officer Graduate Entry programme in 2013 and the Met staff survey) will also be made. 7 The survey was developed and hosted using the Opinion Research Services (ORS) secure online survey portal. 8 These interviews were completed by ORS’s qualitative research team. 9 Participants were randomly selected. Each participant signed a consent form permitting their journals to be analysed for the purpose of the research. 10 3. Results The Summer Academy Training Overall satisfaction Participants - Findings from Tracker Survey 2 (TS2) showed that, overall, participants were happy and satisfied with the level and quality of training they received at the Summer Academy. - Eighty-nine percent of participants (n=40) reported being very or fairly satisfied with the quality of training received and all 45 participants agreed the Summer Academy was a positive learning environment. Overall, how satisfied were you with the quality of training you received during the Summer Academy? Very dissatisfied Neutral Very satisfied 0% 2% 9% 62% 27% 2% 89% - Similarly, the interviews conducted with Police Now participants (n=22) also demonstrated very positive views about the quality of the training delivered at the Summer Academy. Absolutely fantastic. I didn’t know what to expect beforehand but I previously received training from a graduate employer and I was amazed at how much better the Police Now training was. Syndicate leads - Syndicate leads and the Police Now operations team were similarly positive about the training delivered at the Summer Academy. In particular, they commented in their interviews that the training Police Now offers is far superior to the training they received as new recruits. Police Now training is inspiring, all the things that training experienced in my career isn’t. It is dynamic, creative and delivered by practitioners. What worked well? Participants - Participants were asked in the survey and interviews to describe what worked well and what they enjoyed most about the Summer Academy (see Figure 5 for a word cloud of the most commonly reported words). - The syndicate structure emerged as an important feature. Participants thought the syndicate structure gave them the opportunity to bond with their fellow cohort members as well as creating a more relaxed and encouraging learning environment. Syndicate leads were cited as being instrumental to the success of the syndicate model, allowing participants to learn in a safe and supported environment. 11 - The positive learning environment was mentioned by a number of participants. They reported enjoying being surrounded by enthusiastic, interesting, and dynamic frontline officers with expertise in different areas. They noted that the people providing the training were engaged with the topics and were enthusiastic about passing on their expertise. They also commented on the quality of the visiting fellows and guest lecturers. What made it exceptional was having officers who are currently serving being involved; it meant that the training was much more relevant. - Participants also reported enjoying the intensity of learning more quickly via the six- week Summer Academy and the field-training aspect; however, as discussed in the next section, many participants felt they could have benefited from more field-training prior to starting their roles as DWOs. - Finally, participants were particularly positive about the DWO-specific training provided, such as soft skills and problem-solving. Figure 5. Police Now participants were asked what worked well and what they enjoyed most about the Summer Academy. The word cloud above presents the most commonly reported words. Syndicate leads - Consistent with participants’ perceptions of what worked well, interviews with the syndicate leads and operations team suggested the syndicate structure worked particularly well, along with the combination of trainers (e.g., syndicate leads and visiting fellows with expertise in different subject matters). Because the trainers were motivated in their delivery and their knowledge base was very good, interview respondents reported the cohort responded well and were engaged in the training. Police Now trainers were proper street cops – they lived and breathed policing. - The combination of practical learning and classroom based exercises was another aspect of the Summer Academy syndicate leads thought worked well. What could be improved? Participants - Participants were also asked how the Summer Academy could be improved for future cohorts. Firstly, the six-week time period was considered too intense for some, 12 especially in relation to: not having enough time to revise and review information and knowledge; long days causing tiredness and exhaustion; and the ‘rushed’ structure of training days. There was little or no time to consolidate knowledge or go over what I’d learnt. The hours were long; we started at 8am and finished at 7pm, plus travel time, so there was no time to go over anything in the evenings either. - Consistent with these comments, the majority of the cohort (77%; n=35) thought the Summer Academy was too short and 93% (n=42) thought the Summer Academy was exhausting. In addition, many participants thought the Summer Academy overall could have had a larger focus on the practical aspects of the job; one quarter of participants (n=10) were dissatisfied with the balance between classroom and practical learning. The Summer Academy was too short Strongly disagree Neutral Strongly agree 5% 7% 11% 46% 32% 12% 77% - Another suggestion for improvement by the cohort was the field training period, including its timing, structure and organisation. Many participants thought the three week field training period was too short to adequately prepare them for the front line. They also thought the field training shifts could have been longer (i.e., full-day) and more varied (i.e., different shift patterns, including nights). Some participants reported that because they were only able to do half a shift, they could not properly get involved in incidents or, if they did, they finished late and were exhausted in the morning. I think there needs to be further field training. There simply wasn’t enough time to practice key policing skills. - In terms of the organisation of the field training, many participants reported that their teams apparently knew nothing about the Police Now scheme and were unaware that they were supposed to be providing training for recruits. Furthermore, participants reported inconsistencies in the extent to which different team members were willing to help them with their development and training. - Some participants also felt that the field training could have had a wider emphasis on the processes and procedures of overall policing, rather than specifically focused on the DWO role. A number of participants reported since starting their roles as DWOs, because of the heavy workloads of their team, they have been drawn into other police work which they have felt unprepared for. I think the training was good in that it prepared us for the role that was advertised, however I think in reality I have spent a lot of time doing things I wasn’t trained for. - Participants also commented there could have been better integration with the boroughs and more time could have been allocated for feedback sessions with the syndicate leads. 13 Syndicate leads - Syndicate leads agreed with the cohort that for practicality purposes the days should not be split between classroom and field-training and that a full day should be spent out on borough. They also thought that more time could be spent on certain exercises (e.g., role plays) in order to boost confidence. - Some syndicate leads felt they did not have enough time themselves to prepare for the training they delivered, although recognised part of this problem might be exclusive to the first year rollout. Syndicate leads reported feeling exhausted towards the end of the Summer Academy, finding it difficult to keep morale up. Review the 6 weeks duration, as syndicate leads were exhausted afterwards. Does there need to be more thinking and reflection time, and perhaps more recovery time? - These feelings are reflected in participants’ responses to the surveys administered by the Police Now operational team at fortnightly intervals throughout the Summer Academy. Responses to these surveys showed there was a drop in attitudes in the survey administered four weeks into training. In the second survey, 84% of participants (n=56) stated they would recommend the Summer Academy to others compared to 93% (n=62) in the first survey and 97% (n=65) in the survey administered at the end of training. This pattern of results was reflected across other responses; for example, fewer participants in Survey 2 rated the quality of the training they received as “Excellent” or “Very Good” (70%) compared to the other two surveys (84% in Survey 1 and 95% in Survey 3). Confidence and Preparedness Participants - Many participants reported not feeling prepared to start their roles as police officers. Less than half of the cohort (43%; n=19) thought the Summer Academy equipped them with the necessary skills to embark on their career as a police officer and only 36% (n=16) felt they had the necessary confidence10. The Summer Academy equipped me with the necessary skills to embark on my career as a police officer Strongly disagree Neutral Strongly agree 0% 18% 39% 29% 14% 18% 43% At the end of the Summer Academy I had the necessary confidence to start my career as a police officer 7% 25% 32% 27% 9% 32% 36% 10 The evaluation team plans to speak to non-Police Now recruits who joined the MPS in 2015 as a comparison group to determine whether Police Now recruits are more or less confident following training than other recruits who go through standard MPS training. 14 - Confidence appeared to be a particular problem for female participants, with significantly fewer females (20%; n=4) reporting feeling confident compared to their male counterparts (50%; n=12)11. - Developing the confidence and knowledge to deal with the varying aspects of being a DWO was a challenge identified by participants in the interviews. Some put this down to the inevitable nature of starting a new job, while others felt unprepared and unknowledgeable about certain aspects of their role. Line managers - The Police Now operational team conducted a short survey with participants’ line managers following the Summer Academy. Of the 38 line managers surveyed, just over half (n=21) thought the officers were well prepared for the role. Comments from line managers suggested they thought the training was too short and an inadequate introduction to the basics. However, 40% of line managers (n=15) thought their officers’ abilities were just the same as other officers with similar levels of service and 47% (n=18) thought their abilities were better or far better than other officers. - Preparedness was also explored in interviews. Again, just over half of the 25 line managers interviewed (n=13) believed the training had prepared their officer sufficiently. - Nine line managers did not think the training academy had prepared their officers for the DWO role because: they believed the training period was too short to convey proper knowledge about being a DWO and policing procedures; they believed there to be discrepancies between the real-life DWO role and the one the scheme promotes to officers. Furthermore, a lack of general policing experience was identified by line managers as the main challenge to line managing their Police Now officer. A lot of the programme is about problem-solving and DWOs. The only thing for me was, as well as doing that side of the job they also have to learn how to be a police officer. It wasn’t a monumental challenge, but they have to understand the basics of what doing our job is about. - The particular areas in which line managers thought their Police Now officer should have received more training prior to starting their DWO role were: ‘overall’ policing including response team and street duty training (especially for safety reasons); more DWO-specific remits such as problem-solving, partnership working, ASB management and Safer Neighbourhood work; and administrative processes. - Line managers were also asked about the expectations they had of their Police Now officer prior to their arrival. Line managers had mixed expectations of their Police Now officers. Some had high hopes and expected someone with innovative ideas and a can- do attitude, whereas others were more cynical and were unsure about the length of the training period and whether a university degree could assist in generating new ideas or conversely whether it was the high calibre of the recruits instead of the training that was yielding benefits. - Positively, 19 of the 25 line managers said their Police Now officers are meeting or exceeding their expectations. The minority of line managers who did not feel their 11 According to an independent samples t-test, t(42) = 2.82, p = .007, 95% CI [.25, 1.47], d = .87 15 officers were meeting their expectations stated they were anticipating someone with more experience. She is doing very well. However I’m not sure if this is because of the training or due to the high quality of the people the scheme is recruiting…myself and the other sergeants tend to think the latter. Syndicate leads - In addition to the concerns noted by line managers, there were some concerns among syndicate leads that Police Now participants are not receiving the same level of training compared to Hendon and that the six week duration may not be long enough12. Syndicate leads emphasised that Police Now has a responsibility to make sure participants on the scheme are properly trained and any mistakes accounted for. Some syndicate leads suggested training could be residential to promote more opportunities to debrief and bond with cohort members, in addition to welfare duties. The DWO Role The 28-Day DWO Immersion Period Participants - Participants were asked in TS2 how they found the 28 day DWO immersion period. Specifically, they were asked to describe what they learned and what they found challenging (see Figure 6 for a word cloud of the most commonly reported words). - Overall, most participants stated they learned a lot from this period; however, many reported this being a challenging time. The quality and availability of their DWO mentor appeared to be a major factor in how participants found the immersion period, as well as the level of support offered by their team. - Participants who were positive about the 28 day DWO immersion period reported: a high level of support from their team; having a mentor with identical shift patterns; having a mentor who was experienced and knowledgeable, and who was a fellow DWO. Brilliant. My team made me feel welcome and have provided me with support. Having a mentor with identical shift patterns really helped. - Participants who were less positive about the immersion period reported: having an absent or unavailable mentor or a mentor with little experience; having a mentor who was not interested in mentoring; the team having little knowledge of Police Now or what the immersion period was meant to achieve; a lack of motivation and low morale within the team; and a lack of planning by the borough. I didn’t really have an immersion period. I was not given one to one mentoring as my DWO ‘partner’ is frequently absent and had no interest in mentoring me. - Other challenges reported by participants included understanding the different computer systems and procedures, integrating with colleagues on borough, and managing colleagues’ perceptions of the DWO role. 12 Standard MPS recruit training is 12 weeks long and is completed at Hendon. 16 Figure 6. Police Now participants were asked how they found the 28-day DWO immersion period. The word cloud above presents the most commonly reported words. Life as a DWO Participants - The majority of participants (78%; n=35) reported being satisfied with their job as a DWO, 82% (n=37) reported being motivated at work, and 78% (n=35) feel as though they understand the needs of the communities they serve. I am satisfied with my current job Strongly disagree Neutral Strongly agree 0% 7% 16% 62% 16% 7% 78% - In their roles, 82% of participants (n=37) feel they are able to positively influence confidence in the police and 72% (n=32) feel they are encouraged to develop innovative solutions to problems. However, only 53% of participants (n=24) feel they are able to influence local crime patterns in their role. - Almost one third of participants (n=14) feel that serving as a police officer is not as exciting as they anticipated and the same proportion (31%) answered neutral. - Participants were asked in the interviews about the positive and challenging aspects of their roles. Participants mentioned many positive aspects to the DWO role including: getting to know the area and its residents in depth; having the freedom and independence to be responsible for planning their own day; the varied and diverse nature of the role; being able to make positive long-term changes and improve people’s lives; and being able to see projects through to conclusion. - The challenges of the DWO role mentioned by participants included: prioritising busy workloads; a lack of motivation and supervision from their team; the logistical frustrations of trying to solve problems and implement new strategies; complications caused by cumbersome MPS processes and systems; and managing public expectations. 17 - When participants were asked what factors motivate them in their role, the knowledge that they are improving people’s lives and ‘making a difference’, the varied and interesting nature of the DWO role, the dedication of their colleagues, and the daily interaction with diverse people were common factors. Is the DWO role the right choice for Police Now? Syndicate leads - Syndicate leads were positive that the DWO role was the right choice for the Police Now programme. They commented that the role gives the cohort the opportunity to make a difference and allows them to see any impact they are having directly on their community. They also thought the role gives the cohort a good grounding in the skills needed to be a leader; however, in reality, they noted participants will not want to stay in that role forever. Gives leadership and challenge that DWO role bring instead of getting 'lost in the numbers'. Dedicated to that ward gives you accountability, and you set your own goals. Participants - However, there were mixed views amongst the cohort. Some participants were concerned they are not able to experience ‘wider’ policing while working as a DWO, which they felt was negatively impacting on their overall development as a police officer. I feel like I’m not learning as much as I should; I feel like after the two years I won’t know as much as other police officers because I won’t have experienced these confrontational situations and the DWO role is much more slower paced. - Other participants were positive about the DWO role and thought it was the right choice for the Police Now scheme. Some mentioned the greater levels of responsibility that the DWO role brings, along with the ability to work on a more varied range of tasks and to carry out more community engagement. Line managers - Most line managers interviewed were positive about the DWO role; however, as mentioned earlier, many thought participants should have experience in the wider policing role. Indeed, some line managers have actively encouraged their officers to increase their knowledge and experience in other areas of policing in order to equip them to deal with the variety of situations they will encounter as an officer in the future. I sat down with her from the start and talked to her about her expectations, as well as mine and Borough’s and we decided she should immerse herself in many roles and portfolios; she’s not just been a DWO, she’s been given experience in other areas so she can face many scenarios. 18 Impact on the Community Participants - Participants were asked in the interviews to describe the biggest issues their ward is faced with, how they have attempted to gain an understanding of and address these issues, and how they have measured their success. - The biggest issues officers face on their ward include: ASB, drugs, motor vehicle crime, theft and violent crime. They have primarily gained an understanding of these challenges through a combination of talking with their team and local residents and organisations, analysing crime statistics, patrolling areas to witness crimes first-hand, and also through residents reporting crimes as they happen. - Officers have attempted to address these issues by: increasing patrols and police presence in affected areas, dispersal orders, arranging and executing drugs warrants, using intelligence and joint enforcement operations. - Most participants consider their tactics to have been successful in the main, which they reported measuring by looking at whether the number of calls from residents and businesses reduces, as well as identifying trends via mapping software, crime statistics and intelligence. However, there was no evidence of robust outcome based analysis or research and a few were slightly more reserved in their judgements, reasoning it will take time to implement and identify the impact of long-term projects. - One officer considered it “idealistic” to expect the Police Now recruits to make a significant impact in the space of only two years, and that in order to reduce crime on a larger, more long-term basis the Government needs to invest money in community initiatives; for example, the participant explained that “there is nothing for young people to do” in their local area and that without additional help and support in tackling the root cause of crime and ASB, there is only so much the police can do. - When asked about whether they have encountered issues around public confidence in the police, responses differed largely between individuals and wards. Those who had come across negativity or lack of trust explained that this was due to: persistent issues with no long-term solutions in place, which have led residents to question whether the police are ‘doing enough’; a ‘general dislike’ of the police in some wards; language and racial barriers between the public and police; and a perception that there is insufficient police presence in some areas. - When participants were asked how they try to build public confidence in their ward, most of their approaches are seemingly focused around increased communication and interaction with victims, hard-to-reach groups and general residents via various methods (face-to-face meetings, follow-up phone calls and social media). Line managers - Most line managers interviewed felt their officers’ proactivity has had a positive impact on their communities through a reduction in, for example, drug dealing and ASB. Examples of how officers have helped make a difference include: arranging and executing warrants; engaging with the public, hard-to-reach groups and local organisations; and using MPS software and intelligence. 19 - However, some others explained that after only six months as a DWO, their officers are still learning the role and it is therefore too soon for them to have made much of an impact on their community. Furthermore, several line managers described the difficulties involved in measuring Police Now officers’ community impact because there are no official methods of quantifying or determining public confidence that can be attributed back to a particular DWO. As such, the main methods used to evaluate individual officers’ influence are: feedback from colleagues, people in the community, social media and ward panel meetings; the number of complaints about certain issues; and community confidence and satisfaction surveys Stigma and Acceptance Participants - Of the cohort, 55% (n=25) reported feeling a stigma attached to being part of the Police Now programme and 49% (n=22) reported feeling uncomfortable disclosing being part of the programme when meeting new colleagues. There is a stigma attached to being part of the Police Now programme Strongly disagree Neutral Strongly agree 0% 31% 13% 33% 22% 31% 55% When meeting new colleagues I feel comfortable disclosing that I am part of the Police Now programme 7% 25% 32% 27% 9% 32% 36% - Interviews with Police Now participants indicated they thought a lack of awareness and incorrect assumptions about the Police Now scheme was the cause of the stigma they had experienced, though they also explained there are more general prejudices towards MPS graduate schemes. I think it’s because people aren’t fully aware of what it is; they think it’s a high potential scheme and we’ll be promoted in six months. - Those who had not felt stigma suggested either that their team knew nothing about the Police Now scheme, which meant they had been treated in the same way as other probationers, or that their colleagues were more open-minded and accepting of them. - Despite participants feeling some stigma associated with being part of Police Now, the majority of participants (73%; n=33) feel they have been accepted by colleagues on their borough. In the interviews, a few participants explained they had been prepared for negative reactions but were surprised how welcoming and helpful their team has been. - Furthermore, the majority of participants (71%; n=32) feel they are treated with respect by colleagues on their borough. However, this finding suggests that nearly one third of participants do not agree they are treated with respect by colleagues. 20 I feel I have been accepted by colleagues on my borough Strongly disagree Neutral Strongly agree 0% 11% 16% 44% 29% 11% 73% - When participants were asked whether the way they are treated differs to officers not on the Police Now scheme, only 5 of the 22 officers interviewed agreed. They reported that managers and superiors can have higher expectations of them and sometimes offer them more opportunities and responsibility than others. There’s a lot of expectation; a lot of people want the scheme to succeed and are watching us closely just in case it doesn’t, so there is a bit more pressure. Syndicate leads - Interviews with syndicate leads indicated they agreed there is often a burden of expectation and resentment towards those on any positive action/high potential scheme like Police Now. They stated the stigmas are usually attached to the graduates themselves and not just the schemes. - Syndicate leads believe a lot of the stigma Police Now participants will face will be as a result of misconceptions about the scheme; for example, that it is an accelerated promotion scheme. Once people understand what the scheme is (and what is not), syndicate leads believe participants will have no problems being accepted. - Respondents also talked about the need for Police Now participants to be proactive in challenging stigma. Stigma is unavoidable, they will get labelled… Overall, it will make a difference in the first hour. If they excel and show people around them they are competent, then that will be it. - According to the syndicate leads, the syndicate structure was designed to increase resilience towards stigmatisation and enable the participants to better cope with the pressures of their new role. Consistent with this idea, fewer Police Now participants (although still over half) reported feeling a stigma compared to the 2013 Police Officer Graduate Entry Programme cohort: 73% of this cohort reported feeling a stigma attached to being part of a graduate scheme (see Figure 7). This finding suggests the syndicate structure—or some other feature of the scheme—may be protecting Police Now participants from experiencing a higher level of stigmatisation. Although these findings are promising, there are still many challenges to overcome in relation to the integration of historically minority officer groups (i.e., black and minority ethnic officers, females and university educated officers). 21 Figure 7. There is a stigma attached to being part of the [Graduate/Police Now] programme Line managers - Line managers described mixed reactions from within their teams to a Police Now recruit joining them. They described some initial scepticism due to: a lack of awareness of the difference between Police Now and other similar schemes; current and established DWOs either losing their role to Police Now officers or being required to share a previous standalone job; a perception of too many probationers being introduced into the MPS at the expense of more experienced staff; and doubt as to whether having a university degree would make a difference to someone’s policing skills. The general feeling was ‘oh here we go; another of these schemes’. That was definitely the most general view, people thinking they’re going to be leap-frogged. I’ve had to point out to people that it isn’t a fast-track scheme. - Positively, many of the initial negative reactions from existing team members were quickly overcome once the Police Now officers arrived on their wards. Furthermore, almost all line managers felt that their Police Now officer is integrating well into their team, describing them as “popular” and “well-liked”. - Although several line managers did not perceive any differences in the way Police Now officers are treated compared to others within the MPS, some felt that management has higher expectations of them, and that they can have a difficult time living up to these. - More generally, there was a sense that Police Now officers are treated more favourably than those on similar graduate schemes, mainly due to the programme not having a ‘fast track’ element to it. Experience of the Organisation Their team Participants - Police Now participants were asked to describe their team in three words. The most common words reported were “supportive” and “dedicated”; however, “unmotivated” was also notable for its relative frequency. - The majority of participants thought the work of their team is distributed fairly with many praising both their line manager and colleagues for their support and work ethic. 0 10 20 30 40 50 60 70 80 2013 Grad Cohort Police Now cohort % agree/strongly agree 22 - However, some explained that while they are generally satisfied with how work is allocated, there are some issues in terms of workload sharing within their team. Some participants reported that: work is not always proactively delegated and responsibility falls on individuals to volunteer for tasks; being allocated general policing tasks results in DWOs having to balance these with their community role; DWOs are allocated more administrative tasks than others on their teams; DWOs have more responsibilities and therefore naturally have larger workloads. - Discouragingly, only 46% (n=21) feel they are rewarded fairly for the work they do and only 56% (n=25) feel there is a positive atmosphere among the people they work with. I am rewarded fairly for the work that I do Strongly disagree Neutral Strongly agree 2% 9% 42% 44% 2% 11% 46% There is a positive atmosphere among the people I work with 7% 20% 18% 47% 9% 27% 56% Resources and equipment - Furthermore, only 38% (n=17) reported having access to the resources and equipment needed to do their job. When further questioned about resources and equipment in the interviews, participants explained there is at least one resource or piece of equipment they require to do their job that they currently do not have access to. Sufficient IT equipment and police cars were the main reported issues, as well as ward mobiles, harnesses, body-worn video cameras, crime scene cameras, SmartWater, iPads, and mobile fingerprinting devices Leadership and support Participants - Of the cohort, 65% (n=29) stated they had confidence in their own local senior leadership team, two thirds of participants (n=30) agreed that the values of the MPS are reflected in the behaviour of their local senior leadership team and 76% (n=34) agreed that managers on their borough treat all staff fairly, regardless of their background. - Furthermore, 73% (n=33) reported being satisfied with the level of leadership and support they receive from their line manager, yet only 44% (n=20) reported receiving regular and constructive feedback from their line manager on their performance. - When interviewed, most participants were very positive and complimentary about the support and training they have received from their line manager: they generally considered them approachable, helpful and encouraging, as well as being proactive in arranging support and training. 23 I have a lot of ongoing support, day-to-day support; I have a very hands-on line manager who made the transition into the real aspects of the job very easy. My line manager has asked me what I want to do going forward and things have been put in place to help me with my development. - Conversely, some participants felt the training and support provided by their line manager has lacked because of: a lack of actively-provided training and development opportunities; the unavailability of line managers due to workloads, differing shift patterns, long-term sickness and holidays; and line managers’ lack of awareness of the Police Now scheme and additional tasks. It seems to be so much led by yourself, which is a bit poor really considering you’re new and you’ve no idea what sort of training you should go on. I get enough day-to-day support, but training and long-term development not so much. - Nearly all participants felt they would benefit from more on-going training to improve their policing skills, including: more practical experience such as response and street training, training in-line with that received by other probationers such as Hydra, additional in-depth DWO training, and training on MPS-specific systems and processes. - In terms of the training and support offered by Police Now since the Summer Academy ended, officers were typically positive about the ongoing support they have received: many are still in contact with their syndicate lead and regularly use the WhatsApp and Facebook groups to seek advice and assistance. They also considered the Police Now team to be very responsive when contacted – although it was said that the leadership team’s primary focus is now the second cohort of recruits, which has left some participants feeling neglected in terms of the level of support and development provided. Line managers - From a line manager perspective, the majority had no involvement with Police Now prior to being tasked with becoming a line manager, but the majority said they felt at least partially prepared to act as a line manager to a Police Now officer, mainly due to their previous experience in line management. However, 9 of the 25 line managers said they did not feel prepared to act as a line manager to their Police Now officer, mainly because they did not receive enough (or relevant) information about the scheme. I don’t think we got a particularly great deal of input about the programme. We had to read up on it and make our own judgements, which is not always a good thing if I’m honest. There were no seminars or presentations about the programme. - Overall it was felt that more face-to-face and interactive meetings would have been helpful in developing line managers’ understanding of the Police Now scheme and how to support their officer. Comments from Police Now suggested line managers and boroughs were offered opportunities to gain further information about the scheme through a number of briefings, including face-to-face meetings, but uptake was low. - The majority of line managers reported being in regular contact with their Police Now officers via informal daily conversations rather than official meetings. Most considered this to be the most appropriate approach, although one line manager commented they should have ‘proper’ meetings more often, but talked about the difficulties in 24 scheduling these around busy workloads. Other line managers reported sitting down ‘properly’ with their officer around once a week and a similar proportion meet on a monthly basis. Wellbeing Participants - Only 51% (n=23) of participants in TS2 stated they achieve a good balance between their work and private life. Wellbeing was further explored in the interviews with participants. Participants were asked to rate how often they experienced a number of positive and negative emotions. Of the positive emotions listed, “happiness” was experienced most often and “feeling like you are making a difference” least often. Of the more negative emotions, “fear” had been experienced least, while “thinking about work outside of work hours” was rated highest most often. It should also be noted that 9 of the 22 officers interviewed said they had felt stressed at work in the last week. - Participants’ main coping strategies for dealing with stress include: healthy eating and exercise; spending time with family and friends; and talking to colleagues. Encouragingly, most felt they have access to enough emotional support from their colleagues, friends and family; however, a few would like the MPS to be more proactive in urging people to seek support if they require it. Job Satisfaction and Career Aspirations Participants - According to TS2, 75% of participants (n=34) reported being satisfied that their chosen career as a police officer is right for them. Significantly fewer female participants (70%; n=14) reported job satisfaction compared to the male participants (84%; n=21)13. - However, only 33% of participants (n=15) reported seeing themselves as spending their whole career in the MPS: a significant decrease from TS1 where almost two-thirds of participants saw themselves as spending their whole career in the MPS14. At this stage, how satisfied are you that your chosen career as a police officer is right for you? Strongly disagree Neutral Strongly agree 0% 4% 20% 44% 31% 4% 75% I see myself spending my whole career in the MPS 4% 20% 42% 22% 11% 24% 33% - When the cohort was asked about their aspirations for their future career as a police officer, 42% (n=19) stated they wanted to hold a specialist role (e.g., protection, 13 According to an independent samples t-test, t(43) = 2.43, p = .019, 95% CI [.10, 1.06], d = .74. 14 According to a paired samples t-test, t(43) = -4.45, p < .001, 95% CI [-.82, -.30], d = .82. 25 firearms, dog handler), 31% (n=14) a detective role, and 11% (n=5) aspired to work in a senior leadership role (Superintendent or above). - Interestingly, in contrast to Tracker Survey 1 (TS1), fewer participants in TS2 aspired to work in a senior leadership role (a 56% decrease) and more participants aspired to hold a specialist role (a 50% increase). - Furthermore, more participants in TS2 stated it was their intention to complete the two year Police Now programme and move on to a career outside of policing (6% to 9%). Attitudes toward the MPS and Policing The MPS and Policing in London - Participants’ attitudes towards the MPS were mostly positive: 75% (n=34) agreed the MPS is a good organisation to work for, 80% (n=36) felt a sense of loyalty to the MPS, 93% (n=42) stated they are proud to work for the MPS, 64% of participants (n=29) stated it is clear to them what the organisational goals of the MPS are, and 73% (n=33) stated they had confidence in the leadership provided by senior leaders in the MPS. - Participants were asked for their views on the key skills and personality characteristics required to be a police officer in London. “Communication”, “empathy”, “positivity”, “resilience”, “perseverance” and “calmness” were considered the most important skills. - The majority of participants said that most of the skills they listed are the same as those required outside London. However, “determination”, “open-mindedness”, “resilience” and “patience” were considered to be especially important for policing in the MPS due to the city’s size, crime levels and diversity. - Most participants (19 of the 22 participants interviewed) said that they do see their listed key skills within policing in London at present; they have mainly observed these being used by their colleagues. However, some officers explained that despite seeing these key skills within policing on the whole, they do not apply to everyone across the MPS. It depends on the police officer. Some are helpful, happy to take any kind of call, whereas others have been in the service for years and are counting down the days to retirement and don’t want to take on anything an hour before their shift ends. - Furthermore, a minority of participants specified some key skills that are apparently lacking within the MPS, namely diversity, positivity, helpfulness and communication. Policing and serving the public - Existing longitudinal research tracking the job attitudes of police recruits over time— including the research conducted by Evidence and Insight into the 2013 Graduate Entry Programme15—has shown that officers’ attitudes toward policing tend to decline the longer they are exposed to the police working environment.vi Consistent with this research, there has been a decline in some attitudes since the Police Now cohort finished their training and began their roles as DWOs. 15 The most recent findings from this research are presented in Report 1. 26 - Compared to TS1, more participants (7% - 18%) in TS2 agreed that customer service should not be a big part of a police officer’s job and fewer participants (94% - 87%) agreed that officers should treat everyone with the same level of respect regardless of how they behave towards the police.16 - Interestingly, since beginning their roles as police officers, participants reported being less confident (75% - 65%) of receiving a good service if contacting the MPS as a member of the public.17 - Fewer participants (88% - 78%) agreed in TS2 that it is important the MPS workforce ‘looks and feels’ like the diverse community that it services.18 - However, some attitudes toward policing and the public remained unchanged. For example, almost all participants in both surveys disagreed with the statements: “Good police-community relations has no effect on the crime-rate”, “Neighbourhood policing is not ‘real’ policing”, and “Some victims of crime are more deserving of a good service than others” (see Table 1).19 - Despite some declines in attitudes, the attitudes of Police Now participants are still far more positive than MPS staff overall. For example, whereas 65% of Police Now participants agreed in TS2 that they would be confident of receiving a good service if contacting the MPS as a member of the public, only 28% of all MPS staff agreed with this statement.vii In addition, in the most recent survey with the 2013 Graduate Entry Programme cohort, only 45% agreed with this statement after 2 years on the job. - However, when isolating MPS staff with similar levels of service to the Police Now participants (i.e., less than 6 months), a greater proportion of MPS staff (78%) were confident of receiving a good service compared to the Police Now participants (65%). Table 1. Attitudes toward policing and serving the public Question Tracker Survey 1 (n=69) Tracker Survey 2 (n=45) % change Being concerned about customer service should not be a big part of a police officer’s job 7.2% 17.7% 10.5% Officers should treat everyone with the same level of respect regardless of how they behave towards the police 94.2% 86.6% -7.6% If I contact the MPS as a member of the public, I would be confident of receiving a good service 75.4% 64.5% -10.9% It is important that the MPS workforce ‘looks and feels’ like the diverse communities that it serves 88.4% 77.7% -10.7% Good police-community relations has no effect on the crime rate 2.9% 4.4% 1.5% Neighbourhood policing is not ‘real’ policing 2.8% 0% -2.8% Some victims of crime are more deserving of a good service than others 2.9% 2.2% -0.7% There is no point in officers trying to convince some communities in London that the police can be trusted 0% 2.2% 2.2% Note: figures represent % strongly agree/agree 16 Neither difference was significant: t(44)= 1.96, p =.057, 95% CI = -.00, .27, d = .59; t(44)= -1.43, p =.160, 95% CI= -.21, .04, d = .43. 17 The difference was non-significant: t(44)= -1.27, p =.209, 95% CI= -.23, .05, d = .38. 18 The difference was nearing significance: t(44) = -1.95, p = .058, 95% CI = -.23, .00, d = .59 19 All differences were non-significant. 27 Overall thoughts on Police Now Participants - Overall, findings from the survey and interviews suggest participants were very satisfied with the Police Now scheme: 95% of participants would encourage other graduates to join the scheme, 82% talk positively about Police Now when they speak to other people, 82% stated they feel a sense of loyalty to Police Now, and 89% feel they have a strong bond with their fellow Police Now colleagues20. - Furthermore, 86% of participants stated they understand what the goals of Police Now are and all participants agreed the Police Now team are working hard to make the programme a success. - These findings are in contrast to survey findings from the Police Officer Graduate Entry Programme in 2013. Of these recruits, 50% were dissatisfied with the quality of training received during their time with the MPS, and only 59% would encourage other graduates to join the scheme. A lack of communication, poor information provision and an alleged failure to meet programme expectations were commonly cited concerns of this cohort. - The vast majority of participants considered Police Now to be the most effective method of recruiting graduate staff, mainly because they felt it: is well-structured; attracts high calibre people who would not otherwise have considered a career in policing; is more acceptable to other officers than schemes such as Fast Track; provides opportunities for development not available to other probationers; and offers recruits an excellent role. A large part of its effectiveness is how structured it is, and how the ongoing focus is on the cohort. The usual route to joining the police involves a large volume of intakes which makes it hard to monitor people’s performance, whereas our programme means this is a possibility. We very much have access to the tools we need to progress our career. - Some others, although complimentary about the scheme in general, did not consider it to be the most effective method of recruiting graduate staff, mainly because they did not feel being a DWO has given them sufficient experience in wider policing and has, for this reason, hindered their development. I think it is an effective way but think they need to make some changes; I don’t think being in this neighbourhood role for two years is the most effective thing; there are so many things in policing and only experiencing one part is not enough. - In terms of whether or not policing will benefit from Police Now in the long-term, most participants felt it would insofar as the scheme attracts people who bring enthusiasm, innovative thinking, proactivity and the right attitude to the DWO role. Syndicate leads - Syndicate leads were also highly positive about the scheme overall. They reported the Police Now scheme challenges the well-grounded assumptions that policing is a ‘job for life’ by giving participants the option and flexibility to leave after two years. Because of 20 Significantly more female participants reported having a strong bond with their fellow Police Now colleagues compared to male participants, t(43) = -2.36, p = .023, 95% CI [-.99, -.08], d = .72; and significantly fewer females would encourage other graduates to join the scheme, t(43) = 2.30, p = .026, 95% CI [.06, .89], d = .70. 28 this flexibility, Police Now attracts people who have never considered a career in policing before, who are intelligent and who have energy and new ideas. Syndicate leads and the operations team reported that these aspects mean the Police Now scheme has the potential to change policing culture and create a more diverse workforce. Police Now is a good start at challenging people's views. People need to get the idea that it isn't necessarily a 35 year profession. Line managers - Many line managers were also positive about the scheme overall. Those with a good overall opinion of Police Now felt this way mainly because of their positive experience with their Police Now officer, while others praised the training, support and career the programme offers, as well as the type of people being recruited. I would resign and re-join as a Police Now officer if I could do it all again! The training is amazing and the backing and support they receive from the senior leadership team, the possibilities once the two years are up to move laterally or upwards is fantastic. - Furthermore, line managers’ experiences of line managing their Police Now officer has been overwhelmingly positive, with the majority commenting on how impressed they are with their: enthusiasm, keenness to learn, proactive attitude, innovative thinking, intelligence and commitment. They variously described their experience as “thoroughly enjoyable”, “an amazing experience”, “an absolute delight” and “a positive experience”. - However, the most frequent criticism of the scheme was around the two-year timescale of the programme and the fact that officers are free to leave the Service after this time. This, it was felt, could lead to the loss of talented, promising members of staff (and would represent a waste of the resources taken to train them). I am concerned it’s only a two-year scheme; seems a shame to put time and effort into them for them to leave. I know they can go out into the community afterwards but it takes so much input and people are exhausted as it is without more staff to train. - In terms of whether policing will benefit from the Police Now scheme in the long-term, many of the 25 line managers thought it would because it: represents improved recruitment and training of the ‘right’ people; strives to bring continuity to the DWO role; and provides an important focus on the community side of policing. Furthermore, the positive experiences most have had with their Police Now officer have made them feel very hopeful for the future of the programme. - However, a significant minority of participants were uncertain about the long-term benefits of the scheme, with several again reiterating that its two year timescale could lead to the loss of talented police officers from the MPS. - Finally, a number of line managers were concerned that participants’ expectations of their role and what they could achieve have not matched the realities of the organisation. It was felt that this disconnection between expectation and reality could impact on their experience of policing and result in many of them leaving the Service. 29 4. Discussion Summary and Conclusion In summary, the experiences of the first cohort of Police Now participants suggest the Summer Academy training was well-received. The enthusiasm and energy of the trainers, the quality of the material presented, and the strong bond created among participants were common themes throughout the surveys and interviews. In addition, there was general positivity from participants, syndicate leads and line managers around the concept of Police Now and most respondents thought policing would benefit from the scheme in the long-term, most notably because it is attracting the ‘right’ people to the role. The findings presented in this report suggest some recommendations for future roll-outs of Police Now. Firstly, to avoid stigma and misconceptions, clarity around the scheme is needed. In particular, it should distance itself from being labelled or assumed to be a ‘high potential/fast track’ scheme. Externally, syndicate leads and the operations team were positive about the marketing of the scheme and were confident they attracted the right graduates to the programme. Internally, more information and clarity is needed for line managers and the wider team receiving Police Now officers. In particular, some participants reported that their team did not know about the scheme or what the immersion period was meant to achieve, and some participants reported feeling like they had to convince their colleagues they were competent. Line managers agreed that more information about the scheme would have helped them better prepare for their officer’s arrival. They also suggested it would have helped reduce stigma and increased their team’s initial acceptance of the officer. Another area for improvement is ensuring consistency in participants’ opportunities to learn during the 28 day immersion period. Participants who had a mentor who was unavailable or disinterested in mentoring them reported that their learning was impeded. More focus should be placed on ensuring DWO mentors have adequate training and are motivated to mentor their Police Now participant. Thirdly, although both participants and syndicate leads reported many benefits to the intensity of the Summer Academy, the length and timing of the training was also mentioned as an area for improvement. Specifically, participants thought that more time could be spent on field- training and the Summer Academy could be lengthened to make time for this training. It is possible that a restructure of the Summer Academy could alleviate some of these concerns, while keeping the six week intensity. For example, allocating full days for participants to spend on borough rather than splitting the days between classroom and field training could provide the cohort with the variety of field-training experience they thought was lacking in the Summer Academy. Syndicate leads also reported timing concerns. In particular, they thought more time was needed to prepare for the delivery of the programme. Whether these concerns are a feature of the first year roll-out is unclear, but clearly ensuring the people delivering the training are adequately prepared is essential. Interviews with syndicate leads and comments from 30 participants also suggested a review of certain topics should be undertaken. For example, syndicate leads suggested more time could be spent on role play exercises to boost confidence. Participants also reported the training could have spent more time on practical skills/role-plays and less time on other areas, such as soft skills Line managers also thought participants could have benefitted from more training in wider policing skills prior to starting their roles as DWOs. A notable finding from the interviews with participants was that many reported having to get involved in wider policing duties (because of heavy workloads within their team), despite it not being part of the DWO role. This finding suggests the need to ensure the officers are adequately trained to respond to the varied tasks they may be expected to do as part of a team. Further field training and experience of the wider policing role could also help to increase feelings of confidence and preparedness amongst the cohort. The low levels of confidence were a notable finding from the tracker survey, particularly for female officers whose confidence was significantly lower than male officers. Despite these suggestions for areas of improvement, overall the findings suggest the Police Now scheme is attracting to policing people who are proactive, enthusiastic and who have the potential to make a difference to the organisation and to the field of policing. However, it is important to ensure participants don’t lose their enthusiasm, especially when they are faced with a working environment they describe as lacking in resources and low in morale. The survey findings showed that Police Now participants experienced a decline in attitudes after only a few months on the job. The challenge moving forward will be to ensure the first cohort of Police Now officers is well supported as they continue their journey as police officers. Next Steps in Evaluation The next stage of the Police Now evaluation will continue to follow the first Police Now cohort as they continue their two year DWO placements. Tracker surveys will monitor participants’ attitudes and whether they change over time, and interviews will further explore emerging themes. Interviews and surveys with comparison groups—including non-Police Now DWOs and other MPS recruits—will be undertaken to explore the differences and similarities in Police Now participants’ experiences compared to officers who joined the police force through other means. One more interim report and the final evaluation report are scheduled for publication. The details and timescales for these reports are summarised below. Report 3: Interim Progress/Findings 2 – December 2016 - Update on the progress of the evaluation to this point - Mini-literature review – previous longitudinal studies of police recruits and the history of graduate recruitment within the police - Present the emerging findings from the tracker surveys and interviews that will have been completed with the participants and comparison groups to this point, and thus begin to address the following questions: How are the participants’ job experiences and attitudes changing over time? How attitudinally similar are the Police Now participants to non-Police Now officers? 31 - Update on Police Now 2015/16 and the induction of the second cohort of participants. - Overview of immediate next steps and timescales for remainder of evaluation. Report 4: Comprehensive Evaluation Report – October 2017 - Background and Context (both in terms of Police Now, the changing British policing landscape, and the current challenges facing police officers’ in London) - Evaluation Design - Literature Review - Findings from the interviews conducted with Police Now practitioners, and the line managers and mentors of Police Now participants with regard to programme implementation, delivery and the occupational socialisation of the recruits - ‘About’ the Police Now participants (their journey from assessment centre to graduation) - Findings from the data obtained through the tracker surveys/reflection journals/focus groups and interviews with the Police Now participants and the identified comparison groups: How did the participants’ job experiences and attitudes changing over the course of the two-year programme? How attitudinally similar are the Police Now participants to non-Police Now officers? - Case studies of Police Now participants’ evidence-based policing assignments, and any measurable impact they might have had on crime and confidence at a local level. - Findings from any exit interviews completed. - Findings from survey/interviews conducted with the ‘Graduate Partner’ organisations. - Discussion and recommendations. - Limitations with the study and areas for further research. 32 References i Police Now: A longitudinal evaluation of the job attitudes and experiences of ‘Police Now’ participants. Introductory Report (October 2015). Evidence and Insight, MOPAC. ii Police Now: The Case for Change (2014) http://www.policenow.org.uk/downloads/case-for-change.pdf iii Build a Better Met: MPS Staff Survey 2015 iv Fenn, L. (April 2014). Expectations MET? Evidence and Insight, MOPAC. v Neyroud, P. (2011). Review of police leadership and training, two volumes. London: Home Office. vi Fenn, L. (April 2014). Expectations MET? Evidence and Insight, MOPAC. See also Fielding, N.G. (1988). Joining forces: Police training, socialization, and occupational competence. New York, NY: Routledge and Van Maanen, J. (1975). Police socialization: a longitudinal examination of job attitudes in an urban police department. Administrative Science Quarterly, 20, 207-228. vii Build a Better Met: MPS Staff Survey 2015 1 Research into the London Mental Health Street Triage Pilot December 2015 Zoe Hobson, Lynne Grossmith & Paul Dawson MOPAC Evidence and Insight 2 Contents Executive Summary .................................................................................................................. 3 1. Introduction ......................................................................................................................... 4 Policing and Mental Health ..................................................................................................... 4 Pressures on Police Practice .................................................................................................... 5 Reviews on Mental Health ....................................................................................................... 5 National Street Triage Pilot ..................................................................................................... 6 London Street Triage Pilot....................................................................................................... 7 Evaluation of London Street Triage ......................................................................................... 8 2. Methodology ........................................................................................................................ 9 Exploring the Impact ............................................................................................................... 9 Exploring the Process .............................................................................................................. 9 3. Results ................................................................................................................................ 11 Process Findings: Implementing London’s Street Triage ........................................................ 11 Street Triage Impacts and Outcomes ..................................................................................... 15 Wider Learning from Street Triage ......................................................................................... 19 4. Discussion ........................................................................................................................... 24 Sustainability ......................................................................................................................... 24 Training ................................................................................................................................. 25 Promotion of the Service ....................................................................................................... 26 Conclusions ........................................................................................................................... 26 Appendices ............................................................................................................................. 27 Appendix A ............................................................................................................................ 27 Appendix B ............................................................................................................................ 29 Appendix C ............................................................................................................................ 32 References .............................................................................................................................. 33 3 Executive Summary The London Street Triage pilot is one of nine national pilots funded by the Department of Health. Street Triage aims to enable mental health professionals to work alongside the police, providing officers with ‘real-time’ information and advice to ensure people who need mental health support receive it as quickly as possible. The pilot ran in four south London boroughs (Croydon, Lambeth, Lewisham and Southwark) and provided a 24 hour telephone advice/helpline staffed by mental health professionals. Additionally, in all boroughs when clinically required, a face-to-face assessment service was trialled, consisting of mental health professionals attending certain incidents requested by an officer, to conduct an assessment at the scene alongside police officers. Evidence & Insight (E & I), the research capacity of the Mayor’s Office for Policing And Crime (MOPAC), were tasked with conducting research and evaluating Street Triage, running concurrently with the national evaluation. The aims were broadly to explore the processes and potential impact of a mental health Street Triage service, through the interrogation of basic performance data such as usage of the service and examination of the views of those using and experiencing the service. A range of research methods were drawn upon to achieve this (e.g., performance data, police officer surveys, and service user and carer focus groups). Key Findings  It is estimated that one quarter (26%) of occasions where officers encountered individuals with mental health difficulties during their duty resulted in a call to the Helpline. This is a total of 1,179 calls within 12 months. This is a solid start to the helpline, although improved communication and promotion may encourage use.  Between April 2013 and March 2015 London as a whole has seen increases in both the number of criminal incidents involving mental health aspects (64% increase) and the number of vulnerability reports recorded (31% increase) as well as a reduction in section 136 detainees taken to police custody as a Place of Safety (71% decrease). The Street Triage boroughs are consistent with these trends, therefore it is not possible to attribute any changes observed to Street Triage.  Officers provided positive views around Street Triage, appreciating the timely and professional advice and felt that it improved their confidence when working with vulnerabilities including mental health.  Learning was obtained in terms of the implementation of the service – with staff highlighting challenges such as training, limited pathways (such as a lack of designated specialist places of safety) and partnership working. These issues could be ‘designed in’ to any subsequent roll-out of the service to enhance delivery.  Poor communication between the police and mental health professionals emerged as a key theme. This was perceived as a general barrier to effective everyday working and could result in misperceptions between partners (e.g., that officers are using custody as an alternative more than they actually are).  Feedback from officers, mental health professionals, service users and carers indicated mental health training should include human engagement from service users so officers understand how their reactions/body language can affect an individual in crisis. For Street Triage to be sustainable in the future, the focus moving forwards needs to be on training, identifying appropriate pathways for individuals, working with partners to help effective communication of the services provided and integrating this service into current commissioning plans. 4 1. Introduction Policing and Mental Health There has been a long and disjointed history in terms of how organisations respond to people with mental health needs. From a national perspective, there is often little standardisation across services or geographical boundaries, which can make it problematic when trying to provide appropriate care pathways. A number of recent reviews highlight the inadequacies of joint working by the police and health care partners (NHS, London Ambulance Service), particularly around how to effectively respond to individuals living with mental health needs.i The prevalence of individuals experiencing mental health needs in the UK further strengthens the need to prioritise these discussions. It is believed in Britain around one in four people will experience mental health issues each year.ii Within London more than one million residents live with mental health needs, a higher prevalence than in other counties across the UK.iii As a result, the MPS routinely come in to contact with members of public, victims of crime and offenders who are experiencing mental health needs.iv Between April 2014 and March 2015, the MPS dealt with 79,811 emergency (999) or non-emergency (101) calls receiving a mental health qualifying codev and a recent report by the College of Policingvi suggests a typical officer, in a typical force, might deal with an average of 14 incidents per day that are flagged as linked to mental health issues. For those individuals experiencing mental health crisis, the police can act as gatekeepers, with officers regularly facing the tough challenge of having to identify mental health need and responding in an appropriate manner. This raises questions around the very nature of policing, responsibilities and partnership working – a theme that runs through the report. Police intervention may play a significant role in preventing the impact of deteriorating mental health, making the interface between crisis services and emergency services vital to manage and support individuals in need.vii London is not without its own misalignment between services, with a number of different NHS Trusts working in close conjunction, and multiple police officers working largely on their specific boroughs. London also contains 33 Clinical Commissioning Groups (CCGs), each running independently with no London-wide network to communicate and share information, knowledge and good practice. Therefore, London boroughs receive different funding for a range of interventions, with little calibration, creating a very confusing landscape to navigate. This is evidenced on a number of occasions including when officers invoke Section 136 of the Mental Health Act (S136 MHA; 1984) and are required to take the individual to a Place of Safety (POS), and when officers choose not to rely on S136 MHA and take service users voluntarily to Emergency Departments, with some departments accepting mental health patients, and others not. Borough disjoins are demonstrated through the difficulties police officers and NHS staff have in accessing POS (namely 136 suites) within other NHS Trusts when there is no immediate available spaces within their own Trusts.viii New ambitions are being discussed around co-commissioning joint POS to alleviate these issues. 5 Pressures on Police Practice Learning from the Independent Commissionix into how the MPS responds to policing incidents involving mental health highlights the pressures that the police are under. Previous in-depth analysis of police performance data from the MOPAC E&I teamx, demonstrate the complex and demanding aspect of policing mental health related incidents, particularly in terms of:  The individual's journey through the organisation – there is a heavy and rising demand of individuals in contact with the police when in mental health crisis.  Places of Safety - the use of custody as a POS was seen, but often remains inappropriate.  Training – officers continue to feel training is inadequate around identification, needs and management of mental health.  Partnership Working – there are gaps in terms of appropriate access to mental health partners.  Audit and Review – there is a lack of rigorous data collection and capture around mental health needs.xi In addition to the demand and overlap between policing and mental health – the MPS are under considerable pressure to deliver on effectiveness, efficiency, professionalism and to improve public confidence in the police.xii In a time of austerity, implementation and progression of schemes are under threat – highlighting the need for cost effective, partnership and innovative solutions.1 Reviews of Mental Health Over the past years there have been a number of reviews and reports exploring mental health provision with a view to improve treatment and pathways across different sectors. Reviews have detailed how partnership agencies should respond to individuals, with the 2009 Bradley Reportxiii making 82 recommendations for improving the treatment of people with mental health needs and learning disabilities in the criminal justice system. A later independent review, The Bradley Report Five Years Onxiv suggested significant progress had been made over the five years in key aspects, such as the introduction of the Liaison and Diversion (L&D) services.2 In 2013 the Independent Commission on Mental Health and Policing detailed a further 28 recommendations for change, falling under three broad areas for action including leadership, on the frontline and interagency working. In response to these recommendations, and the recognition of both the shortfalls in ‘health’ and that all agencies should engage in collaborative working, the Crisis Care Concordatxv was introduced in 2014. This provided a 1 This is particularly relevant given the potential changes occurring around S75 of the National Health Service Act 2016 between London Local Authorities and Mental Health Trusts. 2 L&D services exist to identify offenders who have mental health, learning disability or substance misuse vulnerabilities on their first contact with the criminal justice system. This way they can be supported through the criminal justice system pathway, and if necessary, diverted into treatment, social care services or other relevant intervention or support services. L&D services aim to improve health outcomes, reduce re-offending and identify vulnerabilities earlier, thus reducing the likelihood offenders will reach crisis-point. 6 national agreement between services and agencies involved in the care and support of people in crisis. The Concordat sets out how organisations will work together better to make sure that people get the help they need when in crisis. Twenty two national bodies including health, policing, social care, housing, local government and the third sector have now signed up to the Concordat, with the aim of focusing on:  Access to support before crisis point through availability of 24 hour services  Urgent and emergency access to crisis care  Quality of treatment and care when in crisis  Recovery and staying well through referrals to appropriate services National Street Triage Pilot The introduction of Street Triage was, in part, to response to elements of the action plans set out through the Concordat, particularly addressing the provision of 24 hour access to professional services and encouraging collaborative working relationships between partner agencies to achieve such goals. As a national pilot, Street Triage attempted to embrace learning from previous reviews providing a comprehensive triage service to individuals coming into contact with the police for reasons relating to mental health needs. This approach intended to provide a more timely intervention from mental health professionals and reduce unnecessary detentions for people under S136 MHA. A key objective was to ensure individuals detained under S136 were not taken to police custody as a place of safety. It was thought the implementation of such a service may also achieve cost saving for police services. The aim of Street Triage was to enable mental health professionals to provide ‘real time’ advice to police officers 24 hours a day, 7 days a week. Advice included a professional opinion on a person’s condition, or appropriate information sharing about a person’s health history. The aim was to assist police officers to make appropriate decisions, based on a clear understanding of the background to these situations. A wider aspect of the Street Triage service was to encourage individuals into better care pathways and referrals to appropriate services through the use of seven day follow ups. Funded for one year through the Department of Health, nationally there were nine Mental Health Street Triage pilot projects. Each participating police force had a local variation of the operational model, predominately deploying NHS mental health nurses in police response cars to assist with mental ill health on the street.3 For example, the British Transport Police and NHS London conducted joint assessments of all cases over the preceding 24 hours and formulated joint plans to reduce the risk of harm and engage relevant care pathways. Differing in their approach, Leicestershire operationalised a Street Triage Car, jointly operated by Leicestershire 3 Many of the pilot areas have commenced their own evaluation of the pilot scheme, each adopting different approaches and methodologies. Additionally, the Department of Health have employed University College London to conduct a national evaluation of all the pilot schemes, appraising the different operating models in comparison to each other. London Street Triage pilot have already supplied information to this evaluation. 7 Police and Leicestershire NHS Partnership Trust. The car was driven by a police officer and contained a mental health nurse from the crisis service. The car provided an initial point of contact for officers first at the scene who encounter incidents with a mental health element, before exercising their police powers. London Street Triage Pilot Commissioned by the Mayor’s Office for Policing And Crime (MOPAC) and NHS England Health and Justice Team, the London Street Triage was a collaborative project between South London and Maudsley NHS Trust (SLaM) and the Metropolitan Police Service (MPS). SLaM Foundation Trust was chosen to participate within the pilot as they recorded the highest uses of S136 MHA for each year (from 2009 – 2012) in comparison to all other London NHS Trusts. The pilot commenced on 31st March 2014 offering a slightly different operating model to many of the other UK pilots. Due to the size of London, limited resources, the number of officers and volume of mental health related incidents, it was decided primarily a helpline and limited face- to-face services would be provided. Helpline A dedicated and direct 24 hour telephone advice helpline was available to MPS officers in four south London boroughs within the SLaM Trust (Croydon, Lambeth, Lewisham and Southwark). Officers could call and speak to mental health professionals to obtain advice, background information on known individuals, and to be aware of the availability of places of safety. Face-to-face Assessments Additionally, a face-to-face assessment service was trialled, operating 24 hours a day. If appropriate, mental health professionals were asked to attend mental health related incidents with the police present. With their specialist knowledge, the professionals provided assessments in person and aided the police in their decision making. The introduction of London Street Triage held a number of aims, including: • Reduce the use of S136 MHA amongst the police • Provide a more timely intervention by mental health professionals • Improve the experience of people who come into contact with the police through either detention under S136 or for other reasons related to their mental health • Ensure that individuals detained under S136 are not taken to a police station as a place of safety • Help officers make appropriate decision making for those people who were not subject to a S136 MHA • To include the service user in the outcome of the contact • Achieving a substantial cost saving for police services 8 It was expected that individuals who came to the attention of the Street Triage Service would be treated and managed within a whole care pathway approach. Collaborative working between health and the police was in order to ensure individuals receive a coordinated approach to address their health and/or social care needs. The implementation of the Street Triage service has provided a platform from which to standardise the mental health response in four boroughs within London, enabling an exploration of the issues that arise. Evaluation of London Street Triage The Mayor’s Office for Policing And Crime (MOPAC) Evidence and Insight (E&I) unit is a team of social scientists who were commissioned to conduct research on Street Triage – specifically to explore the process and potential impact of the service. Results are structured into holistic learning about: 1) the implementation of Street Triage; 2) the impact seen from the service; and 3) wider learning relevant for mental health policy and practice. 9 2. Methodology This evaluation was conducted using a range of methodologies looking at the process and potential impact of implementing the Street Triage service. Exploring the Impact One aspiration of the research was to explore the potential impact of Street Triage on a range of key measures (e.g., the time spent dealing with an incident, number of people in crisis who are taken to custody) through the interrogation and analysis of MPS data systems and NHS data. In the absence of randomisation, a quasi-experimental matched comparison group was the most robust feasible approach. Matched comparison boroughs were identified from factors theoretically relevant to Street Triage, for example volume of mental health calls and socio-demographic factors. It became apparent a number of those best matched boroughs were already receiving an additional service, the Liaison and Diversion (L&D) scheme, at the time of matching. As this may have acted as a confounding variable, four of the best matched L&D boroughs, as well as an additional four of the top matched boroughs acting as the control (no L&D services) were chosen (see Appendix A for further details). Initially the analysis had planned to compare the results from the Street Triage boroughs to those in the L&D matched boroughs and the control boroughs. However, on initial inspection of the data, there were very few differences emerging between the L&D boroughs and the control boroughs in terms of impact measures such as number of crime incident reports and number of calls to Computer Aided Dispatch (CAD) 4, resulting in the decision to combine the findings of these eight boroughs together under the name of “Matched boroughs” (in effect generating one robust control to compare to Street Triage). The analysis presented here compares the Street Triage borough data to that of the Matched boroughs, and a wider pan London view that includes all London boroughs. Exploring the Process This aspect was focussed upon describing the Street Triage process and generating learning. This evaluation focused on the process of embedding a new system within two established organisations (the MPS and NHS). The E & I team adopted a holistic approach, incorporating a range of qualitative and quantitative methodologies including (full methodologies can be found in Appendix B):  Attitudinal surveys of police officers and NHS staff who were involved in the pilot  Focus groups with service users and carers 4 Met Central Command Unit houses the CAD team who receive all 999/101 calls and if necessary dispatch officers/vehicles to attend. When a call is received, an operator takes the details of the call, and assigns it a code which details the nature of the call and any significant factors. One code that can be applied is known as a mental health code (Qualifying Code: 612). 10 The evaluation methods have not been without their difficulties. Response rates to the police officers surveys were low (<6% of respondents, a lower response rate than the typical 10% rate of online survey returnsxvi). This low uptake could be perceived as a finding in itself. Reports were received to indicate the Post-Phonecall survey was not readily distributed to officers who had contacted the phone line due to lack of staff members within the MPS Central Mental Health teams, and ambivalence about who would pick up this work. 11 3. Results Process Findings: Implementing London’s Street Triage Early Implementation The research evidence is clear that well designed and implemented schemes are more likely to get better outcomes. Documenting learning from implementation is also beneficial, gathering learning, insights and enabling replication in similar operational settings. It is well recognised that a key issue when attempting to produce robust evidence for policing is that research faces challenges to implementation within ‘real-life’ policing environments as opposed to ‘clean’ medical trials.xvii One implementation issue identified was that for the first four months of the pilot the SLaM Street Triage team was not fully staffed, making it problematic to cover all the necessary shifts and attend face-to-face assessments.xviii It took up to four months until a full team was assembled, due to difficulties in recruiting staff with appropriate experience and qualifications for this specialised work. Previous researchxix highlights the potential for early implementation ‘teething issues’ and the lasting impact these can have. The method of recording data was challenging as data was collected by both health and police, and specific data requests were often received from the national pilot and local evaluation. This was hampered further by the lack of funding for administrative support, team leader or senior input within SLaM, impacting on the level of commitment at times to the pilot, especially around data and report writing.xx It became apparent during the research that communications around Street Triage to police officers was a key issue in terms of ensuring effective implementation. The pilot was slow to take off, with only 47 calls within the first monthxxi, due to a perceived lack of knowledge and understanding from the officers about the service. To illustrate, six months into the pilot only 49% (n=54) of officers who responded to the survey indicated they had heard about the service. To promote use, officers were provided with a card detailing the telephone number and encouraged to use it whenever they came into contact with someone with suspected mental health needs (see Figure 1). Figure 1. Street Triage contact cards. 12 Procedures were inaugurated to improve knowledge of the service in the form of additional localised briefings by Borough Mental Health Liaison Officers and members of the Central Mental Health team. A number of briefings were conducted throughout the year pilot to remind officers to utilise Street Triage. Early on in the pilot it became clear that officers were contacting the Street Triage team too late – for example, contact was often made after a decision to detain an individual was made.5 If contact had been made earlier then the team could provide more options which may have resulted in a potentially better care pathway for the individual. In an attempt to address this the clinical lead for Street Triage and Lambeth Crisis services also attended the MPS Communications Command6 in Lambeth and spoke to dispatch personnel, informing them of the Street Triage service and encouraging them to remind officers to contact the Street Triage number if suitable when attending incidents. It is important to acknowledge however that police often get involved in mental health crisis incidents as a last resort. It is not always appropriate or possible to contact Street Triage prior to acting, especially if a spontaneous response is required. Although mentioned in the Codes of Practice, there is currently no legal provision to ‘detain’ someone whilst an officer consults the triage service. The model implemented within London was unique in comparison to other pilot sites nationally. It was important to focus on the implementation of such a scheme to enable valuable learning for future models and compare, where possible, to the other methods used nationwide. There is a need to standardise the Street Triage process, to provide a more streamlined approach and ultimately better care pathways for individuals. Use of Street Triage The success of the Street Triage project is difficult to judge. This is often found within evaluations of pilots or newly established programmes. In terms of usage, in total, within the year Street Triage was running (April 2014 – March 2015), there were:  9,434 recorded CAD call’s that received a resolution code of mental health.  A total of 4,453 incidents that officers attended (47%) (Remaining CAD calls required telephone advice or appointments).  Street Triage team received 1,179 calls suggesting officers used the Street Triage service for around 26% of occasions. On average, this is approximately 90 calls to the Street Triage team a month (Figure 2). 5 This is anecdotal evidence only, as this detail of information was not recorded. 6 MET Communications Command answers all 999/101 calls and dispatches officers to incidents. 13 As can be seen from Figure 2, calls to the helpline gradually improved over the duration, peaked in November, then levelled out. Findings from the survey regarding usage were less positive, few officers reported being aware of the availability of the Street Triage service (41%, 85 officers out of 205) and only 47 officers out of 205 (23%) who responded to the evaluation surveys indicated they had actually called the telephone helpline. Those who did contact the Street Triage team called on average three times over the year period. This could be associated with the implementation and communication issues previously highlighted. Figure 2. Time series analysis of CAD calls in relation to calls to Street Triage service The level of usage seen in the Street Triage pilot would appear a solid start to the helpline. However, it remains clear that there were many incidents where Street Triage was not called in eligible cases. On such occasions officers may have drawn on their own expertise, local knowledge or signposted individuals to other services without the need to call Street Triage or there may have already been health and social care professionals or LAS at the scene due to the nature of the incident. On the contrary, there is wider research on the challenges in encouraging staff to use new or innovative programmes (for example the implementation of Body Worn Videos with the policexxii). Regardless of the explanation, more communication and promotion of the service would likely be helpful in encouraging use, especially whilst the new scheme takes time to embed and everyday practices begin to change. Individuals, for whom the police contacted the Street Triage team about, were usually already known to SLaM as current patients (n=866, 74%xxiii). Being known to services enabled the Street Triage nurse to provide up-to-date medical information to the officers, aiding the decision making process. Unlike other services (e.g., Home Treatment Team), when the individual was not known to SLaM, the mental health professional can still provide relevant information based on the behaviours presented by the individual. 14 Officers very infrequently contacted Street Triage when the individual with mental health needs was actually involved in a crime, whether as a suspect (5%) or a victim/witness (2%).7 Instead, over 90% of the contacts with Street Triage were individuals encountered in the community (whether within their own home or in a public place). When first implemented, the majority of calls were for attendances to an individual’s private property with officers asking for advice as they were unable to conduct a S136. However, over time officers started to utilise the Street Triage expertise when they encountered individuals in public spaces and over the year pilot this was the predominant use of the service (n=868, 74% of incidents were in public locationsxxiv). Whilst this could reflect that officers were phoning for advice (n=371, 40%)xxv, the types of engagement recorded by SLaM would suggest that officers used the Street Triage service more to obtain a space within a suitable place of safety (n=488, 52%). On only 8% (n=72)xxvi of occasions did the Street Triage professionals advise the police to invoke S136 procedures.8 This may reflect the anecdotal evidence suggesting officers were calling the Street Triage team once they had already invoked a S136, rather than calling them for initial advice. This may indicate that a cultural shift in officer’s thoughts and behaviours is required to encourage use of the service. It became apparent that certain ‘mental health triggers’ were repeatedly the main issue present at the time of contact from the police. Observations over the year highlight that the two main reasons why officers called the helpline were self-harm (28%) and unusual behaviour9 (37%).xxvii The majority (58%) of contacts with the Street Triage team resulted in the outcome of a S136; however, officers had usually invoked the S136 prior to calling the helpline. Other frequently used outcomes included sending the individual to A&E (10%), providing a community referral to mental health services (17%) or returning the individual to an inpatient ward (3%).xxviii Analysis of a CAD dip sample with a mental health code (n=342) indicated that from the pilot boroughs, mental health repeat callers are a regular occurrence; 41.5% of calls received were from repeat callers. This indicates a level of recurrence and relationship in understanding this demand. Officers only called for additional support a third of the time, and this was usually additional back up from the MPS (55%) or support from the London Ambulance Service (32%). Throughout the dip sample, the Street Triage service is mentioned very infrequently (4% of CAD records) and when it was mentioned, it was rarely contacted (2%, 6 occasions). Street Triage Impact and Outcomes Data reported here examines the impact analysis in terms of Street Triage boroughs in comparison to the Matched boroughs and pan London picture. The time period for the analysis is April 2014 – March 2015, during which time the Street Triage service was in action. 7 Data covering July 2014 – March 2015 (Quarter 2, 3 and 4) only due to data collection methods within SLaM. 8 Anecdotal evidence indicates that in the early stages of Street Triage, some officers were advised to take the patients to A&E rather than use S136. On many occasions officers felt therefore it was inappropriate so used S136 against the Street Triage professional advice. 9 Unusual behaviour is hard to describe, but a few consistent behaviours that are deemed unusual include being semi undressed/naked, shouting using language that does not appear to make sense, and running in and out of traffic or wandering in the road. A clear definition needs to be addressed as part of the new Crisis Line. 15 Key Outcomes Records indicate an increase in both recording of vulnerability10 (Merlin S136) and criminal incidents where there is evidence of mental health issues (CRIS), both in the pilot boroughs, matched boroughs and across London (Table 1). In this way there appears to be consistent pan London trends in terms of a rising demand around mental health in London. The Street Triage boroughs are entirely consistent with the broader picture – although given the wider context it is not possible to attribute any changes observed to Street Triage. This rising level of demand would appear to warrant attention – not only for Street Triage boroughs but for policing (and partnership working) as a whole for London. Additionally, adult Merlin records were only introduced in April 2013; therefore, the findings may reflect the embedding period where officers are adapting to completing Merlin’s regularly. An increase in records may be expected as this practice becomes routine, rather than an actual increase in the number of S136s being conducted across London. When placed under S136, officers have the responsibility to take the individual to a Place of Safety.11 Historically police custody has been used as a POS when the person has not committed an offence and has been solely detained under S136. This tends to occur when there are no other ‘health’ based Place of Safety alternatives available. The pilot boroughs previously had the highest number of 136 detentions in custodyxxix, but recently there has been a significant positive reduction in the use of custody. Across London, between 2008 – 2011 individuals were taken to custody as a Place of Safety an average of 65 occasions per yearxxx; a figure that has remained largely static over longer timeframes.xxxi However in 2014/15 this reduced to just 17 occasions in the pilot boroughs.xxxii Similar reductions (21 occasions) have also been seen across the whole of London in the same time frame, indicating that whilst this reduction cannot be credited to the Street Triage pilot, there has been a move towards a more positive approach to individuals with mental health needs by the police and a desire to find an appropriate response and places of safety for the individuals in crisis. 10 MPS Vulnerability Assessment Framework definition of vulnerability includes being exposed to the possibility of being attacked or harmed, either physically or emotionally. Additional factors to vulnerability may include Mental Health, Disability, Age or Illness 11 The London Ambulance Service has the responsibility to transport the patient to a Place of Safety, with the assistance of the police in the escort. 16 Table 1. Data from Merlin, CRIS and Place of Safety records across London Street Triage aimed to reduce the time officers spent dealing with such incidents, by providing them with the advice and information required to make decisions that would result in a positive outcome. Police data was used to assess whether the time spent at an incident involving someone with mental health needs had reduced during the time the Street Triage pilot was running. For cases when an individual has mental health needs, this may include assisting the individual to A&E or a S136 suite and waiting with them until they are attended to. Analysis of a dip sample of CADs during November 2014 (see Appendix C) provided an average duration from the ‘time of arrival’ (when an officer arrives on scene) to a ‘de-assigned’ time (when an officer has completed their time at the incident). This provides a ‘dealing with’ duration figure. During the pilot period, on average, incidents on the pilot boroughs were ‘dealt with’ in 3 hours and 53 minutes. This is faster than incidents on the Matched boroughs (4 hours and 26 minutes) and examining a random dip sample of 400 across all London boroughs for the same month indicates that the average dealt with time is 4 hours and 44 minutes. Looking specifically at the dip sample CADs for incidents when a S136 was invoked, officers on the pilot boroughs made six S136s in November 2014. On average, officers were assigned to these incidents for 2 hours and 14 minutes. On the matched boroughs however the average time for dealing with their nine S136s was considerably longer (6 hours and 34 minutes). Although there is no specific indication that Street Triage helped reduce these time durations, as it was not mentioned in the majority of these incidents, speculatively it may have had a wider positive impact on relationships between the police and SLaM, consequently improving the S136 process. It is important to remember however, that just because officers are located within the pilot boroughs, they may not have utilised the Street Triage service (only an estimated 26% called) thus not receiving the potential benefits and time saving advantages the service could provide. Additionally, these times need to be caveated, as different boroughs across London are covered Merlin Records with a S136 Apr ‘13 - Mar ‘14 Apr ‘14 - Mar ‘15 % change Street Triage Boroughs 261 385 47% increase Matched Boroughs 381 795 109% increase Pan London 1,721 2,829 64% increase Mental health related CRIS Records Street Triage Boroughs 1,314 1,677 28% increase Matched Boroughs 2,067 2,786 35% increase Pan London 7,605 9,978 31% increase Number of Detections in Custody as Place of Safety Street Triage Boroughs 46 17 63% decrease Matched Boroughs 10 2 80% decrease Pan London 73 21 71% decrease 17 by different NHS Trusts, each with different numbers and access protocols to places of safety, therefore it is expected that there is some variation in timings. Officer and SLaM Staff Attitudes When asked specifically about the Street Triage service, officers and SLaM staff were12:  Generally positive, with feelings of satisfaction around the service received.  Willing to use the service again.  Those officers who called the helpline also felt it had improved their confidence in working with mental health and SLaM staff felt confident in their role advising officers.  Take action on advice provided.  In terms of care for individuals in crisis, the phone line was seen as making the Section 136 process easier, as well as the sharing of information which ultimately helped with the continuation of patient care.  Mental health professionals also felt the officers handled the situation well.  Officers noted a continual complexity with relationships with partnership agencies (health organisations in general). Calling the helpline was seen as reducing the amount of time officers spent dealing with incidents involving mental health – reflecting the findings which indicate Street Triage may help reduce the length of incidents. Going further, analysis suggests that if officers felt that the Street Triage service reduced time spent dealing with incidents, this was correlated with a positive feeling towards the Street Triage service assisting officers to reach a decision (r = .506, n = 149, p = .001). SLaM staff noted that the time spent with individuals could have been reduced further if officers contacted the helpline before they decided to invoke a S136, rather than after. Thoughts around potential services/tools officers would find helpful in the future included:  A continuation of the Street Triage telephone helpline  Mental health professionals attending police incidents  Mental health professionals in 999/101 call centres to offer advice to operators and officers over the radio  Mental health professionals based within A&E13 xxxiii  Better training for police to understand mental health needs 12 Data from Police Officers Attitudinal Survey’s and NHS SLaM Street Triage staff Survey. 13 There are mental health liaison teams based within A&E departments so it is not clear what the officers mean by this comment. It could be speculated that officers would like closer working relationships with these teams. 18 Other learning emerged from staff in terms of the daily hassles of everyday working around mental health. One issue that contradicts other MPS feedback is that when specifically asked about working with mental health, officers did not want access to a Smartphone or Tablet offering information as it was seen as taking up valuable time and providing no firm answer. Officers also did not want direct access to personal medical information, possibly because officers felt this was an intrusion of their privacy, or a reflection that they were not trained medical professionals. Wider Learning from Street Triage As well as providing detailed information about the embedding process of the Street Triage service, elements of the evaluation have provided wider learning around working with mental health in London. A number of issues were identified, focussing mainly on training of officers, perceptions of officers, service users and carers, pathways, confidence in working with mental health and the importance of partnership communication. Training One method of improving officers understanding of mental health needs is through training. Over the years, researchxxxiv has repeatedly suggested that police officers attitudes and response to individuals with mental health needs is strongly mediated by the training they receive. The current research continues to highlight the importance of training, but also that Street Triage is not a plug for the current training gap: at the time of interviewing officers did not find the current MPS training adequate. Our work suggests that officers identified the issues lay less in the amount of training they currently receive, but ensuring content and delivery is in a manner that can be easily understood, learnt, remembered and applied in real life. In particular, a need for adequate training around the identification of, response to, and legal aspects of mental health was Face-to-Face Assessments The opportunity to have a mental health professional attend an incident was used infrequently, with only 99 assessments made over the year period. This may be due to officers not requesting attendance by the mental health professional, or limited capacity within the SLaM Street Triage team, operating on limited resources. Such low usage has meant that it is not possible to explore the element of the pilot scheme in more depth and further research is required if this aspect of Street Triage is to be rolled out in future. However, mirroring officers and SLaM staff views of the telephone helpline, those who used the face-to-face aspect found it:  Positive and helpful.  Appreciated the detailed medical knowledge of the professionals.  The professionals were taking responsibility for the individual.  Resulted in a more positive outcome for the service user too. 19 “Police make the situation hot and heated so they can get the result they want” Service User highlighted. These findings continue to raise the challenge of providing appropriate training around management of those with mental health needs, in a way that embeds the knowledge in officer’s everyday practice.xxxv This is further illustrated by a third of officers reporting they do not feel supported by borough management, suggesting the MPS attitude to the relevance of understanding mental health issues in the work place would require further attention.xxxvi Mental Health service users and carers reported that police need more training to increase the fairness of encounters (e.g., procedural justice training). In particular, service users suggested police needed to obtain ‘real life experience’ to be able to address situations more appropriately. One method of introducing this into training could be through the joint workings of service users and trainers, to provide officers with examples of ‘lived experiences’ from the service users. A further reoccurring theme from service users and carers was the importance that police are able to understand culture and culturally specific mental health needs. Research demonstrates a moderately strong benefit of culturally adapting mental health interventions.xxxvii Specifically, interventions targeted to a specific cultural group were four times more effective than generic interventions.xxxviii Training focused around culturally specific aspects of mental health would aid officers’ understanding of individual’s different reactions, enabling officers to provide a more holistic approach to those they encounter. Perceptions of Mental Health Needs This section provides an understanding of the different perspectives of the service users and carers, as well as the police officers. A unique voice captured within the research—that of the service users and carers—was able to provide original insights and an in-depth perspective on how the police are experienced in times of crisis. Additionally, through survey questions, officer’s perceptions when encountering individuals with mental health needs were attained. These two perspectives should be read in conjunction, as they provide contrasting viewpoints. Mental Health Service Users and Carers Through discussions with service users and carers, three themes emerged around how police are experienced in times of crisis: A Fair Encounter Consensus amongst participants was that the police do not understand the people they are encountering and this may have the end result of officers aggravating a situation making it worse than it needs to be. It was also felt that officers can be overly result focussed (e.g., arrest, charge) instead of considering the wider encounter or individual needs. The perception was that the location of the incident affects the response. When the police were called to a home address, carers suggested the police were more respectful. However, on the street, or in hospital, the police response was perceived as much more “menacing”. 20 “Uniform is very intimidating. In the height of crisis all you are is uniform. It’s like a ‘red rag to a bull’” Service User Fair treatment within the encounter also emerged as paramount.xxxix This included aspects such as body language, mannerisms of the officer, tone of voice and personal contact. It is also worth highlighting that some carers/service users reported to be intimidated by the uniform itself. Some of these aspects cannot necessarily be altered (e.g., uniform) but some are more open to be addressed. The point to highlight here is the range of factors that such a vulnerable group hold important to them during the encounter. Carers reflected that at times police attendance was necessary and they themselves would contact the police. It was recognised that some service user’s behaviours may be seen as ‘threatening’, requiring a response from the police. Whilst carers felt police encounters could be improved, they understood the necessity of their actions in certain situations. Culture & Community Service users and carers perceived they were viewed akin to a minority group. They felt the officers they had encountered lacked local engagement/knowledge, and held narrow views, in particular around culture and mental health. Participants suggested there needs to be education not only in mental health but in cultural backgrounds, so that officers can understand the different ways people react may be due to their culture rather than a mental health related issue. Information Sharing The sharing of confidential information and Care Plans was seen positively, especially if it helped to reduce risk and provided officers with information that could help them approach and “deal with them [service users]” in a more appropriate way. However, it was felt important to keep the information focused on the necessary details and, if the individual has numerous encounters with the police whilst in crises, then service users wanted the information to be fully shared with them in advance. Police Perceptions of Mental Health This section provides an overview of officer perceptions when encountering individuals with mental health needs. Key results relating to officer perceptions are14:  Over a third of officers (37%, n=40) describe fearing for their own safety when dealing with individuals with mental health needs  Not wanting to make the situation worse (55%, n=32)  Worrying about taking the wrong choice of actions (30%, n=17)  Being concerned about the legal reprisals of the actions taken (83%, n=48)  Only sometimes knowing how best to help (78%, n=45) 14 Not all officers who responded to the survey answered every question, therefore the number of respondents (n) per question varies. “Police need to understand cultural issues first - may be wrongly viewed as mental health issues, such as speaking too quickly or passionately is seen as pressure of speech" Service User 21 “I would attempt hands on first to pull them off, then escalate my use of force if that didn’t work. This would be the same regardless of whether they had MH or not” Police Officer – Mental Health and Violence condition  There are frequently lots of unknown circumstances (39%, n=40)  Being concerned the individual will ‘lash out’ and become violent (36%, n=21) This illustrates a mental health encounter from a police perspective. The unknowns and perceived risks appear to run through the situation. Whist there are links between certain mental health issues and violencexl, these links are often exaggerated.xli If police response to people with mental illness is driven by perceptions of dangerousness and potential for violence, this will likely influence the subsequent encounter and police are likely to draw more deeply on their use of force options.xlii Research suggests Taser is used15 considerably more on individuals with mental health needs.xliii The MPS and College of Policing are in the process of producing national guidance documentation for all UK Police Forces that will detail restraint issues, particularly with individuals with mental ill health, and will include information on Taser use. This paper is due for consultation in 2016. To investigate officer’s fear of safety and use of tactical options when encountering individuals with mental health needs, half of respondents to Survey 2 were provided with a vignette regarding a violent incident (n = 48); the other half received a vignette regarding a violent incident when the suspect had mental health needs (n = 63). Officers ranked which tactical option they would consider using. Positively (albeit a very small sample size) results indicated officers reported very little difference in their use of tactical options whether the individual had mental health needs or not. The majority of officers would try verbal de-escalation as a first resort, followed with an open-hand strike. In terms of using a weapon, officers ranked a baton above either Taser or CS Spray. Pathways By the nature of their job, police officers are focused at the front end of the service delivery in terms of mental health, often only being present when an individual is in crisis. In a similar vein, Street Triage by its very name is a triage service, and should be situated at the initial stages of the individual’s journey through the mental health services. It is important to improve this front end service, so that at least the start of the mental health service journey is directed in the correct way. One major aspect in delivering this service would be the earlier uptake (as discussed previously) of officers, as well as implementation of innovative, adequate and responsive health based mental health services. 15 The term ‘used’ in reference to Taser includes the removal of the Taser from the holster, red dot and firing the Taser. “[Verbal de-escalation is] least force. However rapidly switch to Taser if verbal shock/de-escalation is not effective” Police Officer – Violence Only condition 22 When encountering an individual with mental health needs, officers have a few options at their disposal. Officers report they most frequentlyxliv deal with the situation at the scene, or in line with MPS training, offering individuals ‘Time - Distance - Space’.xlv Following the Standard Operating Proceduresxlvi, other options less frequently used include detaining an individual under Section 136 Mental Health Act (S136) and taking the individual to a place of safety, but also delivering the individual to a place of safety without a S136, such as the A&E Department, or arresting and taking the person to custody. Officers highlighted that they would appreciate more options available to them (e.g., access to designated substance misuse centres), and clearer guidelines about which pathway to use in different situations.16 Confidence in Mental Health It is important for officers to feel and act confident in their role, to portray an element of authority in particular around mental health crisis. Confidence here is closely linked with training and much research has shown that officers are negative towards their mental health training. To demonstrate this, when engaging with individuals with mental health needs, only half of officers reported they always feel confident in their actions, which could impact on the encounter and again stresses the need for enhanced training. This call for enhanced training is not a new one and it may be that training such as the Mental Health Briefings and MAST (Mental Health Awareness and Safeguarding Training) may be able to meet this challenge.17 The Importance of Partner Communication Much has been written in terms of the importance of partnership working.xlvii There were many signs of this within the research, in particular a lack of communication and collaboration between agencies. To illustrate, police officers often described resistance to cooperate from partnership agencies as one of the biggest barriers they face, as well as poor communication between partner agencies hampering good practice in day-to-day working. Indeed, this communication between different services can be made more difficult by numerous terms used to describe mental health needs. A lack of consistency and standardisation is not an enabler of partnership working. The findings indicated SLaM staff members were not always fully aware of, or understood specific police related frameworks or models in relation to mental health. Officers also admitted not being fully aware of certain terminologies either. Despite all Territorial Police officers receiving training, from the Pilot Boroughs:  1/4 of officers reported understanding the Vulnerabilities Assessment Framework (VAF)  1/3 had not heard of the ABCDE model 16 This has since been addressed, as the MPS MH Team briefings, delivered in October/November 2015 provided clear operational guidelines. 17 E&I are conducting a Randomised Control Trial on MAST and exploring staff perceptions, due to report end of 2015. 23 The use of a variety of different data recording systems may also act as barriers to partnership working. Both within and between services (e.g., Police, NHS), data management systems often do not integrate with each other, hampering data sharing abilities and collaborative working. Another example of the necessity for effective partnerships is that of meeting needs and providing pathways. This was apparent in the survey when respondents outlined the overlap between substance mis-use and that A&E are unwilling to accept S136 patients, thus leaving officers with a limited number of options. There were a number of internal and external issues around access to 136 suites, as the Street Triage nurse was the key contact for officers when arranging a suitable place of safety. This resulted in Street Triage nurses often being seen in a negative light by some health and police staff members, and having to bear the focus of frustrations on all sides. This pilot highlighted a need for a robust system, both internally and externally around the process for escalation and resolution when there are no 136 suites available. This would help avoid delayed access to suites where patients are left waiting or officers have to cross boroughs and NHS Trusts to find suitable accommodation. 24 4. Discussion The data indicates a heavy and rising demand of mental health within London. There are also pan London trends in terms of rising numbers of criminal incidents relating to mental health and adult Merlin’s recording S136s. Street Triage boroughs were completely consistent with this wider pan London data. There was no 'impact' per se on these issues - equally no signs that Street triage had a negative influence on the boroughs. In terms of usage, figures estimate that officers used the service on approximately 26% of occasions, suggesting Street Triage would appear to have a solid start in terms of usage but with more to do to encourage staff and partnership working. This reports outlines and discusses issues that have arisen throughout the implementation process, which need to be considered for wider roll out across London. To progress, wider issues need to be discussed including sustainability, training and promotion of the service. Sustainability During the final few months of the Street Triage pilot, discussions turned to the sustainability of the service provided once the pilot duration came to the end. Both SLaM and the MPS were keen to continue with some sort of service, whether in the same guise or a condensed model as there was a strong belief that the Street Triage pilot had provided an invaluable service to officers as well as a requirement set out in the Crisis Care Concordat to provide a 24 hour mental health crisis service. The face-to-face service is part of the new service but will possibly be limited due to resources, as two staff members will not always be available. Staff working hours will be flexible whilst the needs of the service are established. This element was seen positively by the mental health professionals from SLaM and the MPS. The new service, which continues to be commissioned by all four boroughs involves a 24 hour telephone line for the use of police officers (both MPS and British Transport Police) and London Ambulance Service. Additionally there will be a free telephone number for service users and carers. At present this amalgamation of the Street Triage service and crisis line is only available within the four SLaM boroughs. Elsewhere in London, this service is unavailable, although individual boroughs may be introducing different schemes and services. There is currently no definitive pan-London plan to standardise the crisis services, nor provide a London wide service to police officers dealing with mental health related situations, however high level discussions are afoot. It is clear that when used, the Street Triage option was welcomed by officers, who appreciated the help and advice provided my mental health professionals in a timely fashion. However, this is in conflict with the reasonably low-level usage of the service over the year period. 25 “Workshops with the police and service users would help to educate them. This is the way you treat us…this is the way you should have treated us…” Service User “Speak in the same language – communicate on a humanitarian level and build trust” Carer “Stop using computer packages like NCALT to deliver important legislation changes and guidance” Police Officer A broader view of the current services also needs to be taken to understand how they currently fit with those using the services. For example, as illustrated in the report many of the callers that the police encounter when in crisis are known to mental health services. They are, in effect, returning through a ‘revolving door’ system. This issue is currently under discussion with the Mental Health Partnership Board Operational Group, who are mapping work to understand what this looks like in London. For sustainability of the service to be effective, it is important to make adaptations to current processes, including training of officers, the audit process and recording of data, and effectual communication of the service provided. Training Officers and NHS staff need to be fully equipped to understand not only mental health situations, but also the legal powers and abilities of each other. Many of these issues could be addressed via appropriate, challenging training to all involved. This is an issue that has been highlighted previously and a common theme running through this evaluation. Both police officers and mental health professionals were able to offer alternatives for future interventions. One of the most common themes emerging is around the involvement of service users and other agencies within the content of the training to provide a more holistic approach. Engagement of this nature may potentially improve the perception of fairness within the police, thus increasing legitimacy and policing by consent. Service users themselves also identified elements that could be easily implemented in the future with few resources. For example, a fair encounter with more respect and humanitarianism from the officers would benefit. To achieve these, it was suggested that officers attended workshops run by service users, or have joint training to address this. In late 2015, there are plans to conduct training sessions on mental health which will involve a session from service users from a charity called ‘HearUs’ in Croydon - a potentially valuable innovative aspect of training. Equally, E&I are in the process of evaluating MAST, which could also yield insights as to improving police training. 26 “Don’t ask for personal details – seems like you’re being questioned. Just ask if I am ok” Service User “Speak directly with someone who had a Mental Health Crisis and what their advice would be for police to assist them” Police Officer Over the past year, work has taken place with members of service user groups within SLaM, in an attempt to develop and formulate a specific set of four questions the police could use. Police would be made aware of such questions and trained to use them when approaching someone they believe to be in crisis. Although not yet finalised, and would need the agreement of the Central Mental Health team within the MPS, the questions would be based around “asking the individual if they are ok”, “providing their own name rather than asking for the individual’s name” and “offering help”. Through the focus groups, participants highlighted that if officers approached in a friendly, non-accusing manner, then they would be much more likely to cooperate. Promotion of the Service Street Triage was promoted through briefings from the Central Mental Health team and Local Borough Mental Health Officers/Leads to frontline officers and the distribution of information cards (Figure 1). However, the importance of effective communications and promotion of the service has been identified throughout this pilot. Moving forward, to increase success, key police officers with experience of working with mental health who can influence practice need to be identified. This often works best with officers who are frontline rather than coming from senior officers or trainers, as has been the case thus far (i.e., VAF training being delivered by MPS trainers). Officers who have had successful interactions with the Street Triage service previously could be asked to encourage other officers to use the service, enabling them to learn by experience. Conclusions The data indicates a heavy and rising demand of mental health within London. Street Triage has helped foster joint working by staff from MPS and SLaM NHS. The pilot fits well with the broader aims of the Mental Health Crisis Care Concordat to improve the standards of care provided to those experiencing any form of mental health distress. Street Triage has had a strong start and sits at the crisis point of ones’ mental health journey. It is important to look forward, focusing on embedding elements of Street Triage into the current NHS commissioningxlviii and MPS processes to establish the best pathways and parity of esteem for individuals in mental health crisis. 27 Appendices Appendix A: Matching Boroughs The four pilot boroughs (Croydon; Lambeth; Lewisham; and Southwark) are all situated in the South East of London, and are covered by the South London and Maudsley NHS Foundation Trust (SLAM). In forming the main control group a robust matching process was used to identify the boroughs within London which were best matched, in theoretically relevant ways, to each of the Street Triage pilot boroughs. Each of the pilot boroughs was compared against all other (non-Mental Health Street Triage) London boroughs in terms of:  No. of calls flagged as MH related (2014)  No. of S.136 (2011/12)  Estimated cases population aged 16-74 years with mental health issues (2010)  Total population by borough (All Ages) (2013)  Crime rates per thousand population (2012/13)  Indices of deprivation average score (2007) For each variable, where the comparison borough was found to be within a certain percentage comparable to the pilot boroughs a score was applied. See Table 2. Summing across the criteria resulted in an overall score for each non-Mental Health Street Triage borough, which represented the strength of its match to the pilot borough in question. The scored comparison boroughs were then ranked and the best match for each Street Triage borough was chosen as the four matched boroughs. Where one non-Mental Health Street Triage borough best matched more than one of the pilot boroughs, the second closes match was chosen. Table 2. Scoring used in borough matching task Before finalising the control boroughs, it became apparent another pilot (Liaison and Diversion (L&D)) was concurrently run across London, and at an enhanced provision in ten boroughs. As Between these Percentages Score Given to Borough >5% 5 5 % - 10% 4 10% - 20% 3 20% - 30% 2 30% - 40% 1 40% < No score applied 28 this may have acted as a confounding variable, the four top ranked of these L & D boroughs were chosen as separate group, creating two subgroups to the matched boroughs. This resulted in eight boroughs identified as the best matched to the pilot boroughs (see Table 3). Table 3. Matched Boroughs Matched Boroughs Liaison & Diversion Boroughs Control Boroughs Barnet (Barnet, Enfield & Haringey NHS Trust) Brent (Central & NW London NHS Trust) Enfield (Barnet, Enfield & Haringey NHS Trust) Ealing (Central & NW London NHS Trust) Haringey (Barnet, Enfield & Haringey NHS Trust) Greenwich (Oxleas NHS Trust) Newham (East London NHS Trust) Wandsworth (SW London & St George NHS Trust) 29 Appendix B: Methodology Surveys Post Phonecall Survey A survey was distributed, via a telephone call or email, to officers who had contacted Street Triage. The survey consisted of 8 short questions asking them about their experience and immediate feedback of the Street Triage service.  Response rate: 212 officers completed the survey (18%) Police Officers Attitudinal Surveys Police Officers attitudes to mental health were investigated through two surveys, distributed 6 months apart. The surveys looked to identify whether the Street Triage pilot had embedded within the pilot boroughs. In both surveys questions were divided into thematic groups including understanding and prevalence of mental health issues and services, barriers to conducting best practice on an operational basis, suggestions for improvements and specific questions on the use of the Street Triage pilot study. An invite was sent out via email to all Local Policing Teams (LRT) and Emergency Response Teams (ERT) in the four SLaM boroughs involved in the pilot. Survey 1 was released in August 2014 and kept live for seven weeks, but response rates fell below the typical 10% response for online survey completion. The survey was sent to approximately 2,233 officers, but only 110 of these officers completed the survey (a response rate of <5%). Survey 2 was released in February 2015 and kept open for eleven weeks to try and increase response rates. Despite numerous reminder emails from Chief Inspectors, and prompts from Borough Liaison Officers, only 95 officers on the pilot boroughs (a response rate of 5%) responded. Additionally, all LRT and ERT officers on the Matched Boroughs received the survey (between 4,000 – 5,000 officers in each wave of the survey). The number of completed surveys was again low with only 172 surveys from Survey 1 and 97 surveys from Survey 2, providing an average response rate of 3%. Whilst matching boroughs in this way provides a more robust study, due to the low response rates received, comparative analysis of the data has not been possible. Survey 1: The majority of respondents (from all boroughs) were:  Male (66%; female, 29%, prefer not to say 5%);  Aged between 25 – 34 years (44%) but respondents ranged from 18 years old to 65+ years old;  Experienced officers – 48% had been in the police service for over 10 years and a third between 5 and 9 years;  The rank of Police Constable (PC) (72%); 30  From Emergency Response Teams (48%) and Local Policing Teams (22%) Survey 2: The majority of respondents (from all boroughs) were:  Male (63%, female = 33%; n=90);  Aged between 25 – 44 years (25 – 34 years – 37%; 35 – 44 years – 38%), but respondents ranged from 22 years old to 64 years of age (n=89); NHS SLaM Staff – Street Triage Attitudinal Surveys SLaM Street Triage mental health professions were asked to complete an online survey regarding their thoughts and opinions of the Street Triage service on offer. This included staff members who answered the 24 hour telephone helpline, and those who attended face-to-face assessments when required. The survey focused on their understanding of the role of Street Triage, relationships with officers and their use of the Street Triage service, types of information provided, suggestions for improvements and more specifically how they felt when attending face-to-face assessments. In February 2015 an invitation to complete the survey was distributed via email to all staff members involved in the running of the Street Triage service over the year period. The survey was live for four weeks and received eight responses. This reflects all of staff members who have worked on the Street Triage service, providing a 100% response rate. The respondent’s demographics are:  An equal gender split (Males = 4, Females = 4)  Aged between 25 – 34 years (n=4; 35-44 years n=1; 45-54 years n=3). Focus Groups The aim of the focus groups was to explore how individuals with mental health needs, who have come into contact with the police, are treated by both the police and NHS staff members. Service Users were recruited via an email advert distributed to a number of Clinical Academic Groups (CAG’s) and user groups within the SLaM boroughs. Participants who responded to the advert had to be currently well, and have had previous (within last few years) contact with the police when in crisis. Where possible, the mental health professional would review the participant’s history and current status to ensure their suitability to participate. Carers of Service Users who have been in contact with the police when in mental health crisis were also invited to attend their own focus groups. Carers are often involved in the service users Care Plans, and in some cases, it is the carer who contacts the police in the first instance. The aim of this focus group was to explore how carers are treated by the police and NHS staff, if they have been present when the Service User is in crisis. Due to the nature of the subject area, there was potential for negative emotions/feelings to be exacerbated through our discussions. Time was provided at the end of the focus group for 31 individuals to talk to the mental health professional on a one-to-one basis if required. If this occurred, SLaM also agreed to follow up the individual the next day with a phone call to the Service User or a discussion with their Care Coordinator. Three focus groups (2 x Service Users, 1 x Carers) were held in March and June 2015 and facilitated by a mental health profession from SLaM, providing expert knowledge if required. Members of the MOPAC E&I team attended as note takers, to enable the Service Users to see that their voices will be heard, but without influencing the discussions. The sessions ran for approximately 90 minutes each.  Six Service Users attended.  Four Carers attended. Topics for discussion within the focus group included:  What are your general views of the police? CPS?  Thinking about how the police have reacted to you/Service User in the past, have the police treated you fairly?  How would they like the police to react on arrival?  How much of your medical information would you like to be shared? What information would you like the police to know prior to attending?  Thinking about the last time you were in contact with the police (because you were in crisis), how did you feel?  Have you called the police in the past regarding the behaviour of your ‘Service User’?  Did the police discuss with you the outcomes of their attendance?  When in contact with the police (because you were in crisis), what four questions you would like to be asked by the police? / Four things the police could do? 32 Appendix C: MPS CAD Data Data extracted from an MPS computer system known as CAD (Computer Aided Dispatch) has been coded according to a bespoke framework. CAD is a method of dispatching emergency vehicles in relation to 999/101 calls. The CAD system also records information passed between the caller and operator, radio communication between operator and attending officers, locations, times and details of the incident being reported. In total, approximately 1,000 CADs, with an Open code or Classification code relating to mental health (code 612) in November 2014 were randomly selected. Four hundred of these CADs were incidents located within the four pilot boroughs. An addition 600 CADs were in relation to incidents located within the eight Matched boroughs, split equally between L & D boroughs and non-L & D boroughs. Although much of the data could be directly extracted from the CAD system (via DARIS), more bespoke information relating to comments and actions are recorded in the free text ‘Remarks’ section of CAD. In order to analyse this data, two MPS Volunteers were trained to code the required data, following a specific framework. This information included details such as whether Street Triage was mentioned or contacted, whether a S.136 was initiated, type of location (private home, street etc.), reason for mental health code, and whether the caller was considered a repeat caller. 33 References i Bradley, K. (2009). The Bradley Report: Lord Bradley’s Review of People with Mental Health Problems or Learning Disabilities in the Criminal Justice System. London: DH; Flanagan, R. (2008) Final report of the independent review of policing commissioned by the Home Secretary and conducted by Sir Ronnie Flanagan. London: Home Office; Cabinet Office (2007) The Cabinet Office Policy Review, Building on Progress: Security, Crime and Justice. London: The Cabinet Office; IPCCC. (2006). Deaths During or Following Police Contact. Available at: www.ipcc.gov.uk; Independent Commission on Mental Health and Policing report. (2013); Department of Health and Concordat Signatories. (2014). Mental Health Crisis Care Concordat – Improving outcomes for people experiencing mental health crisis; Home Affairs - Eleventh Report Policing and mental health (2015). ii MIND (2015). http://www.mind.org.uk/information-support/types-of-mental-health-problems/statistics-and- facts-about-mental-health/how-common-are-mental-health-problems/ iii Independent Commission on Mental Health and Policing report. (2013). iv Grossmith, L., Franklin-Trespeuch, E., & Dawson, P. (2012). Mental Health and the Metropolitan Police Service. MPS Corporate Development. v MPS CAD data. (Extracted April 2015). vi College of Policing. (2015). Estimating Demand on Police Service. http://www.college.police.uk/News/College- news/Pages/First-analysis-of-national-demand.aspx vii Independent Commission on Mental Health and Policing report. (2013); London Strategic Clinical Networks (2014). London Mental Health Crisis Commissioning Guide. viii Mental Health Crisis Care Concordat (2014). http://www.crisiscareconcordat.org.uk/ ix Independent Commission on Mental Health and Policing report. (2013). x Grossmith, L., Franklin-Trespeuch, E., & Dawson, P. (2012). Mental Health and the Metropolitan Police Service. MPS Corporate Development. xi Stanko, E. & Dawson, P. (2016). Police Use of Evidence: Recommendations for Improvement. Springer International Publishing xii Police and Crime Plan: London 2013 – 2016. Mayor’s Office for Policing And Crime (2013). xiii Department of Health (2009). The Bradley Report: Lord Bradley’s review of people with mental health problems or learning disabilities in the criminal justice system. London. http://www.centreformentalhealth.org.uk/pdfs/Bradley+report_2009.pdf xiv Durcan, G., Saunders, A., Gadsby, B. & Hazard, A. (2014). The Bradley Report five years on: An independent review of progress to date and priorities for further development. Centre for Mental Health. xv Mental Health Crisis Care Concordat (2014). http://www.crisiscareconcordat.org.uk/ xvi Grossmith, L. & Dawson, P. (2014). An Armed Response: A survey of Firearms Officers’ opinions of Post-Incident Procedures. MOPAC Evidence & Insight xvii London Criminal Justice Partnership. (2011). An evaluation of the Diamond Initiative: year two findings http://www.londoncjp.gov.uk/publications/diamond_year2_FINAL_050511.pdf xviii SLaM NHS Trust Street Triage Quarterly data reports xix Grossmith, L., Davies, T., Owens, C., Finn, W., Mann, D. (2015). Police, Camera, Evidence: London’s cluster randomised controlled trial of Body Worn Video. xx SLaM NHS Trust Street Triage Quarterly data reports. xxi SLaM NHS Trust Street Triage Quarter 1 data report. xxii Owens, C., Mann, D., McKenna, R. (2014). The Essex Body Worn Video Trial: The impact of Body Worn Video on criminal justice outcomes on domestic abuse incidents; Grossmith, L., Davies, T., Owens, C., Finn, W., Mann, D. (2015). Police, Camera, Evidence: London’s cluster randomised controlled trial of Body Worn Video. xxiii SLaM NHS Trust Street Triage Quarterly data reports. xxiv Data received from GlenDay, V. Clinical Lead at SLaM NHS Trust. xxv SLaM NHS Trust Street Triage Quarterly data reports (June 2014 – March 2015 data) xxvi SLaM NHS Trust Street Triage Quarterly data reports (June 2014 – March 2015 data) xxvii Definition and Data - SLaM NHS Trust Street Triage Quarterly data reports. xxviii SLaM NHS Trust Street Triage Quarter 4 data reports xxix Grossmith, L., Franklin-Trespeuch, E., & Dawson, P (2013). Mental health and the police: Understanding demand and incident management in the Metropolitan Police Service. xxx Apr 08 - Dec 11 - CJU NSPIS - All arrests where the ‘First Arrest Offence’ shows ‘Mental Health Act’ xxxi Grossmith, L., Franklin-Trespeuch, E., & Dawson, P (2013). Mental health and the police: Understanding demand and incident management in the Metropolitan Police Service. xxxii MPS Data Extract (April 2015). xxxiii NHS London Strategic Clinical Networks. (2014). London Mental Health Crisis Commissioning Guide. 34 xxxiv Grossmith, L., Franklin-Trespeuch, E., & Dawson, P. (2012). Mental Health and the Metropolitan Police Service. MPS Corporate Development. xxxv Bather, P., Fitzpatrick, R., & Rutherford, M. (2008). The Police and Mental Health. Sainsbury Centre for Mental Health Briefing 36. xxxvi Grossmith, L., Franklin-Trespeuch, E., & Dawson, P. (2012). Mental Health and the Metropolitan Police Service. MPS Corporate Development. xxxvii Al-Krenawi, A. & Graham, J. R. (2000). Culturally sensitive social work practice with Arab clients in mental health settings. Health Social Work, 25 (1), 9 – 22. xxxviii Griner, D. & Smith, T. B. (2006). Culturally adapted mental health intervention: A meta-analytic review. Psychotherapy: Theory, Research, Practice, Training, 43 (4), 531 – 548. xxxix Independent Commission on Mental Health and Policing report. (2013); Department of Health and Concordat Signatories. (2014). Mental Health Crisis Care Concordat – Improving outcomes for people experiencing mental health crisis. xl Choe, J. Y., Teplin, L. A., & Abram, K. M. (2008). Perpetration of violence, violent victimization and severe mental illness: Balancing public health concerns. Psychiatric Services, 59 (2), 153 – 164. xli Markowitz, F. E. (2011). Mental illness, crime and violence: Risk, context and social control. Aggression and Violent Behaviour, 16 (1), 36 – 44. xlii O’Brien, A. J., & Thom, K. (2014). Police use of Tasers in mental health emergencies: A review. International Journal of Law and Psychiatry, in press. xliii O’Brien, A. J., & Thom, K. (2014). Police use of Tasers in mental health emergencies: A review. International Journal of Law and Psychiatry, in press; Grossmith, L., Franklin-Trespeuch, E., & Dawson, P (2013). Mental health and the police: Understanding demand and incident management in the Metropolitan Police Service. xliv Responses have been collapsed into those who gave a score of ‘frequently’ and ‘for most incidents’. xlv A technique that has been taught for many years in officer safety training for those situations where that is appropriate. As the name of the technique suggests, officers offer distressed individuals time, distance and space to calm down. xlvi Policing Mental Health Standard Operating Procedures, v. 4.2 (2011) xlvii Gilburt, H., Edwards, N. & Murray, R. (2014). Transforming mental health: A plan of action for London. The Kings Fund. http://www.kingsfund.org.uk/sites/files/kf/field/field_publication_file/transforming-mental-health- london-kingsfund-sep2014.pdf; Care Quality Commission http://www.cqc.org.uk/ xlviii NHS London Strategic Clinical Networks. (2014). London Mental Health Crisis Commissioning Guide Perceptions of serious youth violence in London – causes and remedies Research report submitted to MOPAC by the University of West London (UWL) and ARCS Ltd PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT ii ACKNOWLEDGEMENTS The report was written by Mark Liddle and Prof Simon Harding, but the research project has very much a team effort, led by Simon at NCGR. Strong support for the fieldwork and other work strands was provided by Johnathan Green, Rachel Diamond, Junior Smart, Alex Babuta and Nsang Cristia Esimi Cruz (all at UWL). The authors would also like to express their thanks to the many hundreds of practitioners, residents and young people who took the time to share their views with us about violence and exploitation in London. We hope that we have done justice to some of the richness and diversity of their perceptions about these important issues. We are also grateful to other colleagues at UWL, ARCS, and at MOPAC and the London VRU. Their thoughtful feedback, support and advice was much appreciated throughout. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT iii CONTENTS ACKNOWLEDGEMENTS ...........................................................................................ii EXECUTIVE SUMMARY ............................................................................................iv 1 INTRODUCTION ............................................................................................... 1 2 METHODOLOGY .............................................................................................. 2 2.1 Definitions ......................................................................................................... 2 3 FIELDWORK, FINAL DATA-SET ...................................................................... 3 4 FINDINGS ......................................................................................................... 7 4.1 Perceptions of serious youth violence .............................................................. 7 4.2 Gang activity................................................................................................... 14 4.2.1 Girls and young women .................................................................................. 19 4.3 The view from schools .................................................................................... 23 4.4 Feedback from the case study areas ............................................................. 34 4.5 Violence, vulnerability and local context – a vignette ..................................... 41 4.6 Impact of COVID ........................................................................................... 44 4.7 Drivers ........................................................................................................... 46 4.7.1 Drugs ............................................................................................................. 47 4.7.2 Social media .................................................................................................. 48 4.7.3 Trauma .......................................................................................................... 50 4.7.4 Socio-economic factors, debt, exclusion and lack of opportunity .................. 53 4.7.5 Deterrence, lack of repercussions ................................................................. 56 4.7.6 Family factors ................................................................................................ 58 4.8 Protective factors, interactions....................................................................... 59 4.9 Addressing violence ...................................................................................... 60 4.9.1 Projects, interventions ................................................................................... 60 4.9.2 Partnership working ....................................................................................... 64 4.9.3 Policing .......................................................................................................... 72 4.10 Perceptions on how to reduce violence ......................................................... 78 5 DISCUSSION ................................................................................................. 80 6 APPENDIX – METHODOLOGY: FURTHER DETAILS ................................... 84 7 REFERENCES ................................................................................................ 89 PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT iv EXECUTIVE SUMMARY This Executive Summary provides key findings from our final report for the Perceptions of Serious Youth Violence in London research, funded by MOPAC. The research was designed and delivered by the University of West London (UWL), in partnership with ARCS Ltd. The research was designed to focus on the perceptions of key stakeholder groups across London about issues including: ➢ what they see as the key features of serious youth, gang or group violence and related offending in London, and what the impacts of that are; ➢ whether those key features have changed at all recently, and what the current trends and challenges are; and ➢ what is being done to address serious youth violence and related exploitation of vulnerable people, how effective people think those current efforts are, and what else might be done to improve or strengthen responses to these issues. The research involved several methods: ➢ a programme of consultation with practitioners and those working directly on issues concerning serious youth violence (involving direct interviews and wider online surveys), ➢ area-focused research involving collection of feedback from local residents and practitioners in a sample of areas experiencing high levels of serious youth violence, (via direct interviews and targeted online surveys), ➢ collection of feedback from “gang-linked” and “gang-active” young people (involving direct individual and group discussions), ➢ collection of feedback from Designated Safeguarding Leads at schools across London (via an online survey) about violence and victimisation among children and young people, and ➢ collection of feedback from detached workers involved in peer work with young people on the street (via two targeted online surveys), and from young people connected to football and other clubs across London (via 8 separate online surveys). The final data-set includes feedback from 605 respondents, made up of 392 practitioners from across several sectors (including police, local authority, education/schools, and the voluntary sector), 160 young people, 47 local residents, and 6 substance misusers. The bulk of the key work took place between January and August 2021, with some further analysis and follow-up undertaken toward the end of 2021. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD v KEY FINDINGS Perceptions of serious youth violence For all respondent groups, it was clear that issues concerning serious youth violence and exploitation were regarded as having a significant presence and impact. It was generally felt that these issues have become worse, and many of our respondents also suggested that there have been increases in the severity of violence recently. The feedback suggests that in many areas there are high levels of anxiety among young people regarding crime levels, risks and personal safety. In some areas young people discuss these issues on a daily basis, and share information through their own social and personal networks in order to manage and mitigate the perceived risks facing them. Stakeholders working with young people also commented upon the personal impacts of all of this on young people, referring to mental health impacts such as increased sleeplessness, trauma, and increased self-medication on high-strength skunk or other substances. Knife carrying for many young people in some areas now appears as being redefined and re-appraised as something that is a logical response to hostile landscapes, and these feelings of insecurity are matched as well by a perceived lack of safe spaces for young people. Gang activity The term ‘gang’ continues to generate controversy, but across the research the use of the term ‘gang’ was not as much of an issue for either stakeholders or young people as we expected. Some young people felt strongly the term ‘gang’ was negative towards young people in general, and some stakeholders also commented on the way in which the label can be applied much too readily to groups of young people who might just be gathering together in public. The term was also linked by some respondents to issues around disproportionality. Street gangs were said to be very active in all of our case study areas (which was expected given that we chose to focus on high violence areas), but also in areas surrounding schools in some areas that we surveyed. Feedback from some of our gang-involved respondents highlighted recent tensions between older and younger gang affiliates, which can give rise to violence that is less predictable (and more severe) than it was previously (especially in relation to control of county lines). Some street gangs have also proved flexible in terms of methods during COVID, and more generally, have in some cases moved closer to OCGs (organised crime groups). Feedback also suggested that a “survival” narrative or mode is now commonly expressed by those affiliated to street gangs and gang lifestyle - a sense that they will do whatever it takes to survive the streets and what life has given them. The feedback suggests that county lines are thought to equal employment for some young people. This is viewed as a realistic way of making money, as being viable, immediate and highly productive. Entry into county lines is easy, straightforward and quite normalised for some young people. Although it was noted above that money- making is an obvious key motivation for gang activity, some of the feedback that we received was more nuanced about what some young people actually “get” from PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD vi involvement in such activity, with reference being made to self-esteem for example, family-like relationships, and role models. Respondents who work with gang-affiliated young people told us that engaging with some of these issues is difficult, although we were also given examples of good practice. Girls and young women Respondents often referred to the sense in which girls and young women are “invisible” in much of the discourse about gangs and serious youth violence. There is clearly a need for further examination of the role played by girls and young women (and trajectories into and out of such involvement), since that involvement is of key importance to the way in which gang-related activities (such as drug distribution and county lines, for example) operate. Respondents commented on a range of issues concerning girls or young women affiliated with gangs – including the carrying or storing of firearms or drugs, sexual violence or exploitation, and engaging in violence – and most felt that these things took place fairly commonly, and that they now occurred more frequently than in previous years. There was a tension in the remarks made to us about the involvement of girls and young women in the activities of gangs, between perspectives that framed that involvement solely in terms of exploitation or coercion (framing their involvement purely as a safeguarding issue because most of them are traumatised and coerced into involvement), and those that attributed higher levels of agency to girls and young women who had such involvement. More general issues also came up about the sexual harassment of girls and young women, with these issues being commented on both by female research respondents in our surveys and interviews, and by practitioners who worked with girls and young women directly. The view from schools Schools were regarded by practitioners as being of particular importance to efforts to address issues around violence and exploitation, due to their key safeguarding functions, their role in building individual resilience, and the way that they usually operate as places of safety for children and young people. Feedback from the DSL survey highlighted a perception that some issues around violence and exploitation – i.e., bullying, child criminal exploitation, child sexual exploitation, involvement of young people in country lines, and missing children or young people - were fairly common in the areas where their schools were located. Along with most other practitioners, DSLs felt that levels of serious youth violence had generally increased in the area where their school was located, with about half claiming that levels had increased either a lot or a little over the previous 12 months. In terms of contextual factors within families in the areas where their schools were located – such as child abuse or neglect, child/adolescent to parent violence, domestic violence, family mental illness, family substance misuse, and family members suffering from trauma or being involved in offending – over half of DSLs felt that these factors were either “very common” or “common” in the area of their schools, with 81% making such a claim in relation to families suffering from mental illness. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD vii Over half DSLs felt that those issues had become worse in the last 12 months, with family mental illness and domestic violence in particular having got much worse (77% and 84%, respectively). The feedback from DSLs highlighted for us the way in which people in that role are in a unique position to watch young people’s life trajectories play out in some cases over several years, within the same family. There appears to be significant under-reporting of crime in and around schools. Feedback from case study areas In our case study areas we found a level of distrust of “official” surveys into local experience, which they felt would never be acted on or result in any change for them. It was also felt that some local residents were tired of having their areas labelled in the media and elsewhere as “high crime areas”. There were high levels of concern about violence, but also a strong sense of community at the same time in some areas. The research also generated evidence to suggest that in some areas, residents are very concerned about the impact of co-locating groups of vulnerable people in houses in multiple occupation (HMOs) which in some areas we found were clustered together in large numbers in very small local areas. That co-location of vulnerable groups in what are sometimes very geographically concentrated areas can create a dynamic critical mass of both drug dealers and users for example, which crystallises into a self-sustaining and resilient local drugs economy, but which also fuels a variety of forms of exploitation, and the violence which underpins those. Key factors Respondents’ comments about the key drivers of violence ranged widely by level of focus, covering factors at individual, family, peer, school, community and societal levels, but also covering drivers that span these levels. Across all respondent groups, drugs were most commonly mentioned as being a key driver of violence in London, although the precise way in which drugs might be linked to violence was not always articulated. A majority of respondents also made reference to wider factors such as deprivation, lack of opportunity, poor quality services or lack of access to services, family issues, social media, or the impacts of both individual and community trauma. The feedback illustrated how further research to “map” key factors in the context of individuals’ lives, and in the context of an individual’s pathway through and across those different levels over time, could be very useful for the field. Addressing violence The majority of respondents felt that primary prevention (i.e., work delivered “upstream” to prevent violence from happening in the first place) is important, and many respondents made reference in this context to factors having to do with poverty, lack of opportunity, and young people living in circumstances where family and other violence is normalised. Secondary prevention initiatives, which focus more specifically on groups that are at high risk of becoming involved in violence or exploitation, were also regarded as playing a key role in local violence reduction. Respondents also referred to measures that could be categorised as tertiary prevention, involving work with known PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD viii offenders to prevent further violence for example, or criminal justice measures to deter or sanction offenders.Many examples of preferred or “effective” interventions were provided by respondents, although views about the direction of future practice in how to reduce violence were also somewhat “downbeat”. Views about the importance of different types of interventions were strongly related to sector. Partnership working Feedback about partnership working in relation to how to reduce violence suggested quite a mixed picture across all London boroughs, and that the quality of partnership working varies widely in terms of extent, profile, and efficiency. Although “silo working” was occasionally referred to, there were also examples of good practice, and responses to questions about the quality of partnership working were fairly positive overall. Concerning the timescales for available funding, a majority of practitioners noted that short-term funding challenges are a real hindrance to local charities and projects, and it was also suggested that short-term funding eroded effectiveness. Policing Internal confidence amongst police staff appears to be quite variable. A view by many is that recent changes have diluted efficacy, efficiency and professionalism. Some longer serving officers felt that the MPS had changed significantly now and lost too many older, seasoned and knowledgeable staff above Inspector rank. It was noted by some police respondents that engagement with particular communities had suffered over the past couple of years in some boroughs. Feedback from young people concerning their views about the police was mixed, but fairly negative overall (and somewhat more negative than findings from MOPAC’s own survey of young people, for example). Among residents in our case study areas, confidence in the police also appears to be low, as reflected in both our direct interviews with residents and in feedback from our online surveys. Interview feedback and open-ended feedback from some residents who have responded to the survey, is quite negative in relation to experiences around reporting incidents of violence. Some respondents drew connections between disproportionality, racism and trauma at community level when asked for their views about current policing. Our feedback was fairly consistent with findings from other research that has focused on levels of confidence in the police across London. Perceptions of how to reduce violence Levels of confidence were quite low among many respondents, about the scope to reduce violence in London to be effective in the future. Feedback from residents and professionals working in our case study areas reinforced that view, with residents in particular reporting low levels of confidence that future work to reduce violence in their areas will lead to real change. In some residential areas, low levels of confidence about how to reduce violence was clearly related to a perceived history of authorities failing to act, and as many expressed it, failures to honour promises made to provide support or to regenerate areas and develop opportunities. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 1 1 INTRODUCTION This report summarises key findings from the “Perceptions of violence in London” research, funded by MOPAC. The research was delivered by the University of West London (UWL), in partnership with ARCS Ltd. The research was commissioned against a background of ongoing public concern about violence and its impacts in London, and it has been a Mayoral priority to take steps to understand and address some of the complexity of those issues and possible interventions to address them. As part of that priority MOPAC and the Mayor of London’s Violence Reduction Unit (VRU) have been overseeing a programme of research to try and address some of the gaps in knowledge that exist in the field. New research has included studies focusing on violence at neighbourhood level (Kane et al., 2022), violence by children and adolescents against parents (Brennan et. al., 2022), and the recently published Serious Youth Violence Problem Profile (MOPAC, 2022). Our research was designed and delivered as part of that programme. As described in the original specification for this research, MOPAC sought to “commission bespoke qualitative research to gather greater insight into all aspects of serious violence and criminality, and the multi-agency response by those most affected by it whether victims, offenders, community members or practitioners in the field.” So the research was intended to be wide-ranging in its focus, but also largely qualitative, and to collect and examine the perceptions of key stakeholder groups across London about issues including: ➢ what they see as the key features of serious youth, gang or group violence and related offending in London, and what the impacts of that are; ➢ whether those key features have changed at all recently, and what the current trends and challenges are; ➢ what is being done to address serious youth violence and related exploitation of vulnerable people, how effective people think those current efforts are, and what else might be done to improve or strengthen responses to these issues. It is worth noting that the research was not designed to somehow test people’s perceptions against what we know about violence from the official statistics – and to measure the distance between perceptions and “reality” (as complex and contested as the latter notion is in this field). We were interested in what people thought about violence and its impacts, about how serious they thought that some of the problems are and how they might have changed recently, and what they thought about the causes of violence and the effectiveness of measures to address it. Having said that, we do make reference to other research and official statistics where this seems appropriate, and where it can help us to understand and contextualise the feedback about perceptions. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 2 The key research methods used are summarised in section 2 below, followed by a description of the final data-set for the research, in section 3. Findings from the research are detailed in section 4, drawing from across the range of data sources described in section 3. 2 METHODOLOGY Although primarily qualitative, the research used a mix of quantitative and qualitative methods to gather feedback and other data. The study involved a number of key work strands, including: ➢ a programme of consultation with practitioners and those working directly on issues concerning serious youth violence (involving direct interviews and wider online surveys), ➢ area-focused research involving collection of feedback from local residents and practitioners in a sample of areas experiencing high levels of serious youth violence, (via direct interviews and targeted online surveys), ➢ collection of feedback from “gang-linked” and “gang-active” young people (involving direct individual and group discussions), ➢ collection of feedback from Designated Safeguarding Leads (DSLs) at schools across London (via an online survey), about violence and victimisation among children and young people, and ➢ collection of feedback from detached workers involved in peer work with young people on the street (via two targeted online surveys), and from young people connected to football and other clubs across London (via 8 separate online surveys). Although the focus of the research was primarily on people’s perceptions about key issues, the work also involved some collection and analysis of local and pan-London data and documentation concerning violence and related issues, and ongoing linkage of the work to the wider research and practice literature. The research received ethical approval from the Ethics Panel at the University of West London, and this included a review of all of the data-collection tools and supporting documentation used by the research team. 2.1 Definitions In our communications with stakeholders, residents and young people across London, we provided definitions both for the term ‘gang’ and for ‘serious violence’. Concerning the latter, although there is no standard definition of “serious youth violence”, respondents were advised that when we use the term in our questionnaires, we have in mind interpersonal violence that involves actual or serious risk of physical harm, violence involving knives, firearms or other weapons (e.g., acid), but also the threat of these things by individuals to ensure the compliance of others. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 3 We also made it clear that domestic violence was not a primary focus of our research, although it was obviously important to some of the issues that we did ask for feedback about (and we also did ask for feedback about this issue in our survey of Designated Safeguarding Leads, for example). We acknowledged that the term ‘gang’ is contested and subject to wide variations in usage, but we adopted the often used Home Office definition below: A relatively durable, predominantly street-based group of young people who (1) see themselves (and are seen by others) as a discernible group, and (2) engage in a range of criminal activity and violence. They may also have any or all of the following features: (3) identify with or lay claim over territory, (4) have some form of identifying structural feature, and (5) are in conflict with other similar gangs. We tended to use the term ‘gang affiliate’ rather than ‘gang member’ to refer to those connected to a street gang or offending group, while recognising that such affiliations are dynamic, changeable and might involve either a core or a loose connection. 3 FIELDWORK, FINAL DATA-SET The bulk of the key work outlined in section 2 took place between January and August 2021, with some further analysis and follow-up undertaken toward the end of 2021. Some further references to emerging literature were made during report drafting in 2022 prior to submission. The interview strands of the research were implemented as a “rolling” programme throughout that period, and although there were some issues concerning our gaining access to contact details for practitioners (which did complicate matters somewhat), that part of the research generated a unique and rich set of data, and had low levels of refusals or repeated “failures to show” (less than 5%). The impact of COVID19 during part of our work did mean that we were not always able to conduct in person interviews or discussion groups (except with some of the young people), but in our view this did not detract unduly from the quality of the material gathered. We give further details about sector coverage later on in this section. Implementation of the survey strands of the work took place mostly during the period from May – August 2021, with the practitioner survey being undertaken first, followed by the survey of Designated Safeguarding Leads (DSLs), the surveys of residents, and then the young people and “street worker” surveys. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 4 The way in which the young people surveys were implemented had the disadvantage of not allowing us to calculate response rates, and it also limits what we can say about how generalizable the findings are, but it had the advantage of allowing us to tap into small groups of young people who would be less likely to respond to school-based surveys, for example – because many of the youth projects that assisted us, work with young people who have a “street presence”, with some being regular truants, and others being already involved in gang activities. One strand of our young people surveys also focused on detached workers who do “peer to peer” street work with young people (in Southwark and in Waltham Forest), and this allowed us to tap into their views about the cohorts of young people that they work with (which number in the thousands according to sources). For the area-focused part of the research, the team selected a sample of areas across London for undertaking more focused fieldwork, and those areas were chosen largely because current data suggested that they were high violence areas, and in some cases also areas where levels of serious youth violence were increasing. The interview component of our area-focused work did not present any significant difficulties, but implementation of the resident surveys was more difficult than it has been in the past. Our survey work in these case study areas ran into problems with non-response from residents (see the Appendix for further details). In terms of primary data, the key interview strands outlined in section 3 generated a rich and useful spread of data, including transcriptions from: ➢ 97 interviews with practitioners/stakeholders from a range of public, voluntary, community and private sector bodies, ➢ two discussion groups with local multiagency teams (total of 17 participants), ➢ two group discussions with (6) substance misusers involved in treatment services, ➢ a group discussion of safety and violence with participants in a local “Youth Council” body (circa 25 participants), ➢ a group discussion with (6) young people connected with an “outreach” gym project, and ➢ 15 one-to-one interviews with young people involved in with local YOTs, or involved in local projects working with gang-affiliated or “at risk” young people. In spite of some of the difficulties referred to, the surveys described above also generated a useful set of feedback, including: ➢ 97 responses to a survey of practitioners, ➢ 181 responses to a survey of Designated Safeguarding Leads (DSLs) in London schools, ➢ 37 responses from surveys of “Streetbase”1 workers in Southwark and Waltham Forest, 1 Streetbase is a peer to peer street outreach programme, delivered in Waltham Forest by members of the Waltham Forest Youth Independent Advisory Group (and Waltham Forest Young Advisors), and was funded by MOPAC and the Council to run for three years from January 2019 (see https://youngadvisors.org.uk/2020/01/11/streetbase-young-people-listening-and-learning-from-each- other/). The Waltham Forest scheme was based on the Southwark Youth Advisors model (which was PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 5 ➢ 77 responses from a range of small group surveys implemented in youth clubs and at youth events, and ➢ 47 responses from residents in our selected case study areas. The final primary data-set provided the kind of coverage or breadth that we were aiming for, in terms of respondent groups, sector (for the practitioner consultations in particular, although we would have preferred higher numbers from the health sector), and in terms of response from all London boroughs. The final data-set includes feedback from 605 respondents, with that total being broken down by respondent group as summarised in Figure 1. Figure 1 – Total number of research respondents, by group In terms of practitioners specifically, across the survey, interview and discussion group strands, we gathered feedback from 392 individuals, across key sectors (see Figure 2). developed in 2006; see https://youngadvisors.org.uk/southwark/). Our survey respondents included Advisors from each of these schemes. 392 160 47 6 Practitioners Young people Local residents Substance misusers PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 6 Figure 2 – Practitioners consulted in total, by sector The 50 practitioners in the “Other” category were from a range of agencies/sectors, as summarised on the following figure.2 Figure 3 – Practitioners consulted: breakdown of “Other” (N=50) 2 The number for “Health” in Figure 3 is low, although 5 of our interview respondents recorded in the previous figure as “local authority – other” were in fact from Public Health. 185 22 24 68 43 50 Education/schools Local Authority – Community safety Local Authority - Other Police Voluntary and community sector Other 1 5 2 2 4 3 14 6 9 4 0 2 4 6 8 10 12 14 16 British Transport Police Business/commercial Consultancy Health Housing Job Centre Plus / Employment services Local Authority – Children's Services Probation/CRC Youth Offending Service Other public authority PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 7 The team also gathered a wide range of relevant documentation concerning violence and exploitation in London, and we were provided with some key data-sets that were not in the public domain (e.g., raw data-sets from violence-focused surveys conducted by other parties). All of this material was useful for us in understanding some of the issues discussed in this report (although we have not always made specific reference to them in the text). All of our feedback/data and documentation from all of these sources was imported into a large NVIVO project, which made all of the material available for qualitative analysis using conceptual “nodes”, for cross-tabulation with quantitative material (e.g., from surveys), and for the use of search, word mapping, and hypothesis- testing tools. Survey data-sets were analysed using SPSS as well. Further details are provided in the Appendix. 4 FINDINGS Findings from the research are summarised below, under a set of headings linked to our key research questions and to those used in our data-collection tools. 4.1 Perceptions of serious youth violence Across all strands of the research, we were interested in developing a picture of what particular respondent groups thought were the key issues concerning violence and exploitation in their own areas (or the key issues across London more generally), and also whether they thought that these issues had changed at all recently. For all respondent groups, it was clear that issues concerning serious youth violence and exploitation were regarded as having a significant presence and impact. It was also generally felt that these issues had not improved recently, and in fact, most respondents across all groups felt that the issues had become worse. For practitioner survey respondents for example, 72% (n=96) felt that levels of serious youth violence had increased either a lot or a little over the previous 12 months3 (see Figure 4), and among those who felt that levels had either stayed the same or got slight better, most felt that this was only due to the impact of COVID restrictions (which had meant that there were fewer opportunities for public violence to occur). 3 Responses to the surveys were received mostly during the period from May – August 2021, so the previous year for those respondents would have been the 12 months prior to that period. Part of that previous one year period included quite a complicated chronology of COVID restrictions across the country and in London specifically (for a useful graphic summarising all of these restrictions, see https://www.instituteforgovernment.org.uk/sites/default/files/2022-12/timeline-coronavirus-lockdown- december-2021.pdf). Those restrictions during some periods will have had some impact on offending and related issues (and also perceptions of recent trends that were communicated to us). We return to some of these issues in section 4.6. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 8 Figure 4 – Practitioner survey: “Would you say that overall levels of serious youth violence over the last 12 months have:” (% by response category) (N=96) Responses to a similar question in our Streetbase survey provided something of a contrast, with just over a third (13 out of 37 who responded to this question) suggesting that levels of serious youth violence had “decreased a little” in the areas where they worked with young people. A similar number suggested that levels had either increased a little, or increased a lot. Again, many of the respondents who felt that serious youth violence had decreased, noted that COVID restrictions during the period had been the cause of the reduction. Of the 12 respondents who commented on reasons for an apparent decrease, 8 attributed this reduction at last partly to the impact of COVID restrictions, as in the following comments: I believe isolation and the simple act of keeping young people inside has limited the amount of incidents while I've been working (SB-171797910) I think [during] lockdown say incidents decrease but since the lifting of restrictions they have increased again slightly. (SB-171366977) Lockdown restrictions meaning less activity on the road. (SB-171811326) People haven’t really been allowed to go out much with COVID-19 so as much as they are getting frustrated it is probably being taken out at home rather than outside. (SB-171333732) 39% 33% 11% 8% 5% 3% 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% Increased a lot Increased a little Stayed the same Decreased a little Decreased a lot Don't know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 9 Also concerning recent trends, 36% of respondents to our aggregated young people surveys felt that levels of serious youth violence had either decreased a little or a lot during the previous 12 months4, as compared with 31% who felt that levels had increased either a little or a lot. Figure 5 –Young people survey: “Over the last 12 months, would you say that levels of violence in the area where you live have:” (% by response category) (N=64; not answered=13) Many of our respondents also suggested that there have been increases in the severity of violence recently. The following practitioner survey respondents commented on this specifically: The alarming aspects are the ages of those concerned and the ferocity and frequency of attacks. (PRAC-79272177) The amount of violence does not appear to have increased. It is the level of violence that has increased (more fights with weapons etc.). (PRAC-79678632) Those affected are increasingly younger, the violence is increasingly more serious or extreme. (PRAC-80247515) A similar view was taken by some of our young interview respondents. One described the trend in this way: 4 See note 3, above. 8% 28% 20% 14% 17% 13% 0% 5% 10% 15% 20% 25% 30% Decreased a lot Decreased a little Stayed the same Increased a little Increased a lot Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 10 Physical fighting is not really around anymore, it used to be a long time ago but I don’t feel like now. People don’t really fight with their fists now, it’s all with weapons. (A-YP03) Concerning young people specifically, it was clear from all of the feedback that violence and related issues were of great interest and concern to many young people across London. Although this did vary widely across areas, the feedback suggests that in many areas there are high levels of anxiety among young people regarding crime levels, risks and personal safety, especially among those of secondary school age and upwards. In some areas young people discuss these issues on a daily basis, and share information through their own social and personal networks in order to manage and mitigate the perceived risks facing them. Comments such as the following were often made by young survey respondents: Very concerned. I fear for my life, friends and most importantly my family. My young brother and myself getting into the wrong group to make friends can lead us to fatality or nowhere in life. (YP-168472057) I have younger brothers and sisters and they could be a victim of crime. This concerns me a lot. (YP-168472087) I try to avoid certain roads in the evening. . . . I would never let female members of my family walk to places in the evening. (YP-168351220) Those day-to-day conversations that many young people engage in with their peers often involve the sharing of “war stories” about violence and crime, and again in some areas, there are many young people who claim to know someone who has died from a stabbing or other violence. There appeared to be a degree of bravado behind some of the remarks made to us by young people (perhaps especially by groups of young people), and it was not always possible to judge how exaggerated such claims might have been, but this does not detract from the fact that it was a key topic in day to day conversation (and that such conversations about violence took place in areas known to have high rates of serious violence). Many young people interviewed reported that fear of crime restricts their movements throughout their community and determines the choices they make about routes home, socialising, engagement with friends and family, sports, and so on. Others made reference to the impact that concerns about violence had on them: It makes me stressed. I feel that something could happen outside the house or if I’m walking home or leaving the house, something could happen, just random. Crime happens so you never feel 100% safe anyway. (A-YP08) PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 11 Stakeholders working with young people also commented upon the personal impacts of all of this on young people, referring to mental health impacts such as increased sleeplessness, trauma, and increased self-medication on high-strength skunk or other substances. Knife carrying for young people in some areas seemed to be regarded as being a logical response to hostile landscapes. It was seen to offer instant and real protection, but descriptions also suggested that it was seen as a method of empowerment, and something to rely upon. It was also thought to be a necessary and expected accessory which designated knowledge of the streets and that you must be taken seriously – implying that you are not vulnerable and should not be victimised for lack of awareness. For some it clearly provided a “passport to move around” different estates and parts of their borough. The comments made by the following respondent were not uncommon in our interviews with young people, where we focused on concerns about violence, and about steps taken by young people to avoid violence or protect themselves: Interviewer: So do people in this area carry weapons sometimes? Respondent: All the time, knives and that. Interviewer: And, do they actually use them? Respondent: Yes. Interviewer: Have you seen people using them apart from that one time that you saw, or do they threaten with them? Respondent: I’ve seen people get threatened with them, I’ve seen people get chased with them, I’ve seen people get stabbed with them. Interviewer: Would you carry one? Respondent: A weapon? Interviewer: Yes. Respondent: Yes, if I needed to. Interviewer: Have you ever felt you needed to? Respondent: At one point on the estate, yes, the estate was so bad everyone used to have a knife in their room just in case anyone ran up on the house. (A-YP04) Some of the evidence from other research also resonates with feedback of this kind. A recent review of the literature on knife crime suggested that there are several different groups of individuals who carry knives, and the authors describe the way in which knife carrying can be underpinned by a mixture of factors having to do with individual and social adversity (where the weighting of these is different for different sub-groups; Browne et al., 2022). They also note that only a small proportion of those who carry knives, actually use them to attack or threaten others. The feelings of insecurity expressed in some of our feedback were matched as well by a perceived lack of safe spaces for young people, a situation that has been made worse by the closure of many youth clubs and other services for young people in recent years (and during a fairly significant period of austerity). Some interview PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 12 respondents commented at length on this lack of provision. As one voluntary sector representative put it: I would try and reinstate some of the things that are missing or have been taken away from the borough. So there are very few safe use spaces left for young people in [this borough], and so again, where possible, kind of collaborating with young people in the community to set up those spaces feels really important. (B121) It was noted by some respondents however that the provision of “safe spaces” would not on its own guarantee take up by young people, although the perceived safety of facilities and services is clearly a prerequisite for usage of those services. The provision also needs to resonate with the needs and interests of young people, and in particular to be age-appropriate. Some older teenagers noted that they would not be interested in attending a traditional “youth club” if it catered to the interests of younger people, for example. Some young people articulated a sense of being on their own, and that there was an inevitability of outcomes, i.e., whatever you are thinking will probably happen to you as it is inevitable. For some, a sense of fate was reflected in their comments about their perceived life choices as well, with young people occasionally suggesting that they had no real opportunities, and that they had to survive as best they can in whatever way they can (which often includes illegal activity). Practitioners who work with young people pointed out how perceptions of this kind about “lack of choice” left some young people particularly vulnerable to exploitation. That sense of fatalism was also conveyed in some of their comments to us when we asked them what they thought could be done against violence, for example, with many suggesting that there was nothing that could be done. I don’t know any more I think it’s too late to stop it. (YP-169767124) I’ve not long lived here but there’s nothing to do . . . but to be honest nothing will stop the violence. (YP-169883901) Nothing can be done there will always be a way around it or people who are affected in some way where the process of hate towards another kind will continue unless you can stop people from being jealous or hating others you can’t stop crime or reduce it. (YP-170745285) There is a perception among some young people that adults do not understand the situation regarding crime, risks and street life, and the pressures regarding social media, county lines, drugs, knife crime and violence that form part of their own experience. The feedback suggests that young people often do not discuss such issues with their families and very seldom with teachers, although as noted elsewhere in the report, our consultations with young people tended to focus on PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 13 those in areas with high violence levels, or on forms of youth support that catered to more marginalised groups. Some young people have felt their future lies more securely with criminal behaviour as this is a life that is known to them and has clear and understandable rules which are familiar to them. It can also bring in fast monetary returns which are untaxed and undeclared - although some of the feedback from young people suggested that their expectations about this were quite unrealistic – a fact that is sometimes focused on by projects that try to promote better informed decision-making among young people about involvement in gang activity (e.g., school-based awareness-raising interventions delivered by groups such as Growing Against Violence, and street- based work delivered by staff members who have “lived experience” in relation to offending groups and violence). Some of those young people who are involved in criminal activity and have been for a while have obviously travelled some distance from ordinary or “normal” society, and the way that they speak about themselves and their lives illustrates both that level of disassociation, and a sense that there is little or no scope (or interest) in developing such ties. This immersion or embeddedness demonstrates commitment to a chosen career path whilst reducing the stress of switching between the criminal lifestyle and “normal” society, i.e., they simply stay embedded in the rules, manners and presentation of gang life or criminal life on a permanent basis and no longer seek to play a role or participate in wider civic society. This process of “going deeper” or “staying true” can enhance trust amongst criminal associates by displaying dedication to the criminal world. This orientation towards crime, drugs and gangs also seems aspirational for some young people – they see few opportunities in wider society for themselves. Others have aspirations of college and work but this is often said to be “long”. Findings from our survey of detached street workers (the “Streetbase” survey) also suggested the presence of concern about violence among young people worked with on the street. We asked workers about levels of concern about violence that there might be among the young people that they engage with, and just over two thirds of them (25 out of 37) answered either “most of them” or “some of them”. Street workers told us that this concern can be gauged both in terms of what participants actually say to them on the street, and in terms of the way in which the young people actually change their behaviour as a result of such concerns (in other words, we were not inviting respondents simply to guess about what project participants might be thinking about). In relation to the latter, we also asked the workers what they thought the impact of concerns about violence might be, and they (27 respondents) offered a range of useful comments. Four respondents commented specifically on the carrying of weapons, for example, making comments of the following sort: PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 14 Some young people become very paranoid of people they see and the surroundings they place themselves in. This often causes them to carry illegal weapons due to the feeling of being unsafe. (SB-169106890) They have to carry a weapon to feel safe. (SB-171865444) Changing routes to/from school etc. Carrying weapons for safety. No attendance of public spaces: parks/schools. Lack of trust in services. (SB-171709318) Feedback from practitioners who work directly with young people confirmed the widespread presence of those kinds of concerns about safety, with some also linking them to specific reactive behaviours by young people. While most of our young respondents commented on the more negative impacts of violence on themselves or on people close to them, some commented more broadly on the relative acceptability of violence and its potential role in their everyday lives. One of our interview respondents (a 14 year old girl at a youth project in one borough) expressed a view of violence which was striking for its apparent balance: Violence, I think it’s a very necessary thing, people think it’s stupid and they don’t think it’s necessary but I think it is. I think it’s important to have a little bit of violence sometimes because I think violence can be used in the right way. If someone is taking the piss out of you, you should be able to retaliate, and if someone is bullying you or if someone is fighting you, you should fight back, you have the right to fight back. But I think it’s stupid when you’re fighting over postcodes and things like that and you’re hurting innocent people or pedestrians, civilians, people who aren't involved in things or people’s family members, that’s when it gets too far. (A-YP-04) It was also clear from the feedback that in many areas, young people’s “threshold of tolerance” for violent incidents was quite high, and that violence was normalised in their perceptions of community (and family) life. 4.2 Gang activity As noted in the introduction we did question our respondents about gangs specifically, and as noted in section 2 we did use a fairly standard definition of the term in our documents and questionnaires. The term has always generated controversy (and a prodigious literature concerning the term and its use has also developed)5, but across the research the use of the term ‘gang’ was not as much of an issue for either stakeholders or young people as we expected. Whilst young people have a range of alternative slang names or street 5 Some useful reviews of this literature can be found in Francis and Wilcox (2013), Paoli (2014) and Andell and Pitts (2023). PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 15 names they often accepted and recognised that they were affiliated to such groups. Some young people preferred ‘Fam’ implying a family-like bond of trust. Older Gang Consultants6 and some stakeholders noted that whilst ‘Fam’ was often used, the street groups or gangs seldom operated like families as allegiances and relationships were often short-lived, variable and usually motivated by money. Some young people felt strongly the term ‘gang’ was negative towards young people in general, and some stakeholders also commented on the way in which the label can be applied much too readily to groups of young people who might just be gathering together in public. In general, the term ‘street gang’ or ‘gang’ is widely used by young people, practitioners and residents, but there is often some ambivalence about such usage. Some voluntary organisations that work with young people advise caution about such usage, even where they might use the terms themselves in their own work. Abianda, for example – an organisation that offers support to girls and young women affected by gangs and county line activity – advises professionals to use the term “cautiously, being mindful of its history, classist and racist connotations” (Abianda, 2021:2). They do however use the definition of ‘gang’ adapted by the Children’s Commissioner, which is almost identical to the one that we have used in the research (provided above, in section 2.1). Ethnicity of gangs and groups was occasionally raised with the generalised comment that street gangs can involve any ethnicity or group and was certainly not just about “Black people”. Some stakeholders working with young people noted that young people were far less likely to exclude on the basis of race or ethnicity and that some of the street gangs had affiliates from a range of different ethnic groups. Some also felt that street gangs were becoming much more multi-cultural and that this possibly reflected local changes with the arrival of immigrants or new arrivals from Europe and elsewhere. In terms of other feedback we did ask Designated Safeguarding Leads (DSLs) to comment on whether “street gangs” or organised offending groups operated in the areas where their schools were located, and two thirds of respondents had a view on this, as summarised on the following figure. 6 We use that term to refer to gang-linked individuals already known to NCGR (the National Centre for Gang Research, at the University of West London), who we consult about key issues on a regular basis in connection with a range of research projects that we are involved in. These individuals are usually compensated for discussing these issues with us, and feedback sessions are not always recorded. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 16 Figure 6 – DSL survey: Do “street gangs” or organised offending groups currently operate in the area where your school is?” (% by response category) (N=180) DSLs were further asked about the extent to which more specific issues related to gangs took place in their areas – such as young people being used to carry or store weapons or drugs for others, for example - but it was clear that their level of knowledge about such activities was fairly low (with from 40-90% of respondents answering “don’t know” to those questions). Feedback from some of our Gang Consultants and other young people suggested that a “survival” narrative or mode is now commonly expressed by those affiliated to street gangs and gang lifestyle - a sense that they will do whatever it takes to survive the streets and what life has given them. Under this survival narrative, almost anything can be accepted as permissible or forgivable “on Road”. In this way the threshold of what is a credible action to take seems to be lowered, although judgements about this are obviously highly subjective and variable. This narrative also acts as a technique of neutralisation of criminal behaviour. Several Gang Consultants talked about why street and youth violence had changed and become more vicious and more present. They all mentioned that Youngers are often seen by Elders and Olders7 as being “out of control”, i.e., they will no longer be told what to do by Olders and will act on their own or actively defy the line taken by Olders. Violence can erupt as a result of this. Their actions are at times more instantaneous, less thought-through, and more impulsive. As such, violence can be more eruptive and fierce (and also less predictable). 7 There is some variability in the usage of these terms across areas, but Youngers are generally referred to as being aged 12-16, with Olders being aged from 16 to about 23, and Elders over about 23. 36% 11% 54% 0% 10% 20% 30% 40% 50% 60% Don’t know No Yes PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 17 As one respondent put it: It’s more of a lawless society where the Youngers now have taken over. . . .The Elders have no control anymore. They don’t have control anymore. The Youngers are on our streets. The Elders are very, very cautious before they jump in their vehicles, before they go to the shops, before they go splashing out. The Elders are not people who now broadcast their purchases. They’re not out there showing off their Rolex watch in the same way that they used to. . . .The fear of being attacked by Youngers who don’t respect their credibility or merit. (A-FG-PROJ3-01) High levels of stress/trauma/drug use/anxiety are now identifiable and prevalent amongst many young people who are involved in or adjacent to these lifestyles. For some young people they address this by slowly withdrawing from their family or from wider society and increasing their social isolation. The feedback suggests that county lines are thought to equal employment for some young people. This is viewed as a realistic way of making money, as being viable, immediate and highly productive. Entry into county lines is easy, straightforward and quite normalised for some young people. Some young people reported that street gangs are now less focussed on their “patch”, “turf” or territory than before and more focussed on their drug dealing “line” or business. Youth workers often report a shift towards making money as opposed to “defending the turf”. Urban street gangs operating across the capital were said by many young people and by stakeholders to be highly flexible in their dealing, links, their MO, their success and their operations, and some said that they have also increased links and closer links to OCGs (Organised Crime Groups). That “flexibility” was illustrated by rapid shifts in tactics seen during COVID restrictions for example (see 4.6). Profit margins (which may have increased at some points during COVID restrictions) appear to have become more closely protected and monitored. 8 Tighter control of county lines and localised drug dealing was identified as a driver for violence with Olders and Elders keen to keep Youngers in line. Similarly, debt was a major concern for young people as this was almost a guarantee of potential violence (we return to issues concerning debt in section 4.7.4). Increased levels of fraud were noted with young people being actively and regularly pressured into holding illegal funds (from criminal gains) in their bank accounts. This can often amount to thousands of pounds in one go. Again, this scenario can generate stress but also increased violence. Such financial transactions take place 8 It was suggested to us by several respondents who work in the drugs field that COVID restrictions did lead to some drug supply shortages over the short term, and that this led in turn to some price increases. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 18 without the knowledge of parents or teachers and remain hidden and exclusive to young people. Although it was noted above that money-making is an obvious key motivation for gang activity, some of the feedback that we received was more nuanced about what some young people actually “get” from involvement in such activity. One respondent (the leader of a local voluntary sector project working with marginalised young people) described an individual case that they were aware of: You know, that, I knew a young person that was a good kid and I told him, “Why are you hanging around with these negative Olders?” And he was like, “Oh [named worker], I’m just out here man, just trying to make money”. But he’s not really trying to make money, he found self-esteem, he found like family, he found a role model, he found like a father figure, that’s what he found, he didn’t find money. Because and then he – I told him, “You can’t be out with these guys”. (B-014) Other practitioners working with gang-affiliated young people raised similar issues to those expressed by the respondent above, concerning cases where although a young person might be involved in offending or other gang activities, they seemed to be prospering in many key respects, including having a sense of pride in having what they (i.e., the young people) perceived to be effective skillsets, being respected and watched out for by peers, and even having a sense of achievement and inclusion. A few respondents who described these kinds of cases to us also expressed uncertainty about how to proceed and to engage with such people – one asked for example, what his project could offer to compete with those kinds of benefits. Others who mentioned such cases adopted an approach focusing on longer term possible futures, and workers who themselves had “lived experience” of gang activities described how they would work with gang-affiliated young people to help them think about where their current pathway might lead them to be in 5 or 10 years for example. Since the answer to that question is often “dead or in prison”, as one worker put it, either the young person will be attentive to those more negative possible outcomes and make decisions to do something about it now, or they will simply not engage with such interventions – although they may well do so later on, as that sort of change is not usually linear. Feedback of that kind does raise important practice issues about approaches to take when interventions are intended to extricate young people from gang involvement. Interventions that simply take a “gangs are bad” kind of approach without dealing with some of the complexities of gang involvement are probably less likely to engage with people in a way that has much scope for effecting change or transition.9 9 An analogue here can be found in the drugs education field, where awareness-raising programmes that focus only on harms that drug users may experience - and ignore the fact that some drugs can provide positive experiences which young people are aware of and which they might value – are less likely to be effective either in terms of prevention or harm reduction (see ACMD, 2015, for a useful overview of the evidence). PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 19 On the other hand, the work of some of our practitioner respondents (especially a few in the voluntary sector who work directly with young people involved in violence or exploitation) does illustrate that it is possible to adopt a nuanced and flexible approach to these issues in spite of such difficulties, although our feedback also makes it clear that even the most effective approaches to engagement will not be successful with everyone.10 4.2.1 Girls and young women Respondents often referred to the sense in which girls and young women are “invisible” in much of the discourse about gangs and serious youth violence, and the way in which there are often no obvious referral routes for interventions with girls and young women who are gang-affiliated. As one respondent from Probation put it: I think, I mean throughout when we were still, and as we are still slightly stuck in the old gangs narrative, we saw very few girls that were being flagged as, you know, part of a street culture, or violent street culture, sort of set up. . . So I think they remain pretty invisible, and I think whilst we focus, you know, we have this focus on this young man who is the bearer of all risk, we focus on him because we’re in the business of offender management. So the challenge is, isn’t it, that we’re all working on is to widen the focus out to sisters, girlfriends, nieces, children. So we, I don’t feel I have a very, much visibility, actually, about how the borough’s women and girls may be networked in with the young men we supervise. I'm certainly not aware of any specific referral pathways for 18 - 25-year-old women for example, beyond generic services. (B-008) Although we found pockets of expertise in London in terms of knowledge about this area, there is a clearly a need for further examination of the role played by girls and young women (and trajectories into and out of such involvement), since that involvement is of key importance to the way in which gang-related activities (such as drug distribution and county lines, for example) operate.11 One local authority gangs analyst highlighted some of the issues in his own borough: Yeah, so in [this borough], I don't think we have what I would describe as a single female gang member. For the most part, what we’re witnessing is girls being exploited. So we’ve never seen girls being treated as equals in the gang. So I’ve been, so if we ever girls in the music videos, not being, like, not 10 An example of a balanced approach to some of these issues is the multi-faceted intervention described in Davies and Dawson (2022), in their evaluation of the London Gang Exit project led by Safer London. 11 A useful recent discussion of the “invisibility” of young women and girls’ involvement in street gangs and county lines activity in a London context has been offered by Havard et al. (2021). The authors outline a range of tactics used by street gangs to maintain “coercive control” of young women and girls in physical and digital space. They argue that vulnerability to exploitation needs to be understood contextually, and as involving a more nuanced notion of individual agency than is usually adopted in more traditional enforcement-led responses to gangs and county lines. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 20 being exploited for their bodies but actually maybe rapping at the front of the group or they’re actually actively involved in controlling a drug line or something. Then I say yes, we can call these girls gang members and consider them equals to their male counterparts. At the moment, I think they’re just more victims of exploitation by the slightly older male gang members within that gang or group. I don't know if that stretches to other parts of the borough but for [this borough], I think that’s definitely the case. (A-091) It was also noted that the involvement of women and girls in gang activities was quite variable across London by borough or smaller area, and that in some cases this was at least for reasons having to do with culture or ethnicity: [Women and girls are] completely absent. This is all young men. And again I think that’s a cultural thing because young men have lots and lots of freedom, young girls don’t. One of my very best lifelong friends is Bangladeshi female, and she’s got six brothers and one sister, and she always says like, you know, they could just do what they wanted, whereas her and her sister were absolutely, you know, locked up. And that’s a cultural thing that men are able to have more autonomy. And I think that plays a part, if you’re talking about the demographic, that certainly plays a part in why we don’t see girls involved, like we do in Hackney or Brixton, girls are involved in gangs. And, you know, in some situations are seen in very senior roles. But you just don’t see that [in this borough]. (A-078) Our practitioner survey also asked respondents to comment on the extent to which a range of more specific activities involving girls and young women affiliated with gangs took place. A total of 69 practitioners responded to this question; responses are summarised at Figure 7. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 21 Figure 7 – Practitioner survey: “To what extent do the following take place in relation to street gangs/offending groups in your area?” (% by response category) (N=69; percentages of zero are not shown in the Figure) The feedback summarised in the above figure suggests that practitioners tend to think that girls and young women are in fact involved in gang activities in a range of ways, but this is not necessarily inconsistent with the view that was also commonly expressed, about the “invisibility” of such involvement. Part of the overall point made by some respondents was that gangs are traditionally viewed through a “male offender” lens, and that it might be more useful to adopt a focus on gender more broadly when looking at gangs and when addressing issues concerning gang activity. Practitioners were also asked whether they thought that there had been any changes in the involvement of women and girls in the activities of street gangs or offending groups over the last 12 months12, and 70 practitioners responded. Just under half of 12 See note 3, above. 19% 19% 13% 12% 12% 33% 1% 1% 6% 9% 13% 6% 1% 7% 29% 30% 38% 46% 35% 30% 46% 43% 36% 35% 51% 29% 0% 10% 20% 30% 40% 50% 60% Sexual exploitation of girls or young women affiliated with gangs Sexual violence against girls or young women affiliated with gangs Physical violence against girls or young women affiliated with gangs Girls or young women affiliated with gangs committing violence/criminal activity Girls or young women affiliated with gangs carrying or storing drugs Girls or young women affiliated with gangs carrying or storing firearms Often Sometimes Seldom Never Don't know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 22 those respondents were not sure whether there had been any changes, but 16% claimed that there had not, with 37% claiming that there had. Those respondents who felt that there had been a change were then asked to comment on that change and the possible reasons for it, and all of them offered some comment on this (N=26). All of these respondents felt that at least some of the issues listed in the main question had become worse over the past year, and cross- tabulations showed that these respondents were involved directly in work focusing on exploitation or gang activities (with the majority being experienced police officers having those roles). Hence, they were in a position to comment from an experienced standpoint, although they obviously did not all comment on the same trends. About a third of them felt that females were becoming involved in gang activities more often than previously and that they were also starting to do so at a younger age. In the DSL survey, respondents were given the opportunity to skip the section on girls and young women, and 72% did so. Among those who did respond, there were still high levels of “don’t knows”, and so we have not included a table such as the above one for practitioner respondents. Some of the DSLs did offer useful open- ended comments on the involvement of girls and young women in the area where there school was located however. Fourteen DSL respondents suggested that there had been changes in these issues over the last 12 months, and 12 of those respondents provided further feedback about those perceived changes. About half of them referred to increased levels of risk, to an increase in numbers of cases where exploitation was an issue, or to a decrease in the age of girls who were becoming involved in either CSE or CCE. One suggested that female parents in the area were gang-affiliated, and another suggested that there had been an increase of female on female violence in the area. As some of the above comments already suggest, there was a tension in the remarks made to us about the involvement of girls and young women in the activities of gangs, between perspectives that framed that involvement solely in terms of exploitation or coercion (framing their involvement purely as a safeguarding issue because most of them are traumatised and coerced into involvement), and those that attributed higher levels of agency to girls and young women who had such involvement (they actually know what they are doing, and they make their own autonomous decisions about whether and how to be involved). Issues concerning the problematic nature of “consent” among women and girls involved in gang activity have also come up in previous research, where those involved might talk about their involvement as if it is a result of their own autonomous decision-making, but where an outside practitioner might see it as survival in a context of coercion.13 As one police respondent expressed it: 13 Some of the examples provided by respondents are quite similar to those uncovered in previous research – e.g., in some of the feedback provided by practitioners involved in previous work under the Ending Gangs and Youth Violence initiative, as reported in Disley and Liddle (2016). PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 23 It is a really tricky situation for the police to protect these females as they often do not feel like they are victims and pick up on anti-police sentiment from the older males. I think more joint working with charities/more education needs to be done around this. A lot of (male) police officers are drawn into the thinking that if the girls appear to be consenting to the activities then they are suspects and it is all through their own choice and this should really be addressed. (PRAC-79200356) More general issues also came up about the sexual harassment of girls and young women, with these issues being commented on both by female research respondents in our surveys and interviews, and by practitioners who worked with girls and young women directly. As is the case nationally (and as recent events such as the murder of Sarah Everard have brought sharply into the foreground for many)14, this kind of regular harassment was felt to be very common – even a daily experience for some girls and young women. It was also felt that the issue was seldom addressed directly, and that there was little sense of any accountability. One of our young interview respondents described her own experience being “cat- called” by construction workers, for example: Yes, because those construction workers are so creepy, I literally walk this way to school now, they’re so creepy, they’re always watching me and I don’t like it, it’s so uncomfortable. And, the amount of times I get catcalled, even when I’m dressed and completely covered I still get catcalled, even on the days I feel my most ugliest I still get catcalled, it does not make sense and it just shows you it does not matter what you wear, it can happen to anyone. But, yes, I have felt the need to carry a weapon but not all the time. (A-YP06) The idea that it can happen to anyone resonates with findings from other research about the pervasiveness of women’s experiences of harassment across categories of class, ethnicity and age, but it was also highlighted in the remarks that we gathered, that there is often a sense of fatalism about such behaviour – a sense that it’s not worth complaining about, because it’s always happening and it’s going to continue happening anyway. 4.3 The view from schools Schools were regarded by practitioners as being of particular importance to efforts to address issues around violence and exploitation, for several key reasons: First, schools provide opportunities for qualified staff to identify young people who are being victimised or exploited, and to liaise with key authorities about 14 And also perhaps the Child Q case (see (Gamble and McCallum, 2022, for an overview), which highlighted issues not only about the safety of young women and girls, but about ethnicity and the “adultification” of young Black girls – which involves taking a view of them as possible offenders and suspects, rather than as children for whom a safeguarding approach might seem more justified. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 24 interventions to enhance and support children’s safety. Second, schools were described by practitioners as being key sites for delivering interventions that can build individual resilience among children and young people, which can allow them to resist factors that otherwise threaten to pull them into activities that threaten their safety and wellbeing. Third, it was felt that schools can function as places of safety for young people, and particularly for young people whose concern for their own safety is chronic and linked to hypervigilance which in turn is linked to previous adverse experience. A few respondents also commented on how sustained hypervigilance and related impacts from previous trauma or victimisation can have profound developmental implications for young people, and can also have adverse effects on behaviour and attitudes over time – comments which also resonate with what we know from other research about these impacts. Because of this significance, the research team felt it would add value to the study to gather feedback about some of the key issues from Designated Safeguarding Leads (DSLs) at schools across London, as referred to in Section 2. We also felt that feedback concerning (and from) schools should be presented in a separate section, even though there are obvious overlaps with issues presented in other sections. Our survey of DSLs provided a unique strand of data for the research, and involved useful coverage across areas of London and by school type. Further details concerning the survey and responses to it are provided in the Appendix.15 In terms of the current context, feedback from the DSL survey highlighted a perception that some issues around violence and exploitation were fairly common in the areas where their schools were located, although this obviously varied considerably by issue (and by each respondent’s level of knowledge about that issue, with many respondents for some issues answering “don’t know”). 15 A national survey of DSLs in providers in the post-16 education and skills sector is reported on in Education and Training Foundation (2021), which did focus on some of the issues that we asked about in our own survey. A national survey of DSLs in schools was also undertaken while we were writing up our own research, and it focused largely on workload, policies, training, and the perceived effectiveness of ICT usage monitoring (a brief summary can be found at: https://www.judiciumeducation.co.uk/news/Newsflash-Schools-find-key-safeguarding-areas-highly- challenging-according-to-new-Judicium-survey). Our own survey is unique in having a specific focus on London DSLs from primary and secondary establishments (along with a few post-secondary), and also focusing on perceptions of both key safeguarding issues and their occurrence, and key local “drivers”. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 25 Figure 8 – DSL survey – “Regarding violence and exploitation, how often do the following incidents normally take place in your school or surrounding area?” (% by response category) (n=181) DSL were also asked for feedback about changes in a range of more specific violence and exploitation issues, but there were high levels of “don’t know”s in their responses, so we have not reproduced the figure here. A quarter of the items had more than 50% of respondents answering “don’t know”, and for several of these the level was above 90%. DSLs also felt that levels of serious youth violence had generally increased in the area where their school was located, with 50% claiming that levels had increased either a lot or a little over the last 12 months, and 26% claiming that levels had stayed much the same (and 21% claiming that they did not know). 16% 21% 8% 3% 2% 13% 9% 9% 48% 54% 24% 25% 9% 24% 26% 34% 27% 20% 27% 30% 20% 24% 24% 30% 4% 2% 25% 25% 46% 27% 25% 17% 6% 4% 16% 17% 22% 12% 16% 11% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% Physical violence involving children or young people Bullying (including cyberbullying) Child criminal exploitation (CCE) Child sexual exploitation (CSE) Children or young people carrying firearms Children or young people carrying knives Involvement of young people in county lines Missing children or young people Often Sometimes Seldom Never Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 26 Figure 9 –DSL survey: “Over the last 12 months, would you say that overall levels of serious youth violence in the area where your school is have . . . ” (% by response category) (N=178) In terms of contextual factors within families in the areas where their schools were located, DSLs were asked to comment on how common or otherwise some of these factors are. Their responses to those questions are summarised at Figure 10. 0% 2% 26% 34% 16% 21% 0% 5% 10% 15% 20% 25% 30% 35% 40% Decreased a lot Decreased a little Stayed the same Increased a little Increased a lot Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 27 Figure 10 – DSL survey – “How common are the following issues, in relation to families in the catchment area of your school?” (% by response category) (n=180; not answered=1; data labels for values of 2% or less have not been included on graphic) This feedback concerning contextual issues is obviously concerning, with some issues – e.g., family mental illness, family trauma, and domestic violence – having quite high proportions of respondents reporting that they are either common or very common in their areas. Because the survey was anonymous (and could not be linked to a specific named school, for example) it was quite difficult to access official estimates for some of these issues, although these estimates themselves are often quite speculative and do not focus on very small areas.16 They were also asked to comment on whether these issues had changed at all over the last 12 months, and respondents largely indicated that a range of such issues had got worse rather than better. The open-ended comments suggest that respondents knew this because they already knew many of the families in question, and some also tried to keep in touch even during lockdown). Responses are summarised at Figure 11. 16 A useful discussion of some of the issues around estimating the prevalence of such problems can be found in Bywaters et al. (2022), where they discuss estimating levels of child abuse and neglect in circumstances of multiple disadvantage. 8% 5% 25% 28% 12% 19% 9% 44% 21% 43% 52% 37% 42% 31% 34% 38% 25% 19% 31% 31% 32% 8% 17% 4% 10% 6% 14% 5% 8% 3% 6% 10% 7% 8% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% Child abuse or neglect within the family Child/adolescent to parent violence Domestic violence Family mental illness Family substance misuse Family members suffering from symptoms of trauma Family members involved in offending Very common Common Uncommon Rare Very rare Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 28 Broadly speaking the feedback about increases are consistent with what we know from other research and from media reports, suggesting that both violence within the family and mental health issues within the family increased during the whole period of intermittent COVID restrictions. It is interesting that DSL reported that child/adolescent violence to parents also increased during the period specified (with 73% of respondents claiming that that issue got either slightly worse or a lot worse), although it seems likely that all forms of violence within the family could worsen in circumstances were familial proximity was unbroken by school attendance, and where financial stress was also added into the mix. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 29 Figure 11 –DSL survey – “For each of the following issues, do you think it has changed at all in your area in the last 12 months: ” (% by response category) (n=180; not answered=1; data labels for values of 2% or less have not been included on graphic) DSLs are in a real sense “at the coalface” in relation to some of these issues, and many expressed frustration to us, that the scope for them to take steps that will actually result in improvements for young people is quite narrow in some areas – even when the issues include serious victimisation. Examples of the following sort were provided to us by DSLs. Safeguarding Partners do share information well but responses can be poor. One of the most frustrating things is short-term interventions so you don't see impact, changes in personnel (social workers in particular). Poor and ineffective Early Help work. Recent experience with children going missing suggests that this is an area for much improved joined up work with Police and social care and improved communication. (DSL-78505169) Others commented on “repeated meetings” that seemed not to end in any effective action: Sadly, often the only outcome of initiatives is a series of meetings and their respective minutes circulated to a list of well-meaning people. The only way to 28% 36% 16% 12% 34% 26% 46% 54% 37% 52% 47% 33% 44% 17% 9% 6% 26% 38% 9% 20% 4% 7% 21% 7% 4% 24% 11% 30% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% Child abuse or neglect within the family Child/adolescent to parent violence Domestic violence Family mental illness Family substance misuse Family members suffering from symptoms of trauma Family members involved in offending Got a lot better Got slightly better Stayed the same Got slightly worse Got a lot worse Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 30 make a positive difference for these young people is to have a genuine alternative available to them. Sports clubs, social clubs, cadet/military fitness training, mental health walk-ins etc. In short, a huge amount of money, but far less that it costs to try and paper over these issues later in life. (DSL-79209434) I do not think the current cooperative system is designed to provide active help to vulnerable young people. Instead, it seems to generate repeated meetings, often with lots of people present, that repeatedly rehash the same information. There seems to then be no change to the situation. (DSL-78505405) The feedback from DSLs highlighted for us the way in which people in that role are in a unique position to watch young people’s life trajectories play out in some cases over several years. One respondent commented on the kinds of “cycles” that are witnessed, for example: I know and can personally recognise adults in local gangs as ex pupils of my school. I can name many future gang members/criminals that are currently only primary school pupils. We are not doing enough to prevent this cycle repeating. Poverty and ineffective parenting is entrenched. An 8 year old boy just follows his older role models already and repeats their behaviours. (DSL-79519886) As noted earlier DSL comments on gangs and violence in the area of their schools varied considerably, but it was clear that in some schools, those issues affected a majority of the young people attending. One DSL made reference to a “straw poll” undertaken on these issues which reinforced that perception: As a PS [primary school] we are seeing the impact on younger siblings in relation to older siblings involvement in gangs/ violence due to them being exploited by elders. At a recent assembly over 75% of Y6 and Y5 children acknowledged either knowing someone or having a relative involved in gangs.17 (DSL-78156600) Feedback suggests that some schools were not thought to be safe spaces for some young people, however, because of bullying for example (which just under 70% of DSLs felt had got either slightly worse, or a lot worse over the last 12 months). Travel to and from school also appeared to be a source of considerable anxiety for some young people. There appears to be significant under-reporting of crime in and around schools with many crimes not being reported at all (particularly crimes that 17 It is worth noting that although this figure seems high, it should be regarded as broadly indicative at best. Many of those who responded to that effect (presumably by a show of hands, which is not always a very reliable feedback method) may have known or known about even a very small number of such individuals. Also, while a “relative” is obviously defined at some level, “know someone” is not. Still, the point that the respondent was making is that in their area, gangs are very much in people’s perceptions. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 31 concern young people travelling to and from school, which were referred to in feedback both from young people and DSL respondents). In at least one borough, it was reported that changes to school catchment areas had meant significant changes to such bonded group affiliations as now two brothers or two neighbours might go to different schools miles apart. This was said by local gang workers to have changed the dynamics for some local gangs bringing in new unexpected affiliations which were less pronounced simply by shared experience at school. It was also said to have increased the risk of violence for some young people as they now have further to travel to get to school and they must travel across other estates to reach there. However, most DSLs did characterise the quality of their partnership working in a fairly positive manner, as they did in relation to questions about the role and resourcing of safeguarding activity in their own schools. Responses to some questions do illustrate a small “U-shaped” distribution however, which shows that there is variability across areas of London (as we would expect)18. (See Figure 15 in section 4.9.2 for DSL views on safeguarding partnerships more generally.) 18 There can also be variations in responses both within particular boroughs (i.e. between different schools in the same borough), and even within a single school (in cases where staff teams are divided in their views about a particular issue). PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 32 Figure 12 – DSL survey: “To what extent would you agree or disagree with the following statements:” (% by response category) (n=180; not answered=1) Feedback about schools from some of our young respondents highlights a perception that they do not get the full set of information regarding their safety from their schools – for example there are high levels of street robbery around some schools but this is seldom raised according to some pupils. Some young people said that “honest conversations were not had at school” – meaning that issues such the reality of crime, drugs and violence were ignored, downplayed, minimised regularly. Many young people felt talked down to or patronised by staff and adults, when it was the adults who were out of touch, or who “couldn’t handle the truth” about everyday violence and its impact on young people. 82% 53% 57% 3% 28% 24% 0% 3% 3% 0% 2% 1% 14% 14% 14% Safeguarding is treated seriously at my school Safeguarding is well-resourced at my school My colleagues and I have had sufficient training to allow us to address safeguarding issues Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 33 Some young people commented to us that DSLs were there for the school and not so much for them - acting for the school to help protect and safeguard school reputations – and also criticised DSLs as being too remote, detached, middle-class and not from the area they work in, but parachuted in. Thus, it was felt that some DSLs had a poor or limited grasp of the local pressures and lived experience of the pupils, and this seemed to be linked to an expressed reluctance to report or disclose. Some Older Gang Consultants working with young people felt that DSLs were reactive to what they were told by young people and were not in any way proactive. Nor did they foster a proactive safeguarding environment within the school and the community, possibly through under-resourcing or a lack of direction, but possibly that their role was designated and thus boxed-off as opposed to mainstreamed across all staff. One respondent went much further, suggesting that: [T]he safeguarding leads, the people that are in responsible positions to safeguard our young people are very, very blind, deaf and dumb to the issues. They have to be spoon-fed these issues through training, literally, that’s the only way they’re going to learn it. And that’s something that we’ve understood, that’s a really, really big issue. All of these schools in this area, and all of the places locally, all the safeguarding leads, the majority of them are white. And they don’t like to learn from ethnic minorities like myself. . . .We could see that there is a huge disparity and it’s very much perpetuated by white people given positions of power. (A-FG-PROJ3-01) The reference to “spoon-feeding” is obviously derogatory and underlies a broader perspective on disproportionality which was held by a small minority, but which was expressed to us now and again during the research. In terms of safeguarding young people more generally, several respondents (and DSLs) made reference to the need for a broader and more critical view of safeguarding itself – one which moved beyond a focus only on the immediate family for example. “Contextual safeguarding” was sometimes referred to in this context, as being a useful example of such a broader focus. That approach involves risk assessing the environment above and beyond that of the domestic environment, and in particular focusing on potential risks associated with peer groups, neighbourhoods and public space, schools, and also online space.19 19 Carlene Firmin at the University of Bedfordshire has pioneered the development of contextual safeguarding frameworks – see in particular Firmin (2020) for a detailed discussion and background. Evaluation evidence concerning contextual safeguarding approaches is still accumulating, although recent studies include Lefevre et al. (2020) which focused on a contextual safeguarding system implemented in Hackney. Although that study concluded that “It is too early for robust evidence to be generated regarding improved service user experiences or enhanced child-level outcomes”, it did highlight the extent to which the contextual safeguarding framework can facilitate the development of new processes and practices that are well placed in the longer term to generate measurable positive outcomes. It is also worth noting that that framework is much more closely aligned with public health PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 34 In our discussions with the range of practitioners and professionals that we consulted as part of the research it did seem to us that there was less sense of professional cohesion amongst DSLs as compared with some other professional groups such as social workers, or youth workers. Opportunities for the development of a stronger professional identity do not appear to have been embraced or acted on. 4.4 Feedback from the case study areas Although many of the issues that we focused on in the area-based part of the research overlap with those discussed in other sections of the report, we decided to present some of the key findings in a separate section. As we noted in section 3, the response rates for the survey components of our feedback collection on sampled estates were poor, and for us this was a significant finding in itself, which led us to probe local community representatives (and in some cases local councillors and other professionals who worked in the area) to find out why in more detail. According to some of our police contacts there were local concerns about using survey links or survey QR codes, because levels of fraud and local scams had increased markedly during lockdown, and had made residents highly suspicious, and unlikely to trust surveys that required the use of these methods for questionnaire completion. At the same time, there seems to have been an increase in the use of these consultation methods, and some areas were “flooded” with multiple consultations during the period we were attempting to generate interest in our own surveys. We were made aware of surveys implemented by local Councils during the same period, that generated zero responses. It was also pointed out to us that in some of the areas we were looking at there were significant problems with “the digital divide”, and at a time of increasing financial hardship, this problem was more apparent than usual20. While we have experienced this in previous research as well, we did not in this case also use hard copy questionnaires which could have been completed by those having no access to computers or smart phones (the implementation of a hard copy component does involve significantly more resources). (and whole systems) approaches to violence and exploitation reduction work, than are more traditional approaches to safeguarding young people within the family. 20 The term refers to the division between those who have access to computers and the internet, and those who do not. Also referred to as “digital exclusion”, research suggests that such exclusion is complicated but determined primarily by income and age. Concerning the former, ONS data suggest that only 51% of households earning between £6,000 and £10,000 had internet access in the home, as compared with 99% of households with an income of over £40,001 (https://www.ons.gov.uk/peoplepopulationandcommunity/householdcharacteristics/homeinternetands ocialmediausage/articles/exploringtheuksdigitaldivide/2019-03-04). The impact of the “divide” worsened for young people in particular during COVID restrictions, because those from households with no access to the internet (and/or to computing equipment) were disadvantaged in terms of continuing with their education (https://www.cchpr.landecon.cam.ac.uk/system/files/documents/digital_divide_research.pdf). The PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 35 But another significant reason for non-response had more to do with the way in which residents in our sample areas seem to have perceived our choice of their areas for consultation, and our reasons for consulting them about violence in particular. In all the areas we found a level of distrust of “official” surveys into local experience, which they felt would never be acted on or result in any change for them. This perception was reinforced by the fact that many residents felt that they had been “consulted to death” on some of these issues, and that they had also been given promises about possible changes to their estates or areas, or possible extra funding for local community groups and so on, which had never materialised.21 It was also felt that some local residents were tired of having their areas labelled in the media and elsewhere as “high crime areas”, and that some resented having another survey come in and reinforce that label by asking yet more questions about violence. As a local councillor in one of the areas put it: I do think residents are fed up with their estate and area being consistently classed as a place riddled with crime. Residents take pride in their community and have been working well to bring people together. (B-165) In a sense, the final group of 47 survey respondents was more like an interview sample than a survey one, although the responses themselves are informative and useful, and did add value to the final data-set. The respondents were predominantly White (38 out of 47), compared with a London average of about 60%), with 29 being female, and 18 being male. The questionnaire asks about the resident’s overall level of satisfaction with their area as a place to live, and also about whether things have got better or worse (or not changed) over the past 12 months.22 In relation to overall satisfaction, 25 out of 47 respondents claimed to be “fairly satisfied”, with 12 claiming to be either “fairly dissatisfied” or “very dissatisfied”. In terms of levels of violence and related issues in their areas, it was obviously expected that residents would perceive that these levels were high, given the way in which we sampled the areas in the first place. Local residents did perceive rates to be fairly high for the full range of issues that we asked them about, and concerning problem is particularly acute in some parts of London, and has led to the development of the Digital Inclusion Innovation Programme for example, and to a recent mapping exercise of efforts to address digital exclusion in London (https://loti.london/wp-content/uploads/2022/02/LOTI- Digital_Inclusion_Report.pdf). 21 Again, there are some interesting overlaps between our findings here and those presented by Kane et. al. (2022) in their study of neighbourhoods and violence. 22 See note 3. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 36 violence in general, 25 out of 46 respondents described levels in their own areas as being either very high or high. Many residents also felt that violence had gotten worse in the areas that they lived in, with 21 out of 47 respondents suggesting that it had either got slightly or got a lot worse. Respondents also reported similar increases across a range of specific issues that we asked about (including knife crime, verbal abuse/threatening behaviour in public spaces, on-the-street violence, people dealing drugs in public, robbery/theft from the person, and similar issues), but there were high levels of “don’t knows”. More generally in terms of whether their area had got better or worse as a place to live over the past 12 months23, residents also reported a worsening of conditions. Twenty-three out of 47 respondents felt that their area had gotten either slightly or a lot worse, and 16 felt that it had stayed the same. Five out of 47 respondents felt that their area had gotten “slightly” better. It was also of particular interest to us to learn whether respondents were themselves concerned about levels of serious youth violence in the areas where they lived, so we asked specifically about this after reminding respondents how we were defining “serious youth violence”. Respondents were asked “How concerned are you personally about levels of serious youth violence” in their area, and 36 out of 47 respondents claimed to be either “moderately concerned” or “extremely concerned”. Only 6 out of 47 respondents claimed to be “not at all concerned”. Those who suggested that they did have some level of concern about serious youth violence were routed to a couple of questions about whether that concern had any impact on the way that they lived their lives, and what sort of impact that this was. In responses to the former question, 25 answered “yes”, and 21 answered “no” (one respondent did not know). Nineteen residents commented specifically on the impact that concern about violence had on the way that they lived their day to day lives, with most of these comments being about how the resident is afraid to move about the area during certain times of the day or night, or how they avoid particular areas altogether. Others made it clear that they would prefer to live somewhere else, and some residents referred to multiple impacts in the same response: I feel I can't take my children out around certain areas. I might not necessarily be a target but I'm worried about getting caught in the crossfire. I also am concerned my children will grow up thinking this is normal and soliciting drugs is acceptable. 23 See note 3, above. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 37 (RES-170863110) Having been the victim of street robbery I'm always scared coming home after dark but also now worry even in busy public spaces. I've seen guns being shown off to girls by young men, walked passed blooded pavements, watched untold acts of threatening behaviour in public spaces. All these things make life often intolerable and my adult children keep advising I should move away for a safer life. (RES-168688870) The impact is immense, having a disability means that one is limited to the immediate area. Following a recent assault in my local park means I now also feel unable to access local green spaces, as they are unsafe, due to drug paraphernalia and all the other crimes that are associated with it. I have to constantly alter my schedule in and around those who are offending and being offensive, if I want to be outside at all. (RES-169948041) I do not feel safe walking alone and feel the need to use the car even for short journeys. The lack of CCTV cameras and police presence makes me feel even more unsafe as the public do not step in to assist if there is something 'brewing'. The amount of drugs being dealt has dramatically increased with people being moved on, only to be replaced by others within the same gang. I wanted an escooter for fun but I am afraid even to ride it in the park in fear of being robbed. I am afraid of wearing gold jewellery or using my mobile phone in public in fear of being robbed as crime is now taking place openly during broad daylight, criminals no longer wait for sun down It’s much easier to commit 'quick easy targeted crimes' during the day. (RES-169721627) We are due to have a child soon and have no desire to raise them in this area. (RES-169124463) The feedback also suggests that alongside these kinds of perceptions about the severity and trends of violence in their areas, there is a sense of community and “locatedness” for many residents. The survey questionnaires did ask about overall levels of satisfaction with the area, and also about levels of trust in neighbours, whether people felt they could socialise with their neighbours, whether people from different backgrounds “get on well in this area”, and so on. For most claims of that sort, at least half of respondents answered that they either agreed or strongly agreed, and in the open-ended comments (as well as direct interviews), residents in some of the areas expressed frustration that their areas had been stigmatised in the media for being dangerous and out of control. The feedback highlights an interesting mix of concern about personal safety on the one hand, and on the other, a sense that high violence areas can also be places where there is a sense of community and where residents can find support from one another. More generally, the feedback highlights how difficult to interpret “feelings of safety” can be, and how those perceptions might interact with varieties (and the context of) social connectedness. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 38 Previous research has shown how areas of very high crime can also have high levels of resident connectedness (see Bottoms and Wilson, 2007, and Bottoms 2009), and also how visible local efforts to address difficulties (e.g., through local projects and community activism) can act as “control signs” which can in turn affect feelings of safety (see Innes, 2014 for a discussion). We also know from other research that different groups can have widely varying thresholds for “feeling safe”, with some traumatised or vulnerable groups – including gang-affiliated young people for example – struggling to feel safe anywhere.24 It is also worth noting that in some of our case study areas, we found examples where residents and community workers were involved in efforts to address local issues (such as projects to support children and young people for example) and where those efforts appeared both to facilitate oasis-like pockets of local safety, and to provide visible signs to other residents that positive steps were being taken to improve conditions. Some local projects in these areas also appeared to be effective in terms of generating positive changes for those who they worked with (although we were obviously not in a position to evaluate them), and some were also inspirational for our team members who spoke with residents and workers in those areas – not least because such projects sometimes functioned with very little outside support or funding, and they were staffed by committed and determined individuals. Views among residents about policing were mixed (see also section 4.9.3), but some residents did describe events involving the police that were particularly negative. One resident interviewee described a stop and search incident which she witnessed, for example, involving a local boy from the estate, where half a dozen undercover police officers searched the boy on the street on his way to school. Apparently they found nothing, but then left him crying and picking up his own school books. The resident was somewhat politicised by the incident, and subsequently discovered that many of the estate’s young people were also stopped and searched on their way to school. Members of the research team were not able to speak with officers involved in those incidents, but clearly, if even a few such incidents generate these kinds of perceptions among local residents (and among young people), that will obviously have implications for police-community relations and for the effectiveness of policing in the area more generally. We did hear quite a few descriptions of this kind, although again, the specific details could not usually be corroborated. In some of the areas that we focused on, there had also been a clear impact on communities when young people in particular were lost to local violence - this is often not recognised and would benefit from further research, but it was clearly visible to the research team on certain estates and in ongoing conversations with young people. There is often a sense of idolising anyone who has “fallen” or “passed” due to street violence – as realised and visualised by the erection of public murals and public grieving events – although those practices are also sometimes linked to reminders of the need to take revenge and engage in violence. 24 See Van der Kolk (2014, Chapter 5) for a useful discussion of the issue. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 39 Further study of what we have called “community trauma” could be quite beneficial for the violence reduction field. The notion is relevant for several reasons beyond just looking at the impacts of local violent deaths of young people – it is relevant also to issues around some of the other correlates of high violence areas, and since many of these areas also have higher levels of residents from BAME groups (and are more likely to be deprived as measured on the usual scales), the issue is also relevant to efforts to understand links between trauma and disproportionality – links which have motivated the authors of some recent reports on this to have separate sections that discuss community trauma (see for example, IOPC, 2022). Finally, both interview and survey respondents in some local areas made reference to connections between “HMOs” (houses in multiple occupation) and what we would call the vulnerability landscape in those areas. What has happened in some areas of London – Redbridge and Croydon in particular – is that because of the availability of relatively cheap private rental properties, practices have developed that involve the placement of vulnerable groups (e.g., substance misusers, individuals who would otherwise be homeless, young people in care, and individuals with mental health difficulties) in those high density accommodations. That co-location of vulnerable groups in what are sometimes very geographically concentrated areas can create a dynamic critical mass of both drug dealers and users for example, which crystallises into a self-sustaining and resilient local drugs economy, but which also fuels a variety of forms of exploitation, and the violence which underpins those. As a police respondent in one borough described it: A lot of our problem in [this borough], the drugs problem is down to the housing of the users and the dealers. Right in the middle of the town centre, we’ve got a few roads, a block of roads that are all HMOs and bail hostels and cheap hotels and they are full of drug dealers and drug users. It’s the bail hostels, the HMOs, the cheap property and the hotels, that’s where the market is coming from….The lion’s share is coming from the cheap property. (B-FG, INT6) A senior manager of a voluntary sector organisation described connections between placement in HMOs, vulnerability and exploitation: PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 40 I just mean that we need to look at all the systems and not think that if we put a little youth violence project over here we’ll solve the problem when we’re not solving the systematic issues. And whether that is funding or addressing it differently or whatever we’ve got to do, we will keep feeding this if we don’t take care of some of those most vulnerable people in our society that I suppose as a local authority, and I’m going to say ‘we’ because I don’t see that we’re doing different parts of this role; we are not doing our jobs, corporate parents, the way that we treat kids in care. . . I mean, it boggles my mind how we can allow young people in care to be pushed out to a bedsit in an HMO in some crappy part of town, with no interaction, no guidance, no authority figures. So, obviously when somebody comes along and says, would you like to jump on my county line, but you’ll have to go and live in Lincolnshire, well, in a way, they’ve kind of already done that, haven’t they. So, you know, it’s they’re institutionalised into thinking, OK, well that’s doable, but this time I’m going to make some money, so, you know, why not? (A122) A local authority drugs worker made more specific reference to the vulnerability of rehabilitating drug users being placed in HMOs: And we constantly have a problem with people that we treat, that become abstinent or come out of rehab and then they end up getting housed in an HMO round the station. And, from my perspective, realistic how long do we help somebody early in recovery from drug addiction. And of course the dealers will of course target them, absolutely. (B-FG, INT2) Many HMOs are required to be licensed, but depending on the borough, this sometimes applies only to larger HMOs (where 5 or more people from 2 or more unrelated households live and share facilities) – these are subject to Mandatory Licensing, but that is not always enforced effectively. Respondents to the residents’ survey also mentioned HMOs as being responsible for increases in a range of crime and related problems in their own areas: [The estate] has become a catchment area for [the Council] to dump vulnerable people without supporting them sufficiently. The council also does not enforce HMO licensing. This has created a toxic environment filled with people who are unable to cope with the stresses of the pandemic. . . Encourage [local projects] to support their vulnerable tenants and to police against antisocial behaviour. Police against idling cars to deter the drug dealers. Enforce HMO rules to deter slumlords from renting a single house to 25 single men. (RES-168698054) PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 41 One of the many local data-sets that members of the research team received during the study, was a small database with detailed material concerning vulnerable individuals placed in a sample of HMOs in a small area in one borough. Our analysis of that material (and of responses to questions that we asked local stakeholders about it) highlighted how complex the range of individual needs can be, and how difficult the management issues are that this kind of co-location can introduce. Our investigation also showed how some of these difficulties were not anticipated by local professionals in advance, and how difficult to address they can be once they become rooted in an area. One implication of this for local violence and exploitation reduction work seems to be that key stakeholders who focus on HMOs and related developments could profitably be brought onto multi-agency planning or strategic groups if they are not already there, and that a shared understanding of how developments in housing can be linked to the local “vulnerability landscape” can also help in the adoption of contextual safeguarding approaches (and the management of vulnerability by key public agencies). 4.5 Violence, vulnerability and local context – a vignette In order to illustrate some of the issues referred to above, we focus in this section on an example which highlights key issues concerning vulnerability, local context, and violence and its impacts, as well as some of the interconnections between them. We had other similar episodes while we were gathering feedback during the research, but this particular session was fairly lengthy, and allowed us both to gather useful feedback, and also to see some interactions which don’t normally occur in the same fieldwork exercise. We held a group discussion on the street with young people attached to a local youth club, in an area that was characterised by high levels of violence and gang activity. We had arranged to visit the project and hear the views of project staff and young people, but on arrival it became clear that an incident was developing in the immediate area of the club, concerning the activities of a rival group that was apparently planning to target some of the club’s attendees. The young people seemed very concerned for their own safety, and they were discussing how to respond to the perceived threat, and which routes they might take when they left the area and so on, which could help them to avoid being targeted. The club itself is regarded by many local young people as being a kind of sanctuary, a safe place which is also overseen by project workers who have lived experience of violence and gang activities. When asked about the events that seemed to be playing out in the area, the project manager explained: It’s just a little issue, like some of the people that have got fights with another school, a notorious school, they’re going to come down and try to attack some of these kids . . . .So the students from there are going to come down here with their knives to come and attack some of the kids from here. We’re just mitigating whether they’re going to come or not or what their plan is, or where PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 42 they’re going to go and what area they will be hiding in order to catch some of the young people here. Just a bit of like, looking at the logistics, their logistics, and understanding how they would attack. Like I know these streets like the back of my hand, so if I was to attack any person, how would I do it? And then I’ll mitigate it from that to say ‘that’s how they would do it if they wanted to get away with it’. If they didn’t want to get away with it, then it’ll just be a bare-faced attack. (A-FG-PROJ3-01) While waiting to see if members of the rival group would make an appearance, a discussion ensued between the researcher and the group of young people (all males, aged around 14-16) about the area, the local school, and about violence and its impacts. The young people made it clear that concern about violence was something of a constant for them, and that they tended to group together for their own protection. Male 5: Everyone’s frightened. Deep inside, people don’t be reassured, maybe don’t show it, but everyone is frightened about everything. Male 2: You’re not really safe anyway. Male 5: Exactly. No matter who it is, you could be the most respected person in the whole school and will have the most security ever around you, you could be one of the rich earnings people with the most amount of security, but at the same time, the one time where you’re no longer safe, alone, someone is going to take you. Someone is going to rob you. Male 5: All we’re trying to do, just come after school, like either just straight go home or come [to the project]. We feel safe here, we know what we do or just go home with friends just in case anything happens. Male 2: Like, in a big group of friends, so nothing will – Male 5: – Exactly, yeah, just in case anything happens then we can defend ourselves. We asked them about their local school and about the level of support that they had from teachers about issues concerning their own safety, and the young people commented at some length on those issues, noting that the teachers were not from the area, and did not really know what was going on. Some felt that the school was not aware of the issues locally, and that were more likely to punish than to offer support on these issues: Male 3: But the school pretty much doesn’t believe in it. The school talks about it only if so many people in school talk about it also. But if no one talks about, they’re not talking. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 43 Male 4: Only talk about it, if we talk about it. They don’t care. Male 2: The school won’t know about it because like most of the teachers, they live like out – they don’t live – they don’t live in [the borough]. Interviewer: They’re not from [the borough]? Male 2: Exactly. They pretty much don’t know really about what’s going on. But he said that they don’t care. Male 3: You say the teachers really care about us, but even if, they’re like, “Oh, just go home, just go to …” These, man, rather than like helping us, helping people with problems and helping them with their issues, what they do, they just reprimand them. Male 4: Exactly, basically, they just say stuff and not even act. They’re the ones meant to be helping us but they don’t do it, and even if they do, they’ll just give us a punishment. Interestingly, during the discussion two teachers from the nearby school came walking through the area, in an effort to break the group up and encourage them to go home (or to “shoo them on”, as the project manager put it later). The teachers did not ask any questions about what the group was doing, or why they were there. Issues concerning ethnicity and class also came up during the discussion, with the project leader suggesting that those at the school who are responsible for safeguarding students are “White middle class” and not at all aware of the circumstances in which many of the students live. Similar comments were made by the young people, and taken together, the feedback highlights some of the issues raised in current discussions about the importance of “cultural competence”. As it turned out, the rival group did not make an appearance - and we did not have enough information to piece together a background history to this event (e.g., whether it was about retaliation for previous events or not) - but in response to our observation that the young people near the club seemed quite fearful about the perceived threat, the project manager noted: Yeah, absolutely they did, because they know that there could be a potential chance that they will be stabbed and they could lose their lives today. So, you know, with that in mind, you can see a lot of the kids find their sense of security just by being in our presence, just by being, you know, next to [the project]. This example also illustrates a key point that we return to in section 4.8, which is that some of the key drivers of violence are also interactive – that is, they can reinforce one another in a cyclical way over time. In an area where groups of young people engage in attacks on other groups, levels of concern about individual safety increase PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 44 or are reinforced, and can ripple across whole communities. This can lead to increased weaponisation – where young people in particular feel not only motivated to group together for self-defence, but to arm themselves for their own protection. Some of our other feedback also suggests that where actual events of public violence regularly do take place, they are a powerful reinforcing signal to all groups that violence is an ever-present threat, and violence in this way becomes normalised in those areas. A normalisation of violence in turn means that thresholds of tolerance are raised, which can mean that incidents are less likely to be reported if they do not reach that threshold of perceived seriousness (which is partly why none of the robberies referred to by the young people above were reported). Serious woundings and murders in such areas can become community events which are traumatising at both individual and community levels, and further impacts can include changed behaviour in public spaces, as residents and users avoid certain areas. Those changes can in turn have economic impacts, if they include a withdrawal from the local night time economy for example, or an abandonment of high streets. Those latter changes can then increase the scope for drug markets to flourish, as an abandonment of public space in some areas allows it to be taken over by others, and can also increase antisocial behaviour and levels of violence in those areas. 4.6 Impact of COVID There was clearly one overriding impact of COVID that was identified across all respondent groups, which was adverse effects on mental health – of young people in particular, and especially those living in deprived areas where they also experienced abuse or neglect (or domestic violence and parental substance misuse) at home. DSLs in particular identified very high levels of trauma and adverse experience in the home lives of children and young people attending their schools – 52% felt that child abuse and neglect were either common or very common among families in their catchment area, and the figure was 80% for family mental illness, 48% for family substance misuse, 67% for domestic violence, and 60% for family members suffering from symptoms of trauma. All of these things were said to have either got a lot worse, or slightly worse, over the last 12 months, with levels worsening in particular for child abuse or neglect within the family (63%), and domestic violence (77%). DSLs pointed to severe mental health impacts from these trends, on young people’s mental health, which in turn they related to future behavioural problems that are corrosive of effective learning. During COVID restrictions some issues died off initially before then quickly altering in their presentation as drug dealing crews and street gangs adjusted their operations and movements. It was necessary that drug dealing activity alter quickly given that young drug dealers were more easily spotted and arrested by the police in circumstances where there were few people on the street. Numerous strategies appear to have been adopted in response to those circumstances - it was suggested to us that some drug dealers even signed up to become taxi drivers (or simply masqueraded as taxi drivers), for example, so that they could move around and deal directly from cars without attracting suspicion or attention. Cases of this kind were described by a youth worker in a voluntary sector project, and also by a local PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 45 authority serious violence worker in another borough (A-164, A-157). Other lockdown tactics emerged to ensure drug supply was maintained including more localised dealing, less travel on public transport, creation of fake IDs to provide reasons for movement outside, and a greater switch to ordering and dealing drugs online.25 Whilst online ordering and dealing of drugs has hitherto been the premise of recreational drugs, steroids and so forth, lockdown prompted some drug dealing crews and street gangs to move more into dealing online. Some respondents noted that during lockdown, the rules for young people were less clearly defined, articulated or stipulated, partly because most young people would have otherwise been in school during periods when COVID required closure, but also because there had not always been an effective adult presence in their home lives during this period (or only a sporadic presence). As a result, it was claimed that young people were more out of control; less likely to accept instruction/direction or discipline. One voluntary sector worker described to us how groups of 12-14 year old boys were “roaming the streets” during some of these periods in one borough, and that they were “absolutely terrifying” because they seemed “out of control”. We were not able to identify a proxy measure of some kind in the existing data that would have allowed us to test such statements, although DSLs did report to us in their survey responses, that child/adolescent to parent violence had probably got worse in the 12 months prior to them completing the questionnaire – 6% of DSLs claimed that this problem got “a lot worse”, and 37% that it got slightly worse during that period; 36% felt that the problem had stayed the same (with less than 1% claiming that it got slightly better, and 21% claiming not to know). In their open-ended comments some DSLs also described how some children from their schools have parents who are simply absent because they work long hours; some of these families even have relatively comfortable levels of disposable income, but the young people use the time to gather in parks and take drugs. In some cases there are parents who themselves have mental health difficulties and are more easily pressured or controlled by their young children. One DSL described to us cases where families with disabled children did not get the support that they needed to work with their children, and where autistic young people for example, may hurt or control family members as a result of their autism and may need a different type of intervention (DSL78538881). Difficulties of this kind, some DSLs claimed, became worse during COVID restrictions. Another DSL noted that some families in the area of their school struggle to meet the needs of their children at the best of times, and that during periods of school closure this situation became even worse: 25 Cases where drug dealers masqueraded as joggers or as NHS workers were highly publicised during 2020 (see for example Drug dealers posing as joggers and NHS staff in Covid-19 lockdown); other examples of changes in methods were given to the research team – one, for example, where a dealer went out in public dressed in a thobe and kufi, and carrying an armful of Islamic texts (along with a backpack full of drugs) to reduce the odds of being searched. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 46 Parents who struggle to meet the needs of their children struggled more with children at home full time and children missed out on daily support in school (beyond just education, e.g., emotionally responsive adults) (DSL-75434130) Other respondents also noted that there was active recruitment of young people into street gangs during lockdown, and that opportunities were taken to groom, entice and recruit new players into “the game”. One police respondent linked the apparent increase of recruitment to financial factors (see also section 4.7.4): Anecdotally we have heard from our partners in youth work that gang recruitment has got a lot worse in recent months, including young people carrying drugs. It has become increasingly easy to recruit young people into gangs because of families’ financial struggles which gangs are exploiting. Our partners are saying that structural issues that cause violence are the same, and punishment is not working as a deterrent. This is because the present needs of young people are immediate & paramount (e.g., money for their families or status) and are more important than the threat of a potential prison sentence. Nationally police forces are reporting that there is more recruitment into County Lines, which they say is a child protection issue. (PRAC-79797999) This recruitment during lockdown altered the affiliation of some young people with street gangs. This will eventually lead to increased competition and rivalries within street gangs which may take months to play out but which will impact upon behaviours and activities as new skills and thresholds are introduced. 4.7 Drivers Respondents’ comments about the key drivers of violence ranged widely by level of focus, covering factors at individual, family, peer, school, community and societal levels, but also covering drivers that span these levels.26 Some respondents limited their comments to factors that operate at an individual level, but a majority of respondents also made reference to wider factors such as deprivation, lack of opportunity, poor quality services or lack of access to services, family issues, social media, or the impacts of both individual and community trauma. The feedback illustrated how further research to “map” key factors in the context of individuals’ lives, and in the context of an individual’s pathway through and across those different levels over time, could be very useful for the field. Some highlights from the feedback concerning factors of this kind are provided separately below. 26 A very useful framework for mapping drivers of violence can be found in Cordis Bright (2015). A recent MOPAC (2022) publication also contains some very sophisticated and elegant analyses of links across factors of this kind, produced using sets of official violence data. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 47 4.7.1 Drugs Across all respondent groups, drugs were most commonly mentioned as being a key driver of violence in London, although the precise way in which drugs might be linked to violence was not always articulated. Comments about drugs made it clear that this factor spans the levels referred to above. At an individual level for example, some referred to the way in which users become involved in illegal activities and violence in order to secure the resources to gain access to drugs, and others focusing on the way in which drug use itself can increase interpersonal violence because of its direct effects on individuals and relationships. Some just mentioned drugs as part of a long list of factors associated with violence, without talking about what the precise causal link might be. One police respondent described some of the links between drugs and violence in this way: I wouldn’t say they’re necessarily all involved in violence. I mean, the job that I’m dealing with at the moment is a firearms job, so that’s obviously linked to violence. And a lot of the jobs that we do around drug supply, a lot of the individuals have history of violence, violence, history of criminality. A lot of them carry weapons and a lot of the young ones carry weapons for their own protection, from other people, so they don’t steal their commodities. So, yeah, I would say a lot of it is revolving around violence. [Interviewer: And how much of it revolves around drugs and drug dealing?] I’d say a lot of it is, it sort of comes hand in hand. A lot of them deal drugs or are involved in the supply of drugs and that’s how they, they sort of fund their lifestyle, their criminality, it kind of starts with drugs. And then that kind of links in with the violence and the gangs. Obviously there are feuds and, again, it goes up and down, depending on what’s going on. Usually when there’s a major incident, like a murder or an attempted murder kind of aspect, then there’s usually quite a fall out in the subsequent weeks and months after that. (A-066) Another interviewee from a local authority exploitation team also placed drug markets at the heart of the issues that they were dealing with: Our issue is the drug market and it’s about exploitation and making money out of drugs and poverty and all the other things that are linked to that. (A-075) PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 48 So long as there is a market for drugs, there will also be competition to supply that market and benefit from the profits that can be made from doing so, and as many respondents suggested, violence is a key component in that competition. Still others note that drugs are linked to the negative impacts of trauma, and that they are used for numbing and escape purposes – but this also increases impulsiveness and risk-taking behaviour, and erodes problem-solving capacity. One of our young interviewees described how drug selling on her estate was occasionally linked to violence (and her comments also highlight how serious some of the violence is that young people sometimes witness in their own areas): I think a lot of violence comes from drugs. I think that’s the root of it, the gang war. I saw a boy, a man, a young man, one time, he was shotting27 on the estate and he had just started shotting actually, and then he and another boy, man that was shotting on there before, that had left, had run away, God knows where, and come back, had been like, “Oh so now you’re shotting on the estate, yeah”, and beat him literally to a pulp, beat him, ‘til the guy was [fanned], like collapsed outside. The guy managed to run to the Police Station and collapsed outside the Police Station, head injury. But a lot of the violence does come from selling the drugs because you’ll have the boss, then you’ll have the runners, who are the ones that go out and actually sell the drugs to the people. And then the boss will basically supply the drugs, come onto the estate to supply the drugs and then leave. And then you will have other people also that want to sell drugs and if someone else starts selling drugs on someone else’s turf, that can create a big, big war. (A-YP-07) Links between drug use and family issues were also commented on by respondents, with some noting how substance misuse within families also led to some neglect or abuse of children and young people. Just under half of DSL survey respondents reported that substance misuse was either common or very common within families of children at their schools, and 42% felt that this issue had go with worse or a lot worse in the last 12 months. 4.7.2 Social media The influence of social media/ networks continues to be central, and is woven into all aspects of life for most young people. This has led to new forms of vulnerability as well of course, and both practitioners and young people described many of these to us in their feedback. A DSL noted, for example, that there is real concern at the DSL’s school about game apps such as Discord, which are being used for grooming purposes: We are extremely worried about grooming via gaming - discord servers and groups set up where young students anticipate they are corresponding with 27 The term usually refers to low level drug dealing. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 49 people their own age... however on a number of occasions we have had them groomed into unacceptable behaviours. (DSL-78533738) The role, importance and significance of social media continues amongst all young people, including those who are gang-affiliated. This is used for self-promotion, gang branding, buying and dealing drugs, gang recruitment, surveillance of others; ascertaining and determining loyalty; blackmailing others, fraud, etc. Music such as drill remains a key expressive output for some young people across London, especially for working-class boys in some social housing estates. For many this seems to operate mainly as an artistic outlet of self-expression of their lived experience. However, a small percentage of drill content is highly targeted towards certain individuals or crews and articulates highly specific violent incidents which have actually occurred. Such targeted humiliations are viewed as provocative, and are thus often generators of violence. Some of those connections were highlighted by one of our young interviewees: I think that most of these gangs are influenced by what they see on social media, having to keep up with a certain lifestyle and through music, like drill music and stuff, they have beefs and it all gets shared on social media and I feel like they’re constantly having to keep up a persona that they show on social media amongst each other. (A-YP08) Respondents noted that some young people can be successful in relevant forms of music production and generate significant amounts of cash income from music publishing and production companies who will provide a sizeable cash advance to those deemed worthy of further investment. We were informed that some young men can find themselves in receipt of advances ranging from £10 – 50K for music recording, although we were not able to confirm some of the details provided to us. One of our practitioner interviewees described some of these developments, noting that the scope for both involvement in violence and money-making has been broadened by social media: . . . with the advent of like YouTube and drill music, which literally translates into stab music. That is the best way, it’s a way for the young person to get notoriety within their community, within their area, and they see it as, you know, a legit way to go. Especially when you’re seeing, like, local drill artists. We just had a meeting today, one’s been signed, another one’s been signed to [a prominent music company] with money in hand. (A-091) PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 50 Feedback from DSLs suggests that there have been few issues concerning drill music at schools across London however - with 74% of DSLs suggesting that drill music had not raised any concerns about young people at their school, and 14% suggesting that drill had raised such concerns (N=165). The open-ended feedback about the issue was also reflective of wider debates about drill music (and its causes and impacts), with some DSLs expressing their own concern about apparent glorification of violence, for example, and the creation of inappropriate role models for children and young people. Other DSLs suggested that drill music was a reflection of the reality of many young people’s lives in circumstances of severe deprivation and high levels of daily violence.28 More generally in terms of links between social media and violence, feedback from many respondents suggest that what is key is the speed at which social media operates, and in certain circumstances this makes it act like an accelerant of violence, where complex perceptions about status and respect, loyalty, and of course business, interact. 4.7.3 Trauma Links between violence and trauma were much referred to by interview respondents, and also by respondents to the practitioner and the DSL surveys in particular. The feedback that we gathered made it clear that the way in which practitioners think about trauma and its possible links with violence and exploitation is not always consistent with what we know about these impacts from other research. Much of the feedback was quite ACE-focused rather than what we would call “trauma-responsive”, for example, and it was sometimes suggested that dealing with the impacts of trauma would require the inclusion of psychologists or clinicians on violence reduction teams – as if only they have the skills to engage with negative impacts of trauma, and other workers can at best be knowledgeable about such impacts in their dealings with vulnerable people (e.g., to avoid re-traumatising them). When asked about what being “trauma-informed” meant for interactions between project staff and vulnerable young people at a project in one borough for example, the project manager suggested that work focusing on the impacts of trauma was for others – we are not psychologists (C-027). 28 Some of these issues have been developed at greater length in some of the wider research on this topic. See for example Faulkner’s recent Knife Crime in the Capital (2021), which provides an evidence-based account of links between the content of some drill music and actual killings or attacks over time in specific areas of London. See also a recent publication by Hall et al. (2022), which focuses on links between grime and drill music lyrics, and serious violence and gang activities, and concludes that violent lyrics are a “reflection of lived experience” and the impact of adversities on Black boys and men. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 51 Strictly speaking ACEs are static factors in the backgrounds of many people – and especially those involved in violence, as we know from decades of previous research – but the impacts of trauma are dynamic, and can be affected profoundly even by non-clinical staff. Having an impact on what young people in particular are dealing with in the present should arguably form a more conscious part of violence reduction work (and be much more clearly articulated within theories of change that are used by violence reduction partnerships), since the scope for having positive individual impacts can be sharply constrained by the impacts of trauma. Purely ACE-focused approaches were also described by some respondents as being too “backward-looking”, and those who worked with gang-affiliated young people in particular often said to us that they prefer “forward-looking” approaches that explore options and possible futures with those they work with. Comments of this kind again suggest a kind of “either/or” understanding of trauma and its relevance to practice, and that kind of understanding also tends to close off avenues for understanding protective factors in individual timelines which allow for the development of resilience for example. Some of our interview and survey respondents did feel that attention to the presence of trauma symptoms in work to address violence and exploitation was not adequate, however. A DSL commented on the issue in this way: The rhetoric is “trauma informed” but this does not play out in decision making or strategic responses. The reality of the everyday lives for some YP does not correspond with the threshold document in terms of a response. The school are seeing more traumatised children in crisis whose needs are not being met by partners including SEN, CSC and CAMHS. These children are further punished by mainstream schools who are not equipped to meet their needs and don't get sufficient support, resource or understanding from some partners to support them holistically in school. The reality for some of these children is that school end up holding them or managing behaviours to reduce the high risk of exclusion and are not able to educate them and due to the risk factors identified at Y2 / Y3 the trajectory for these children is poor and increases the risk of them being criminally and sexually exploited as they get older. (DSL-78156600) Others suggested that dealing with the impacts of trauma is not well facilitated by a system which sets high thresholds for intervention, especially in times when budgets are stretched. As one interview respondent expressed it: But I just think if you don’t kind of meet certain thresholds sometimes, then – and if in general for mental health and stuff like that, trauma wouldn’t necessarily hit the marker for threshold. So, it can often get lost, but I think if you have any sort of youth worker, youth services worker, children’s services worker … So, anybody that really works with young people and children should have some form of training around trauma. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 52 Because I think if you can identify trauma early, then you won’t necessarily have half of the kind of issues that you’ve got with exclusions and stuff. Because there’s a lot of young people, especially in our borough, that have witnessed, seen or been involved in stuff that would be impacting on their mental health. A bit like soldiers. (A-095) On the other, some respondents argued that the impulsive or violent behaviour of some young people is often anchored in key symptoms of trauma, which are in turn related to adverse childhood experience. As one DSL expressed it in relation to behaviour at school: Secondary schools are often focused upon results and therefore are not addressing the high level of students who have suffered significant trauma that will impact negatively on behaviour and their ability to learn. (DSL-78677973) A focus on trauma was in some cases anchored in analysis of information that professionals had gathered in relation to their own clients, and where that information was interpreted as being relevant to subsequent pathways into serious violence. As one local authority gangs worker described their own analysis of background information on cohorts involved in serious violence in their own borough: I think a lot of the factors, especially amongst young people, I did life course – I did a piece of life course analysis of our top ten serious youth violent offenders, which normally means they're serious youth violent victims as well. And I think on a longer term we identified that there was strong evidence of domestic abuse and child neglect, which is a form of domestic abuse, in the home at a young age that they experienced. There was also a high level of cannabis use in their young teenage years. There was – they were involved in minor antisocial behaviour from a young age so it's those things but really what my long-term plan is for this borough, the only way to tackle violence in the long term is domestic abuse. I think that is where we really have to focus almost all of our resources on a strategic level to really, if we really want to drive down violence in this borough. (A-129) Although the perceived links between trauma and violence (and/or exploitation) were not always clearly articulated by respondents, some did offer comments in that direction (as did the latter respondent, who attributed some causal power to experiences of domestic violence in early life – we comment further on this in section 4.7.6). A voluntary sector interviewee focused on a trauma-linked interpersonal aspect to the “recruitment” of young people into gang or offending activity, for example: PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 53 It is over-simplistic to see this phenomenon as “recruitment”. It is traumatised young people finding each other and acting out much of their interpersonal trauma patterns - seeking out approval and attention, seeking out status and power, and acting out warped notions of care-receiving and care-giving. There is a transactional interaction in large part fuelled by a lack of self-esteem and self-confidence that needs to be defended against. (A-075) 4.7.4 Socio-economic factors, debt, exclusion and lack of opportunity A majority of respondents who commented on key drivers of violence made at least some reference to poverty or deprivation, and and/or to related issues such as lack of opportunity, debt, and access to education. Some respondents also alluded to the key role played by a lack of opportunities in the lives of young people – given a lack of alternatives, as some young people see it, their option to become involved in offending activities is for them the only realistic path that is open to them: A lot of the young people who work with are involved in crime, but they don't want to be. They just feel they have no other options. (PRAC-78783548) An environment of deprivation and perceived lack of opportunities can also make it easier for young people to be lured into participating in offending or gang-related activities, as many of our respondents commented. As one police respondent put it: I feel that this is a losing battle as often the dealers or gangs will provide the young persons with phones, game consoles, Gucci gear, drugs and so on. This is by far much more than what we can offer them. Until we can at least take this power away from those recruiting we will not beat this problem. I have stopped the same youths, black, white, Middle Eastern or Eastern European and often those individuals will wear the same two items of clothing which will be Hugo Boss, Gucci, Prada or other high value brands. This is provided by the recruiters and these kids will not want to take them off as is shows their "status". Even when they have little to nothing else to wear. They will wash these clothes and put them on still wet as it is the only status clothing they have. (PRAC-79210135) Comments of that sort suggest not simply that young people in some areas might have a desire to own specific “high value” brands (although they might), but that such items provide a real and visible contrast to what those young people might be able to acquire for themselves otherwise. A local authority gangs worker also commented on the importance of status or “aspiration”, but also suggested that even though drugs are commonly viewed as being a key driver, poverty is a more fundamental (“core”) factor underpinning violence: PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 54 So I think those are the issues we need to start tackling. That actually it’s quite an aspirational thing now for some young people and obviously they’re being drawn into it through poverty and what not but, yeah, for me, I just think we need to get away from this conversation about drugs fuelling the violence and we need to start talking more about the core factors, which are poverty. And then, from the poverty, there’s this aspirational thing where you can make money really quickly, through the music, through drug dealing. (A-091) Some respondents who highlighted the importance of poverty also linked it with both lack of opportunity and lack of education, as in these comments from a local authority community safety officer, and from a police officer: And I would like to see much more, we know what the drivers are for people going into violence, lack of education, poverty, so why aren’t we doing that instead, why are we spending hundreds of thousands of pounds on football clubs and boxing and all that nonsense. (A-078) If you look at where crime and where violence and where gangs are embedded, they tend to be in areas of very high social deprivation. They tend to be in areas with significant poverty and other issues. Low school attainment. All of those types of things. (A-126) Many respondents also suggested that in a context of poverty and deprivation, the development of “debt bondage” at an individual level was much more likely to occur, and a majority of those who work with gang-affiliated young people commented on the importance of debt bondage to individual pathways into gang activities and violence. In relation to debt bondage, violence can also become an ever-present threat both for gang-affiliated young people and their families. This local authority community safety respondent highlighted the importance of debt bondage both to individuals and families already under extreme financial stress: PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 55 Debt bondage plays a keys aspect in child criminal exploitation issues in [this borough]. We have high numbers of young people reportedly involved in drug dealing / county lines; with many of these young people making disclosures of professionals that they are working to pay off debts to criminal gangs. We have received several reports of families paying drugs debts to gangs, putting immense financial pressure on families who, in some cases have had to borrow the money themselves. Young people in [this borough] are trafficked across various parts of the UK, but also within London. Drug dealing is often the catalyst for issues around serious youth violence with young people in conflict with rivals over territory and notoriety. The pressure debt bondage plays on the individual and the family both mentally and physically is extremely worrying. (PRAC-78743797) About half of the practitioner survey respondents who offered open-ended comments on issues concerning debt bondage (i.e. 24 out of 48 who offered these comments) felt that this was of key importance, particularly to involvement in drugs markets and/or country lines. Lack of opportunity is also strongly linked both to exclusion from school, and to issues around disproportionality, and many respondents commented on the importance of each factor to involvement in both violence and offending more generally. In relation to exclusion and ethnicity, some respondents pointed out how rates for Black boys in particular (for both temporary and permanent exclusions) are far higher than for any other group.29 It is again worth noting how our discussions with practitioners and others about the key drivers of violence and how they might be addressed, illustrated the interconnections between those factors and the way that they can impact across different levels. Poor school performance or behaviour is linked to exclusion for example, which in turn reduces the scope for ETE opportunities and for longer term stable employment. Respondents did not always comment on such linkages without probing from us – sometimes beginning with responses like “it’s all about the family” or “it’s all about drugs” or words to that effect. Others verbally mapped out quite complex sets of drivers and protective factors and how these interacted and affected one another (and even how they could be addressed in terms of actual programmes or interventions) but those kinds of exchanges were obviously not typical. 29 These claims are borne out by the official data concerning school exclusions. In one borough Black pupils are excluded at a higher rate than any other ethnic group, and at a rate that is about 2.2 times higher than for White pupils (for both temporary and permanent exclusions). If the figures are grouped into White and non-White, the exclusion rate for non-White pupils is 5.6 times higher than for White pupils for temporary exclusions, and 3.4 times higher for permanent exclusions – full data-sets can be downloaded at: https://explore-education-statistics.service.gov.uk/find-statistics/permanent- and-fixed-period-exclusions-in-england PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 56 4.7.5 Deterrence, lack of repercussions Feedback about violence and young people also tended to vary in terms of the way in which individual decision-making, choice, and responsibility were characterised, and it was clear that different models of offending behaviour underpinned that variation. It is a similar variation highlighted in both academic and public debates about criminal justice, and about sentencing and deterrence, for example. In our own data, on one side there were those who argue that young people who engage in violence will make decisions to engage in such behaviour if they think it will profit them and that there are no realistic consequences of being punished for it. It is claimed that current increasing rates of violence are at least partly due to the fact that young people do not feel that there are any consequences for bad behaviour, and they therefore act in their own self-interest in the absence of fear about repercussions. As one police respondent expressed that view: The only way to stop youths from committing violence is when they have fear of repercussions. Currently, there are no repercussions. We deal with the same minority (but growing) of people. These people often have charges of serious offences on their record yet here they are being dealt with by police again. I've spoken to numerous young people and it is demoralising. There is no fear of the police, nor of any consequence of the criminal justice system. (PRAC -79170331) Another police respondent to the practitioner survey suggested that longer sentences could deter offenders: Longer sentences for convicted offenders. If a person carries a knife clearly any preventive/diversionary measures have failed. The emphasis on treating these often young people as 'victims' of poor circumstance is erroneous, they are contributory factors but law and order must be maintained for the greater good of society. The public perceive that some areas of London are becoming lawless. They have the wherewithal to kill &/or cause serious injury and must be dealt with robustly whenever identified or this endless cycle will continue. (PRAC-79076283) One police respondent (to the practitioner survey) also referred to a necessity for young people to have some “fear of the police” or “fear of repercussions” for violent behaviour, when suggesting what the key characteristics of violence reduction should be: 1. Overhaul of the criminal justice system. There must be consequences to their actions. Over the years we have focused on trying to intervene and deter but it doesn't work. I remember years back a mandatory 5 year custodial sentence was to be implemented for carrying a knife; that could have helped but never came to fruition. 2. More police on the streets. More robust policing with significant support from politicians and senior management. . . . The only way to stop youths from committing violence is when they have fear of repercussions. Currently, there are no repercussions. We deal with the same PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 57 minority (but growing) of people. These people often have charges of serious offences on their record yet here they are being dealt with by police again. I've spoken to numerous young people and it is demoralising. There is no fear of the police nor any consequence of the criminal justice system. . . . I fear we will not solve this issue for a long time; if at all. Unless you are part of the agencies dealing with these people first hand it is impossible to imagine the reality. (PRAC-79170331) Views of this sort tended to be expressed more by police respondents, but not exclusively so. One manager of a voluntary sector organisation suggested that repeated “slaps on the wrist” for knife crime can in effect “teach” young people that they can act without consequence, for example: I think the violence has definitely got worse. I think if you just look at the level of violence and I think in this case stats don’t lie. I think if you look at the stats in terms of definitely knife crime, we are definitely in a worse place. I think the hardest thing is - the biggest question is why? Why has it got worse? Ultimately, I do believe that we don’t hand out harsh enough sentences for young people. So, if a young person under the age of 18 - so they’re still part of the Youth Offending Team, is caught with a knife, you can get a warning. It’s up to three times that young person needs to be caught with a knife, for them to receive a custodial sentence. Well, the third time could lead to someone’s death. So, I could leave my house with a knife and get a slap on the wrist, the second time leave my house with a knife and get a slap on the wrist, the third time kill someone and then go right, 14 years. What are we actually doing the first time to make them know, you shouldn’t be doing this again. So, if we keep giving people chances to do the same thing, we’re humans, we’re programmed, we’re robots, it’s OK, I’ve done it once and I never got in trouble, I’ll do that again. So, ultimately how are we educating these young people not to make the same mistake? So, if you get caught with a knife what do we do? Maybe here’s a prison sentence. Maybe it’s workshops around knife crime. (A-124) The view that a “lax” criminal justice response to violence is a key driver of violence is of course not inconsistent with often expressed views in the wider population, and it was also the case for our own respondents that arguments about the need for stronger enforcement approaches were sometimes made alongside arguments for “upstream” prevention. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 58 Perspectives that were more exclusively criminal justice focused and individualised obviously also rest uneasily with recent developments in the violence reduction field, and with public health approaches which incorporate consideration of broader (and more complex) factors at both community and societal levels. It is also a key aim of public health approaches that an understanding both of key drivers of violence and of interventions to address violence are carefully anchored in evidence. 4.7.6 Family factors Family factors that have been described as being relevant to the development of serious youth violence have already been referred to in some earlier sections of this report – such as on Figure 10 and Figure 11 in section 4.3 for example, which summarise the responses of DSLs to questions about key families issues in the area of their own school – but many respondents singled out one or more of these issues as being key drivers of violence. In particular, domestic violence or abuse was sometimes referred to as being relevant in several ways. Domestic abuse was described as being an effect of previous abuse, but also as a cause of the kind of damage to young people who witness it, which can in turn increase the likelihood of those young people becoming engaged in violence themselves (domestic or otherwise). Part of what respondents said could “cause” this future violence was the way in which domestic violence can normalise violence in the eyes of young people who are exposed to it, but respondents also made reference to the impacts of trauma on such young people. The importance of domestic violence as a driver of future violence was also referred to by practitioners who indicated in our survey that interventions to support young people who witness such violence are of key importance to include in local packages of violence reduction measures (see Figure 13 and Figure 14, which summarise survey respondents’ views on preferred violence reduction interventions). That feedback on domestic violence specifically is also in keeping with feedback from practitioners gathered in other research (e.g., Crest Advisory, 2019). Respondents also referred to other family-related factors as being significant for the evolution of violence, including a lack of role models (especially positive role models for young boys), substance-misusing parents (which can lead to neglect and/or an absence of early childhood nurturing), and sibling or other family member involvement in violence or gang activities. Some respondents also referred to a combination of family factors: So, how do you get a 15 year old that’s say mum and dad, dad’s in prison, mum’s a drug addict, all they see every day is negative stuff. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 59 When they wake up mum’s drunk, mum’s on drugs. When they leave their house on the local estate, people are selling drugs. You become a product of your environment. So, if you see it every day you become desensitised and it becomes normal, that’s your normal. And therefore, it’s really hard when it becomes normal to see there’s anything wrong with it. That’s why I believe that a lot of young people are really desensitised to what they do. (A-124) Of course, the converse of views such as the above can also be found, in respondents’ descriptions of families as being sites where other (protective) factors are in place or can be supported to develop, where such factors can reduce the likelihood of involvement in violence. Consistent parenting and nurturing of children for example, can promote the development of resilience and self-confidence over time, and the presence of positive role models can provide examples to children and young people of possible (positive) futures for themselves. Some of our respondents also commented on the way in which connections with non-parental “significant others” within the family (such as siblings or grandparents) can also give children and young people the kind of connection and support which can build resilience and reduce vulnerability. We return to some of these issues in the following section. 4.8 Protective factors, interactions Taken together, the feedback concerning drivers such as those referred to above also makes it clear that single factors do not usually operate in isolation from other factors, and that they are often interconnected – and in some cases, mutually reinforcing. Drivers of violence can coalesce both at the individual level and more broadly (e.g., at community level or within a school), where they can also interact with factors which have a more positive impact on the development of violence (i.e. “protective factors”). In a lot of our feedback the focus was very much on drivers rather than protective factors, with some respondents noting that an absence of a particular driver would reduce violence – in other words, it was felt that the absence of a risk factor is what was important for reducing the scope for violence to occur. Some also pointed specifically to factors which they felt were important for such reduction however – such as improved individual resilience of people in vulnerable groups for example, positive educational achievement, enhanced employability, greater “social cohesion” in communities that suffer from high levels of violence, and so on. Not many respondents offered clear accounts of why they thought that certain factors were either a risk or had some importance for protecting individuals and communities from violence however, even though much is made about the importance of theories of change for underpinning such work and an understanding of why such work might be effective. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 60 This is linked to the way in which some respondents were also unclear about issues around the effectiveness of interventions to address violence, and why particular interventions that they favoured might be expected to generate positive outcomes. Although not all respondents commented specifically on drivers or protective factors, it is to some extent possible to read those perceptions off of comments about interventions to reduce violence and why these might be effective or useful. We turn to those issues in the following section. 4.9 Addressing violence As noted in section 2 a key focus of the research was on perceptions about current efforts to address issues around serious youth violence and exploitation, about the value and effectiveness of such efforts, and about what else could be done to address violence across London. We questioned all of our respondents about these issues, and key themes in that feedback are outlined below. 4.9.1 Projects, interventions Given the wide range of issues focused on in the research it was expected that respondents would also range widely in their views about the kinds of interventions that might be effective in addressing those, and feedback about these issues from across all the strands of the research was in fact quite varied. We asked respondents about specific interventions that they might be aware of in their own areas or elsewhere (which they thought were promising), and also about the kinds of interventions that they thought might be effective (even if there were no such interventions in their area that they knew about. Concerning specific interventions, respondents were not always aware of actual initiatives that were being implemented in their own areas (to address issues around serious youth violence or exploitation). We would expect this given the range of interventions that actually exist in this field of work, but in some cases it did suggest to us that some interventions could perhaps generate more traction or buy-in by doing more public relations work. The resident surveys in particular did not generate very much in the way of references to specific projects or interventions that the respondents thought were promising, with some even saying in response to our questions that they were not aware of any local initiatives. Many of our survey and interview respondents did refer either to types of intervention that they thought were most important, however, or to specific projects or initiatives that they were aware of. The majority of respondents felt that primary prevention (i.e. work delivered “upstream” to prevent violence from happening in the first place) is perhaps most important, and many respondents made reference in this context to factors having to do with poverty, lack of opportunity, and young people living in circumstances where family and other violence is normalised. The actual schemes referred to in this category ranged quite widely again, from more universal interventions – e.g., GAV- style (“Growing Against Violence”) interventions in schools to raise awareness PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 61 among young people of violence and its impacts – to interventions designed to provide better and more widely accessible youth services, initiatives that aim to strengthen social capital, and so on. Respondents also made reference to secondary prevention initiatives, which focus more specifically on groups that are at high risk of becoming involved in violence or exploitation. Some football schemes are of that sort for example, because they aim to occupy or divert young people who are on a negative trajectory in terms of violence or offending. Other schemes that involve trauma-responsive work to support young people living with domestic violence or parental drug use were also referred to. Schemes designed to build resilience among particular victim groups were also referred to, where the aim was to protect individuals from violence or exploitation in the future. Respondents also referred to measures that could be categorised as tertiary prevention, involving work with known offenders to prevent further violence for example, or criminal justice measures to deter or sanction offenders – unsurprisingly, police respondents usually made reference to the importance of interventions in this category. That typology obviously involves overlaps – GAV-type schemes are universal for example, but they also engage with at-risk young people in particular areas, since within a particular class room there might be a mix of young people, some of whom are on the periphery of gang activity for example, and some of whom are already involved in gang activity or violence to some extent. In terms of how important respondents felt that various interventions were, we asked for feedback about these in our surveys and also in our interviews. In general, views about the importance of different types of interventions were strongly related to sector. Across all sectors, some kinds of interventions were endorsed quite widely by respondents. Family support and parenting programmes were deemed by 95% of respondents to the practitioner survey to be either “very important” or “important”, for example, with only 45% feeling the same way about the use of knife wands in schools. Figure 13 summarises responses across the whole respondent group, to the question “For each intervention listed below, how important do you think it would be for it to be included as part of an overall approach to reducing serious youth violence and exploitation?” PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 62 Figure 13 – Interventions to address violence and exploitation: percentage of practitioner survey respondents regarding them as being either “very important” or “important” (N=96) We also compared responses to these questions between broad sectors, to see how much variation there might be underneath the kinds of figures presented at Figure 13. As expected there were some clear differences in the way that practitioners from different sectors placed importance on different kinds of interventions to address violence and exploitation. We provide one example here, which compares responses by police respondents to the practitioner survey, with responses that DSLs gave to the same question in the DSL survey. In order to make the comparison, we calculated “strength of agreement” scores for each group in relation to each type of intervention, and plotted those as bars for each group at Figure 14.30 30 Those scores are calculated by assigning values to each of the possible responses (“very important”, “important” and so on, but excluding “don’t know”s) and then averaging those for each group for each intervention type. In this case a score of 3 would represent the strongest level of importance possible. We would point out that scores of this kind should be regarded as indicative 95% 94% 93% 92% 91% 90% 90% 89% 82% 76% 76% 74% 67% 65% 55% 45% 0% 10%20%30%40%50%60%70%80%90%100% Family support and parenting programmes Capacity building and resilience training in schools (to help young people reject violence) Provision of youth services Access to good quality mental health services Diverting vulnerable people away from the criminal justice system Support for young people experiencing domestic violence Supporting offenders upon release from custody Meeting the needs of vulnerable offenders Measures to increase trust and confidence in the police Access to good quality drug treatment services More CCTV in hotspots or high crime areas Improved street lighting Access to good quality alcohol treatment More police on the street Using knife arches/wands on the street & transport Using knife wands/arches at schools PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 63 It can be seen from those figures that DSLs tended to place importance on broad social interventions and support measures, particularly those that provide support for vulnerable young people and families. DSLs appear not to have placed much importance on target-hardening and physical measures such as CCTV, street lighting, and in particular, knife wands. Police respondents were likely to highlight the importance of physical measures and to place less importance on social programmes or support programmes – although they felt that capacity-building programmes in schools and family and parenting support programmes were important (in fact there was little difference between the two groups in their ratings for those interventions).31 only, since Likert scale data are, strictly speaking, ordinal data. There is an ongoing debate about this and related issues in the literature (see Norman, 2010 for a brief summary and useful comparison of both parametric and nonparametric approaches to analysing Likert score data). 31 It is worth noting that the size of the two groups was quite different, with police respondents numbering 45, and the DSL group numbering 96. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 64 Figure 14 –Comparison of “strength of importance” scores for DSLs and police, relating to various interventions to address violence and exploitation Differences of that kind came out in more detail in the interviews and in open-ended feedback in the surveys, where responses across the two groups also made it clear how views about interventions might be linked to views about the causation of problems around violence and exploitation. As noted earlier police respondents were more likely to see the value of wider criminal justice measures designed to deter potential offenders, for example (e.g., having more severe sentences for weapon- related violence), and to comment on the need for role models and strong family upbringing. 4.9.2 Partnership working Issues concerning partnership working are obviously of considerable importance to violence reduction activity, mainly because issues concerning violence and exploitation cut across traditional agency boundaries and policy areas. The issues 2.2 2.4 2.1 1.9 2.0 2.7 2.7 2.3 2.5 2.4 2.3 2.0 2.4 2.2 2.2 2.4 2.2 2.1 1.9 0.9 1.2 2.5 2.7 2.5 2.7 2.8 2.4 2.4 2.9 2.5 2.6 2.6 0.0 0.5 1.0 1.5 2.0 2.5 3.0 More police on the street More CCTV in hotspots or high crime areas Improved street lighting Using knife wands/arches at schools Using knife arches/wands on the street & transport Capacity building and resilience training in schools (to help young people reject violence) Family support and parenting programmes Measures to increase trust and confidence in the police Provision of youth services Support for young people experiencing domestic violence Access to good quality alcohol treatment Access to good quality drug treatment services Access to good quality mental health services Supporting offenders upon release from custody Meeting the needs of vulnerable offenders Diverting vulnerable people away from the criminal justice system Police DSLs PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 65 are also anchored in complex “causal chains” which cannot be addressed by single agencies acting in isolation from others. As noted in the previous section (and also in section 4.7), some of the perceived drivers of violence problems have socio-economic and complex area-based correlates, for example, as well as individual ones concerning family and community history. That complexity is acknowledged within most “public health approaches” to violence reduction, and such an approach clearly underpins official violence reduction frameworks and budgets in London, and is also enshrined in recent key policy documents.32 Related notions such as contextual safeguarding and “whole systems” theory also have implications for the involvement of partner agencies, as suggested both in the documents referred to above and in respondent comments of the following sort: Better join up, and a single lens and pathway. Compartmentalisation or silo working must be avoided as children do not recognise the artificial boundaries we put in place and can fall through the gaps. (PRAC-80247515) Safeguarding harm outside the home is so complicated and involves so many more people than harm inside the home and that kind of wider understanding of that feels like it’s the right way forward. (B-075) Feedback about partnership working in relation to violence reduction suggested quite a mixed picture across all London boroughs, and that the quality of partnership working varies widely in terms of extent, profile, and efficiency. Partnership working also continues to be characterised by strong political and agency interests in some areas, which can have a significant effect on direction and sense of purpose. Some practitioners felt that the skills required for partnership working have declined or diminished significantly in recent years across all agencies and organisations including the police, with some claiming that those involved in violence reduction often pay lip service to it rather than getting actively “stuck in to it”. Some interview respondents pointed to an erosion of strengths in partnership working which had previously been developed in the community safety field, further to the Crime and Disorder Act 1998, which of course had made partnership working a requirement for identifying and prioritising crime and disorder issues of concern to local communities. That focus was felt by some to be closer to public health approaches to violence reduction, and that it also allowed greater scope for community involvement than was currently happening in some areas. One local authority community safety manager noted that partnership working had traditionally 32 Key background documents include those released by Public Health England (2019), the cross- party Youth Violence Commission (2018; see also Irwin-Rogers et al., 2020), and the World Health Organization (2020). PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 66 been a key strength of community safety partnerships (CSPs) but that some of those skills had been lost in the wake of reductions in numbers of staff over the years; as they expressed it: Community Safety is diminished and relegated – not now a useful way forward but it should be if it was restored to its former vision and way of working. . . Community safety needs re-instating and a boost. (A-012) “Silo working” and disjointed services were mentioned by several stakeholders as being a debilitating aspect of partnership working, and where that was thought to be an issue it was felt that fractured services led to a fractured sense of purpose. Links between silo working and budgets were also pointed to: Well, I think the silos are the way that we created this system, isn’t it? When I think about [this borough], so you’ve got all of these departments that say that they’re working together but when it comes to resources, that’s when they are going to go back to their silo, isn’t it? When there is a young person that needs some support from education and a bit from social services, people will start talking about their budgets and what they can provide, etc., rather than looking at long-term, what will this mean if we do this piece of work now, if we did this now, what would be the impact? Well, we’d probably be saving ourselves from a lot of other work, a lot of other stuff that we just react to, but we can see it happening. (B-122) In terms of information-sharing, issues concerning GDPR appear to have eroded the effectiveness of data-sharing practices for partnerships in some areas, and to have made some staff more risk averse about such matters – declining to share or “dragging their feet” on such matters until the pressure to share information becomes strong enough to outweigh perceived downsides to sharing. Respondents sometimes referred to a perceived lack of leadership or direction here, and a lack of clarity or information about best practice. Where strong data-sharing practice seems to exist it is often informal and can turn out to be short-lived, as one or more of the transactional partners moves on. I just think silo working is an absolute issue in every area, whether that’s internally, within strands within the Police, whether that’s expert intel, you know, I just think that the more we can look to problem solve whilst making sure that we’re compliant in relation to data protection, everything like that, which is quite simple to get through, you know. . . . As long as you’ve got a good knowledge of GDPR most of this stuff is doable with the proper information sharing agreement that you can make sure, you know. (B-156) Respondents also referred to the need for more shared IT systems across partners, as these systems currently do not always allow for seamless sharing from system to system. This was sometimes even within a single organisation and made worse by “silo working” in local government and other large organisations such as health and police. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 67 A related point is that crime analysis ability across local partnerships and CSPs seems to have been depleted – at least since periods where a substantial amount of analysis was required for the generation of area community safety profiles and so on. Some suggested that cuts had led to an erosion of the Crime Analyst “profession”, and that staff involved no longer had a sense of professional unity. Partnership working also appeared to be functioning very well in some areas, and particularly at operational levels, and in relation to case work and problem-solving in particular: I’m very lucky on this . . . I’ve arrived to the, a very, very sleek functioning system, whereby there are multiple meetings and engagements between all of our partners. From Social Services, Housing, DWP, Probation, you name it, they’re there, and there’s a really, really collaborative look at how we reduce risk and how we look after people. There’s also, we’ve got the integrated offender management units, so I’ve used IOM as well. So, you know, our relationship’s both internal with our partners and external with other government agencies, you know, it’s really, really strong here. And it’s just seen as a, these are people on a list that are a risk, at risk of violence or committing violence and how do we manage that best. So I’ve been really, really impressed with what I’ve inherited, so I can’t claim any credit but it works really well here, in terms of our partnership working and sharing information where we can. And it, I suppose, it’s a bit of a dare to share principle. . . . A few years ago everything was siloed, keep it to yourself. And then we realised, off the back of various inquests . . . . and the same with frontline policing, that the more you can actually share, the more chance you have of achieving results. So it’s about sharing safely in the right forums and that’s what’s kind of in place here in terms of governance. (B-065) There is a perception that there are too many voluntary sector agencies involved across the whole crime landscape making funding difficult and leaving the third sector space over-crowded and with too many people doing the same thing. Many frontline agencies or small, new agencies were just not getting access to funding. This was often because they cannot produce three years of audited accounts, let alone good quality evaluation reports. However, many stakeholders felt more could be done here to help fund and resource such groups as they often are the ones with the best reach into the young people who are most active on the streets or most difficult to reach. Concerning the timescales for available funding, a majority of practitioners noted that short-term funding challenges are a real hindrance to local charities and projects, and it was also suggested that short-term funding eroded effectiveness. The London VRU was also commented on in remarks about partnership working in violence reduction, and although the VRU received some criticism for “trying to please too many masters” or for being “too remote from communities and real-life PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 68 issues”, it also received praise both for the way in which it organised and deployed funding opportunities for promising scheme to address violence and exploitation, and for the way in which it has promoted smoother partnership working in some areas. In relation to data analysts for example, several respondents commented positively on the way in which efforts had been made to “pull together” analysts from different parts of London, in order to generate stronger overall data-sets. Schools, colleges and universities were not thought to be involved in partnerships to the extent that they could be, although DSLs obviously need to engage in a degree of partnership working as part of their work. Their perception of partnership working and its effectiveness was fairly positive, although views again did vary from area to area. Their responses to our main questions about their own partnership working are summarised at Figure 15. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 69 Figure 15 – DSL survey: “As a DSL, you will occasionally work with external partners on safeguarding issues. To what extent would you agree or disagree with the following statements:” (%, by response category) (n=180; not answered=1; to improve readability, data labels for values of 2% or less have not been displayed) While interview respondents were quick to raise issues about the difficulties of partnership working in relation to violence reduction, respondents to our practitioner survey were fairly positive about particular aspects of partnership working, in their responses to some of our questions about the quality of that work. 17% 18% 12% 7% 8% 17% 12% 9% 55% 54% 42% 33% 26% 41% 33% 29% 19% 19% 26% 34% 16% 23% 33% 24% 7% 8% 16% 19% 39% 14% 11% 24% 4% 3% 11% 4% 11% 9% 0% 10% 20%30% 40% 50%60% 70%80% 90%100% Together with our safeguarding partners, we are an effective problem-solving group Our safeguarding partners have a shared vision and goals Our information-sharing arrangements work effectively Together with partners we have pooled resources to achieve the greatest advantage With partners, we have sufficient staff and other resources to address safeguarding issues effectively With partners, we have sufficient training and development opportunities to help us address safeguarding issues effectively Our safeguarding partners operate in a trauma- informed way Local authority “threshold levels” assist me in my role Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree Don't know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 70 In relation to key issues such as joint problem-solving, having a shared vision and goals, information-sharing, and consultation with local communities for example, from 70% to three quarters of respondents either “agreed” or “strongly agreed” with the relevant statements, which is clearly quite a positive response overall. Levels of agreement were somewhat lower for statements concerning some of the other aspects of partnership working - such as a shared focus on cost-effectiveness, for example, or having a shared understanding of links between trauma and violence. Details are summarised at Figure 16. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 71 Figure 16 – Practitioner survey: “It is often noted that work to address issues around serious youth violence and exploitation requires the involvement of a range of agencies. To what extent would you agree or disagree that partners involved in addressing these issues in your own area:” (%, by response category) (n=94; not answered=3; to improve readability, data labels for values of 2% or less have not been displayed) 47% 40% 42% 31% 44% 35% 24% 33% 36% 41% 29% 33% 26% 33% 33% 30% 18% 24% 16% 26% 34% 27% 32% 29% 11% 8% 8% 14% 5% 17% 29% 20% 9% 14% 15% 14% 7% 9% 9% 18% 17% 15% 17% 9% 11% 9% 10% 11% 7% 9% 9% 6% 14% 6% 9% 9% 7% 7% 10% 9% 5% 4% 3% 5% 5% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% problem-solve together effectively have a shared vision and goals share information effectively pool resources to achieve the greatest advantage have sufficient staff and other resources to address these issues keep track of key interventions and their effectiveness keep track of the cost-effectiveness of their violence reduction work adopt a public health approach to violence reduction work ensure that their work is informed by the interests and concerns of local communities have a shared understanding of connections between violence and trauma have a shared understanding of the importance of gender, to the contours of violence and its impacts base their violence reduction activities on evidence about what works Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 72 4.9.3 Policing Although policing was not a direct focus of the research, it is clearly of central importance to efforts to address violence and exploitation, and issues concerning policing obviously came up a lot in the feedback. In terms of direct feedback about policing from young people, we did ask young people in our surveys to give us their opinions about the police, and some of that feedback is highlighted below. The percentages of respondents who claimed that they had either a very bad opinion or a bad opinion of the police seem quite high, and it is also not very positive that the most common response (selected by 39% of respondents) was “neutral opinion”. Again, it is hard to extrapolate to wider populations of young people, but these figures do seem quite negative. Figure 17 –Young people survey – “Overall, what opinion do you have of the police?” (% by response category) (n=77) It is useful to compare these figures with findings from MOPAC’s own Youth Voice Survey (Ramshaw et al., 2018), since that survey covers (almost 8,000) young people aged 11-18, and covered all boroughs in London. Since that survey is administered to school-based populations we would expect that feedback would differ in some key ways from our own feedback, since our questionnaires were completed by young people who are more likely to be somewhat older and also to have been excluded or otherwise not be attending school. That expectation seems to be borne out in relation to perceptions about the police - while 43% of our sample had either a very bad or a bad opinion of the police overall, only 14% of the MOPAC respondents had a “bad” opinion of the police. 14.3 28.6 39.0 11.7 3.9 2.6 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 Very bad opinion Bad opinion Neutral opinion Good opinion Very good opinion Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 73 The other survey question where we did ask directly for young people’s views about the police involved asking for each respondent’s level of agreement with each of seven statements about the police or police practice. The feedback was fairly mixed, as indicated on the following graph. On the one hand, it does appear that whatever they think of the police, most respondents would make sure that incidents of serious violence were reported to the police (if they saw any such incidents) – 57% either agreed, or agreed strongly with the relevant statement. A sizeable proportion of respondents (though not a majority) also felt that “the police will be there if I need them” (with 38% either agreeing or agreeing strongly), although the response relating to “the police keep me safe” was a bit more equivocal (with 30% either agreeing or agreeing strongly). Forty-three percent of the MOPAC respondents felt that the police “will be there when I need them”, and 47% felt that the police “can protect me from crime”. Levels of agreement with the statement “The police treat everyone fairly whatever their skin colour” were somewhat more worrying in our own survey, with only around 12% agreeing or agreeing strongly, and just over two thirds of respondents expressing disagreement or strong disagreement. That finding also resonates with some of the comments made in response to the open-ended questions (see below). Figure 18 –Young people survey – “Do you agree or disagree with the following statements about the police?” (% by response category) (n=77) 4 8 3 7 8 25 11 26 14 9 32 22 32 33 27 31 21 33 30 16 21 25 22 32 18 18 12 18 14 21 36 9 19 10 9 4 4 1 3 5 8 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% The police do a good job in the area where I live The police are helpful and friendly towards young people The police treat everyone fairly whatever their skin colour The police will be there if I need them The police keep me safe If I saw any serious violence I would make sure that it was reported to the police The police respond effectively when serious violence is reported Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree Don’t know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 74 Again by way of comparison, the MOPAC youth survey respondents were somewhat more positive than our respondents, with 37% agreeing that the police “treat everyone fairly”, and only 27% disagreeing (Ramshaw et al., 2018). Although the open-ended questions did not ask directly for comment about the police, some of the comments made are also relevant to issues around policing. The comments below were made in response to the question: “In just a few words, what do you think is the most important thing that could be done, to stop or reduce violence?” Police or policing are mentioned 22 times (with further mentions using other terms – such as “law enforcement”), and there is roughly an even split between positive and negative comments. Again there is a view that police need to be more understanding of young people, and that the police need to pay attention to racism and relations with the community – but there is also a strong theme about the need for more policing, and more targeted policing (e.g., arresting drug dealers, being on visible patrol more often, and so on). Eighteen respondents made further comments in response to the question: “If you would like to make any more comments about any of these issues, please use the space below”, and some of those comments make direct reference to the police – for example: ➢ Fuck the police. ➢ I’m satisfied with police. They are always there when we need them. ➢ less stop and searches. ➢ police create the violence with picking on certain groups they do nothing. ➢ Stop and search is getting out of hand in the area. ➢ The acceptance of racism in the police force is necessary to move forward. The rest of the main young people surveys do not include much on the police, although one other question does ask for feedback about the frequency of police arrests in the respondent’s own area. Responses to that question are summarised below. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 75 Figure 19 –Young people survey – “The next few questions are about things that happen in many areas of London. Please say how often these things happen in the area that you live in.” (% by response category) (n=77) The respondents do seem to see arrests quite a lot, although it is difficult to assess these numbers on their own. (It is also interesting that it seems to be a regular feature of everyday life for these young people, that they see either drug-taking or public use of alcohol.) Comments about policing and racism were not limited to our young respondents, with some practitioners and residents also referring to a need to address disproportionality and what they saw as the racism that underpins it. The following comments from one of our DSL respondents illustrate such a view: The unnecessary aggressive and racist attitudes of the police towards young people of BAME heritage is fuelling some of the issues of violence and trauma that is being experienced in communities. This is also leaving some young people and their families traumatised as well as criminalised and Local Authority Safeguarding arrangements and expertise do not have either the means or sufficient expertise to challenge or address this. This is a systemic issue and a consequence of institutional racism. (DSL-75487072) Internal confidence amongst police staff appears to be quite variable. A view by many is that recent changes have diluted efficacy, efficiency and professionalism. Some longer serving officers felt that the MPS had changed significantly now and lost too many older, seasoned and knowledgeable staff above Inspector rank. 36% 9% 35% 12% 6% 22% 29% 19% 27% 43% 22% 12% 10% 18% 13% 22% 29% 19% 27% 43% 5% 32% 13% 14% 13% 0% 10%20%30%40%50%60%70%80%90%100% You see people drinking alcohol on the street Someone gets robbed You see someone using drugs You see the police arrest someone You see a fight Almost every day A few times a year Once or twice a week A few times a year Never PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 76 Some police respondents felt that new BCU arrangements across the capital have affected ability, productivity, confidence and outcomes significantly over the past few years, although some boroughs report moving on a bit now from such an ‘unwelcomed interruption’. Others report that the situation has still not settled. There was some criticism as regards the impact of moving to tri-borough BCUs in terms of local partnership working with CSPs with several areas saying this generated some confusion. It was noted by some police respondents that engagement with particular communities had suffered over the past couple of years in some boroughs, and fallout from public issues about the Matrix was referred to in this context. Policing resources both centrally and in each borough were thought to be very variable with a commonly heard critique that resources go to key boroughs and hotspots only, whilst other areas are left to fend for themselves. Some also commented that resources more generally were spread too thinly across London. Confidence amongst young people towards the police is recognised by stakeholders as being very low, and our own feedback from young people generally supports that view, although expressed lack of confidence in the police obviously varies by area and also by age and ethnicity. Among residents in our case study areas, confidence in the police also appears to be low, as reflected in both our direct interviews with residents and in feedback from our online surveys. Interview feedback and open-ended feedback from some residents who have responded to the survey, is quite negative in relation to experiences around reporting incidents of violence. Views on stop and search were quite mixed across respondents, with some arguing for more but more focused stop and search activities, and others arguing that the negative impact of stop and search on community and youth relations (especially in some BAME communities) is not outweighed by any particular gains that the police might get from that activity. Comments such as the following were typical: More strategic use of stop and search. Being searched and delayed on the way to school or after school several times a week and having nothing illegal on their person is traumatic. (DSL-78801890) Other respondents also drew connections between disproportionality, racism and trauma at community level (e.g., the DSL quoted on the previous page), and a small number linked policing and the criminal justice system itself with wider social control of non-White groups. Again, these were minority views, but strongly expressed to us on several occasions. It is worth noting that some of the above feedback is consistent with findings from other research that has focused on levels of confidence in the police across London. A recent review by the Police Foundation found that overall attitudes toward the police have become more negative in general recently, for example, although there PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 77 have been clear differences in those attitudes across different ethnic groups (Muir et al. 2022). Similarly, a survey of Londoners conducted by YouGov (Mile End Institute, 2022) suggested that just under half of all respondents (N=1,114) had either “no trust at all” or “not very much trust” in the Metropolitan Police, with trust levels again varying significantly by ethnicity – the percentage for respondents from BAME communities was 54%, for example, and a large majority of those respondents also felt that the Metropolitan Police service is either “definitely” (43%) or “probably” (29%) institutionally racist. MOPAC’s own Public Attitude Survey returns also show that perceptions of fairness and trustworthiness of the police among Black and Mixed ethnic groups fell further recently and from a lower starting point than elsewhere in the UK. In response to the key question about police “fairness”, for example levels of agreement fell for the latter groups, from a point one year ago that was already significantly lower than for respondents from White groups.33 Concerning stop and search specifically, it is well known that people from BAME communities are more likely to experience it, more likely to be subjected to use of force as part of a stop and search encounter34, and less likely to claim that stop and search as currently practiced is justified or should continue. In a recent IOPC public perceptions tracker, only 28% respondents from BAME communities claimed the latter for example, as opposed to 43% of white respondents (Naseem, 2021). 33 Details can be found on MOPAC’s Public Voice Dashboard. The survey is noteworthy for its methodological rigour and for the robustness of the final data-sets that it generates, and these are therefore often used for a variety of purposes in the literature. 34 Gaps in the available data concerning use of force during stop and search mean that claims about disproportionality are indicative only, but reports such as the recent HMICFRS (2021) publication on force offer some support for such claims; the authors suggest that “despite limitations, the data suggests a disproportionate use of force” by ethnicity. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 78 4.10 Perceptions on how to reduce violence As part of the research, we asked all respondent groups for their views about how to reduce violence and the form that work takes. We also asked respondents for their views about what the likely trajectory of such work might be, and how likely they thought it would be effective. Overall, the research team found it quite striking that levels of confidence were so low among many respondents about the likely effectiveness in reducing violence in the future. Interview respondents were fairly restrained in their expectations about how to reduce violence, in either their own areas or London generally, and survey respondents were even less positive. Practitioner survey respondents were asked to rate their level of confidence that future efforts to reduce violence and exploitation would be effective, from 1 (meaning not at all confident) to 5 (meaning very confident), and the average was about 2.8 (for those having a borough focus in their own work) and 2.2 (for those with a pan- London focus). Respondents were also asked how likely they thought it was for each of a range of factors, that it would blunt the effectiveness of efforts to reduce violence in their area or in London generally. For “lack of resources”, 94% of respondents felt that it was either likely or highly likely that this would reduce the effectiveness. Figures for other factors included “lack of understanding across partners about the causes of violence” (77%), “lack of awareness across partners about what would be effective” (78%), “lack of longer-term planning and oversight” (82%), “lack of political will to do what it takes to address these issues” (84%), “lack of shared commitment across partner agencies to reduce violence” (73%), “work to address serious violence and exploitation is not properly joined up” (89%), and “lack of understanding across partners about public health or ‘whole systems’ approaches to violence reduction” (77%). Responses from practitioners to those issues are summarised in the following figure. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 79 Figure 20 – Practitioner survey: “For each of the following factors, how likely do you think it is that it will undermine the effectiveness of future work to reduce serious violence and exploitation?” (% by response category) (n=94; to improve readability, data labels for values of 2% or less have not been displayed) Perceptions about the future reduction of violence in London were also provided in more detail by practitioner survey respondents in their open-ended comments about these issues (which were provided by 60 out of 97 respondents). Feedback from residents and professionals working in our case study areas reinforced that view, with residents in particular reporting low levels of confidence that future efforts to reduce violence in their areas will lead to real change. Interestingly, respondents to the residents’ survey expressed views about the likely trajectory of future general violence efforts in their own areas, which were quite similar to the views of practitioner survey respondents. The question for residents was worded in a slightly different way, but 41 of 43 people who responded to this question felt that “lack of resources” would make it either likely or very likely that the effectiveness of reducing violence would be undermined, and for “lack of political will to do what it takes to address these issues” the number responding in this way was 40 out of 43. In some residential areas, low levels of confidence about the future reduction of violence was clearly related to a perceived history of authorities failing to act, and as 67% 39% 41% 43% 63% 30% 44% 32% 27% 37% 37% 39% 21% 43% 45% 45% 4% 16% 16% 13% 10% 22% 9% 15% 6% 4% 4% 7% 5% 3% 6% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% lack of resources lack of understanding across partners about the causes of violence lack of awareness across partners about what would be effective lack of longer-term planning and oversight lack of political will to do what it takes to address these issues lack of shared commitment across partner agencies to reduce violence work to address serious violence and exploitation is not properly joined up lack of understanding across partners about public health or “whole systems” approaches to violence reduction Highly likely Likely Unlikely Highly unlikely Don't know PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 80 many expressed it, failures to honour promises made to provide support or to regenerate areas and develop opportunities. One resident made reference to a lack of services, and to a need for investment and “upstream thinking”: There is no simple quick fix, it will take investment and upstream thinking which the government don't seem to be interested in. Locally only so much can be done without the money and staff to be effective but ironically a happy and prosperous community would be able to contribute more. It's a very sad situation and I'm glad I'm not a young person living here at this time. (RES-168777014) Other residents also referred to a need for longer-term investment: Solving youth violence will take long term solutions and needs long term commitment and funding. This isn't an issue to keep on chopping and changing approach. (RES-169127472) The council and partner agencies do nothing but tick boxes to collect funding while ignoring the causes of many of the problems in the community because that requires too much long-term commitment. (RES-168698054) Some residents simply claimed that they did not know what future efforts to address serious youth violence might look like, or that some of the causes of such violence were perhaps too difficult to address: I just think there is currently a serious lack of resources or political will to do anything about these problems. I am centre-left and I feel that the Police services have been excessively cut. There is a total lack of any community policing. I would like to see more services to help young people - clubs, sports and recreational facilities. I know there is a problem with the breakdown of families as well and absent fathers... how do you address that? (RES-169115529) 5 DISCUSSION Although there are clearly some limitations concerning generalisability for some of the findings (because of the way in which respondent numbers were achieved for several of the fieldwork strands), the research has generated a strong and unique data-set covering a wide range of respondent groups, and focusing on a variety of key (and interconnected) issues. Using mixed methods has also allowed for the development of a final data-set that has both depth and scope. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 81 While some of the feedback confirms or elaborates what we already knew from other research, some of it has also usefully either highlighted some key gaps in existing knowledge, or yielded new insights into how some of the key determinants of serious youth violence might interact and influence one another. In the former category, the final data-set again highlights the importance of learning more about how girls and young women are connected with serious youth violence and gang activity, and what the key features are of individual pathways into such involvement. We would suggest that what has been referred to as the “invisibility” of girls and young women in gang activities could be addressed not just by examining those forms of involvement directly, but by viewing those activities (and serious youth violence more generally) through a gender lens, in order to developed a more nuanced evidence base which can allow us to see gendered interactions and pathways. At the time of writing MOPAC is already responding to address some of these issues, by commissioning new work to examine pathways into gang activity taken by girls and young women and to identify factors which might protect against or facilitate the involvement of girls and young women in such activity. That research will be timely and highly relevant to violence reduction in London. Another gap that has been highlighted by the feedback that we have gathered also concerns the identification and theorisation of protective factors which might insulate otherwise vulnerable groups from becoming involved in violence – factors which perhaps strengthen individual resilience and reduce vulnerability. It is crucial to understand factors of that kind in much greater detail, since we know that many young people who are exposed to whole sets of violence drivers do not in fact take pathways that lead them into involvement in gang activities or violence. Some of the feedback offers us hints about what those factors might be and how they might have positive effects at key points in an individual life history – such as descriptions of “significant others” in the lives of young people for example, individuals who perhaps offer a combination of individual connection and support, key role models, and some continuity of interest at key points where a young person might otherwise lack those things. As noted earlier, we already know that simple equations such as poverty=violence are not very useful, but a deeper understanding of protective factors of this kind would allow us to understand individual pathways into and out of violence much more clearly, and to design interventions to allow us to test such connections more specifically. The research has also shown us how some small areas of London are particularly affected by high levels of serious youth violence alongside the presence of a whole range of drivers at both individual and community level, and a study of individual pathways into and out of violence could usefully allow for some focus on the interplay of factors at different levels – the way in which individual factors resonate with the impacts of poverty and/or disproportionality for example. Other developments in some of those areas can add even more complexity to these problems - by co-locating vulnerable groups for example, as in areas with a high density of HMOs. Further research focusing on shifts of this kind in the “vulnerability landscape” in small geographic areas over time could help us to understand how PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 82 new and concentrated risks can sometimes be formed, and help us to anticipate such changes and develop partnership structures for avoiding and/or managing them. In relation to issues about links between violence and trauma, the feedback suggests that much “trauma-informed practice” in this field is quite backward looking and overly fixated on static past events (or sets of events), rather than on current struggles that young people in particular might have right now. Not everyone with sets of adverse events in their childhoods or adolescence struggle with the impacts of trauma (though many do), and ACE-focused assessments can therefore lead to inappropriate targeting and to a focus which has been shown in other evaluations not to combine well with efforts to establish trust relationships between young people and service providers for example. A more general point about that issue is that “trauma-informed practice” is arguably moving in a direction that is less positive than it could be in this field. The current Youth Endowment Fund call for new delivery and evaluation of trauma-focused interventions is to be welcomed, but the description of the proposed work is arguably unnecessarily focused on efforts to actually address trauma directly (e.g., via clinical methods). Some of our respondents have also commented on this, but it is possible for violence reduction efforts to address trauma-related drivers without having a clinical focus at all – in our view, some prison-based interventions have quasi- therapeutic outcomes (including aspects of trauma resolution) for example, in some cases even where they do not focus on trauma at all. We also know from the wider literature that interventions designed to address some of the impacts of adverse childhood and adolescent experience can be both effective in terms of reducing a whole range of problems including violence, but that such interventions can also generate significant financial savings in the longer term. Much of this evidence comes from interventions that were designed to address specific adversities such as abuse, neglect or domestic violence for example, rather than to address adversities that co-occur in particular areas in the manner we have referred to above in relation to the notion of “community trauma”. Further work is necessary to help us understand how drivers of violence can coalesce and reinforce one another in very localised communities, and also how community activism and local pride in an area can be harnessed to build resilience and a broader and deeper sense of connectedness in such areas. There is some evidence that social cohesion can be a protective factor in high violence areas, but it would be useful to know more about which forms of social cohesion might have the strongest violence reduction potential, and how these can be developed over time. This leads to a more general final point, that some violence reduction in London could benefit from stronger anchorage in robust theoretical frameworks. We have examined a wide range of theories of change that have been developed in this field, and our respondent feedback overall also suggests that there is some scope for key stakeholders to be given more support for considering why they think that a particular intervention might have the impact that they hope it will, and how identified drivers of violence might have their causal effects. Theories of change that have been developed to map out violence reduction and its impacts are in some cases PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 83 less useful than they could be, for underpinning local violence reduction strategies and for raising awareness about how programmes can be effective. This could in turn help to focus future research and to inform evaluation frameworks, but also to do more nuanced and finely-tuned cost-benefit studies. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 84 6 APPENDIX – METHODOLOGY: FURTHER DETAILS We provide further details here concerning sampling, selection of case study areas and fieldwork within them, and data-analysis. Sampling For all of the fieldwork (interviews, focus groups, and online surveys) we worked through our own existing networks to build contact lists but were also assisted where appropriate by colleagues at MOPAC and/or the London VRU. The latter two organisations each have extensive mailing lists, and some of those contacts were used to communicate with local stakeholders about the research, and to invite their participation in it. In some cases where representation by sector was thought to be imbalanced, we made further contacts (some through our MOPAC colleagues, and others through our own networks) to improve coverage by further targeted communication with key stakeholders in that sector. For the survey of Designated Safeguarding Leads (DSLs) we first worked through a voluntary organisation that had good existing school links in London, and we supplemented those respondents by working through the London community safety networks. In cases where we did not receive any responses from a particular borough, we made separate contacts directly with key stakeholders in that borough to enlist their help in encouraging response. For the surveys of young people and of street workers, we worked through our own networks of youth clubs and youth support organisations in order to generate participation, and we also attended youth events (e.g., football events) to set up stalls where we provided information about the surveys and how to participate. Decisions about which respondents to interview and which to involve in one of the surveys was dependent on assessment of the achieved sample of interviewees as it developed over time (with it being our aim to have an achieved sample that was broadly balanced by area and sector for example), but also on practicalities (such as switching to the survey for those who could not easily be contacted for interview for example, or by interviewing some practitioners who were simply very keen to speak with us rather than use a survey link). Because of the way in which the surveys were conducted, it is not possible to calculate response rates. For each survey there was some internal cascading within organisations or agencies, and invitations to be involved were forwarded on our behalf by MOPAC or the VRU – where because of GDPR constraints we did not see the full contact lists ourselves. Our own contact lists also grew incrementally over time, and the survey responses were in any case anonymous. In terms of survey design, both the practitioner survey and the survey of Designated Safeguarding Leads were designed using JISC software, and the other surveys were designed using the SmartSurvey Enterprise platform. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 85 The young people and the resident surveys used both survey links and QR codes. For the surveys of young people we worked with local youth projects and/or had a presence at key events for young people, where we distributed survey links and/or QR codes. All of the young people and residents’ surveys were incentivised. The surveys and the interviews and discussion groups all had bespoke questionnaires or topic guides, and although we have not attached them to this report (as there are several dozen across all methods), they covered broadly the same issues that are reported on in the main sections of the report. DSL survey response Concerning the spread of DSL responses across London boroughs, responses numbered from 0 (5 boroughs) to 25 (for Wandsworth). The average number of DSL responses per borough was 6. Figure 21 summarises the type of school that respondents worked in, and PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 86 Figure 22 summarises the level of the establishments worked at. Figure 21 – Respondents to DSL survey: type of school or college worked in (N=181) 38% 1% 3% 7% 44% 2% 4% Academy Alternative provision school Free school Independent school Local-authority-maintained school Pupil referral unit Other PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 87 Figure 22 – Respondents to DSL survey: level of establishment worked at (N=181) Interviews Almost all of the interviews were digitally recorded, and transcribed fully for importation into NVIVO for analysis. Most of the stakeholder interviews were conducted using Microsoft Teams or similar, although some were conducted “on site”. The interviews with young people were all conducted face to face, either in local youth clubs or other community locations, or in agency premises. Project summaries and information sheets were provided to stakeholder respondents in advance, and for the interviews with young people we provided relevant details on the day. Our standard details concerning confidentiality and anonymity and reporting were also given again verbally prior to each interview, for all respondents. Case study areas In selecting the case study areas we used an assessment tool developed by the GLA Community Intelligence Unit (CIU) (in conjunction with MOPAC and the London VRU); we examined ward rankings based on multiple sources of data relating to crime, public perceptions, deprivation, and public health. Using four different approaches to combining data on those key issues, the tool generates scores which are then used to rank wards in terms of overall need. The intention was to conduct interviews and online surveys in all six of the sample areas selected – Angell Town/Loughborough estates, Lambeth; Brandon estate (Newington Ward), Southwark; Noel Park estate, Noel Park ward, Haringey; Fairfield ward, Croydon; Grahame Park Estate, Colindale ward, Barnet, and Shadwell ward, Tower Hamlets. After several months we dropped one of the areas (Shadwell ward) on the advice of local contacts, but we did conduct interviews and online surveys in all five of the other areas. 59% 32% 3% 6% primary secondary post-secondary Other PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 88 While the interview component of the case study work was fairly straightforward to conduct and very useful in terms of the feedback generated, the online surveys were very difficult to implement due to a fairly negative local response. In all we designed 7 surveys - 2 for Lambeth (one each for Angell Town and Loughborough estates), 2 for Graham Park (one for adults and one for young people), and 1 adult survey for each of Brandon Estate, Noel Park Estate, and Fairfield ward - and after much consultation we launched 5 of them. Two received zero responses in spite of assistance from local contacts, although in the end we did receive 47 responses from across 3 areas. There were several reasons for this low response, as we note in the main report at section 4.4. Analysis Survey data-sets were analysed using SPSS, but were also imported into NVIVO so that the open-ended material was easier to analyse, and also so that cross-tabbing and cross-referencing by topic area could be undertaken more quickly across data types. For understanding the final data-set we adopted a reflexive approach to our thematic analysis, in keeping with the approach pioneered by Braun and Clarke and branded as “reflexive thematic analysis” or RTA35. That approach involves six stages (data familiarisation, generating initial codes, generating themes, reviewing themes, finalising thematic framework, and reporting) although in practice these stages are not strictly chronological. We also truncated their second stage because we began with an initial coding structure which was linked both to our research questions and to the structure of our data-collection instruments. But RTA is paradigmatically flexible, and it allows analysts to vary quite a bit in terms of where they sit on the scale from essentialist to constructionist for example, what the balance is between deductive and inductive analysis36, whether it is felt that respondents’ comments are a reflection of their experience or constitutive of it, and the extent to which it is felt that analysis should focus on those comments in terms of surface or latent meanings. 35 Braun and Clark first outlined the RTA approach in their 2006 paper, but have since elaborated on the approach at much greater length (2012, 2013, 2014, 2020). 36 With inductive approaches tending to use open-coding for example, as in “grounded theory” approaches to thematic analysis (famously popularised by Glaser and Straus, 1967), and deductive ones tending to use preconceived coding structures. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 89 7 REFERENCES Abianda (2021) Abianda’s reflections on the term “gangs”. 31 May, 2021. London: Abianda. Available at: https://www.abianda.com/post/what-does-abianda-mean- when-we-say-gang-read-our-reflections-on-the-term-gang-to-find-out-more ACMD (Advisory Council on the Misuse of Drugs) (2015) Prevention of drug and alcohol dependence - Briefing by the Recovery Committee. Available at: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attach ment_data/file/406926/ACMD_RC_Prevention_briefing_250215.pdf Andell, P., and Pitts, J., (2023) The Palgrave Handbook on youth gangs in the UK , Palgrave (forthcoming). Bottoms, A.E. (2009) Disorder, order and control signals. British Journal of Sociology, Vol 60 (1):49-55. Bottoms, A.E. and Wilson, A. (2007) Civil Renewal, Control Signals and Neighbourhood Safety’ in T. Brannan, P. John and G. Stoker (eds.) Re-energizing Citizenship: Strategies for Civil Renewal, Basingstoke: Palgrave Macmillan. Braun, V., and Clarke, V (2006) Using thematic analysis in psychology. Qualitative Research in Psychology 3(2), 77–101. Braun, V., and Clarke, V. (2012) Thematic analysis. In: Cooper, H., Camic, P.M., Long, D.L., Panter, A.T., Rindskopf, D., Sher, K.J. (eds.) APA Handbook of Research Methods in Psychology, Research Designs, vol. 2, pp.57–71. American Psychological Association, Washington. Braun, V., and Clarke, V. (2013) Successful Qualitative Research: A Practical Guide for Beginners. Sage Publications, Thousand Oaks. Braun, V., and Clarke, V. (2014) Thematic analysis. In: Teo, T. (ed.) Encyclopedia of Critical Psychology, pp. 1947–1952. Springer, New York. Braun, V., and Clarke, V. (2020) One size fits all? What counts as quality practice in (reflexive) thematic analysis? Qualitative Research in Psychology. Volume 18, 328- 352 - Issue 3: Quality in qualitative approaches: Celebrating heterogeneity. Brennan, I., Burnley, N., Cutmore, M., Holt, A., Lillis, J., Llewellyn, J., MacLeod, S., Shah, M., Van Zanten, R. and Vicentini, L. (2022) Comprehensive needs assessment of Child/Adolescent to Parent Violence and Abuse in London. London: Violence Reduction Unit. Available at: https://www.london.gov.uk/sites/default/files/comprehensive_needs_assessment_of_ child-adolescent_to_parent_violence_and_abuse_in_london.pdf. Browne, K.D., Green, K., Jareno-Ripoll, S. and Paddock, E. (2022) Knife crime offender characteristics and interventions – A systematic review. Aggression and Violent Behaviour 67 (2022). PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 90 Paul Bywaters, P., Skinner, G., Cooper, A., Kennedy, E. and Malik, A. (2022) The Relationship Between Poverty and Child Abuse and Neglect: New Evidence. Nuffield/University of Huddersfield. Available at: https://www.nuffieldfoundation.org/wp-content/uploads/2022/03/Full-report- relationship-between-poverty-child-abuse-and-neglect.pdf. Cordis Bright (2015). Preventing youth violence and gang involvement: A review of risk and protective factors. London: Early Intervention Foundation. Available at: https://www.eif.org.uk/files/pdf/preventing-gang-and-youth-violence-risk-protective- factors.pdf Crest Advisory (2019) Serious Violence in Islington: A Deepdive. December 2019. (Powerpoint presentation, provided to the research team.) Davies, T., and Dawson, P. (2022) London Gang Exit: Final Impact Evaluation Report. London: MOPAC Evidence and Insight. Available at: https://www.london.gov.uk/sites/default/files/final_lge_pdf_190122.pdf. Disley, E., and Liddle, M. (2016). Improving Understanding of Urban Street Gangs: An Assessment of Local Experience in Ending Gang and Youth Violence Areas. London: Home Office. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/49180 2/horr88.pdf Education and Training Foundation (2021) Survey of safeguarding staff in post-16 providers in England: March 2021 - Report of a survey into the skills, needs and views of safeguarding staff in the learning and skills sector. Available at: https://www.et-foundation.co.uk/wp-content/uploads/2022/02/Safeguarding-Survey- Report.pdf Faulkner, S. (2021) Knife Crime in the Capital - How Gangs are Drawing another Generation into a Life of Violent Crime. London: Policy Exchange. Firmin, C. (2020) Contextual Safeguarding and Child Protection: Rewriting the Rules. London: Routledge. Francis, C. and Wilcox, P. (eds.) (2013). The Oxford Handbook of Criminological Theory. New York: Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199747238.001.0001. Gamble, J. and McCallum, R. (2022) Local Child Safeguarding Practice Review, Child Q, March 2022. City of London and Hackney Safeguarding Children Partnership. Available at: https://chscp.org.uk/wp-content/uploads/2022/03/Child-Q- PUBLISHED-14-March-22.pdf Glaser, B., & Strauss, A. (1967) The Discovery of Grounded Theory: Strategies for Qualitative Research. Chicago: Aldine. PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 91 Hall, B., Khan, R., and Eslea, M. (2022) Criminalising Black Trauma: Grime and Drill Lyrics as a Form of Ethnographic Data to Understand “Gangs” and Serious Youth Violence. Genealogy 7: 2. https://doi.org/10.3390/genealogy7010002 Harding, S. (2020) County Lines: Exploitation and Drug Dealing among Urban Street Gangs. Bristol: Bristol University Press. Havard, T., Densley, J., Whittaker, A., and Wills, J. (2021) Street gangs and coercive control: The gendered exploitation of young women and girls in county lines. Criminology & Criminal Justice 2021: 1–17. HMICFRS (2021) Disproportionate use of police powers - A spotlight on stop and search and the use of force. Available at: https://www.justiceinspectorates.gov.uk/hmicfrs/wp- content/uploads/disproportionate-use-of-police-powers-spotlight-on-stop-search-and- use-of-force.pdf Innes, M. (2014) Signal Crimes - Social Reactions to Crime, Disorder, and Control. Oxford: Oxford University Press. IOPC (Independent Office for Police Complaints) (2022) National stop and search learning report – April 2022. Available at: https://www.policeconduct.gov.uk/sites/default/files/Documents/publications/OFFICIA L%20IOPC%20National%20stop%20and%20search%20learning%20report%2020% 20April%202022.pdf Irwin-Rogers, K., Muthoo, A. and Billingham, L. (2020) Youth Violence Commission Final Report. Youth Violence Commission. Available at: http://oro.open.ac.uk/72094/1/Youth%20Violence%20Commission%20Final%20Rep ort%20July%202020.pdf Lefevre, M., Preston, O., Hickle, K., Horan, R., Drew, H., Banerjee, R., Cane, T., Barrow, M., and Bowyer, S. (2020) Evaluation of the Implementation of a Contextual Safeguarding System in the London Borough of Hackney - Evaluation report. London: Department for Education. Mile End Institute (2022) Polling London - Londoners’ Priorities ahead of the Local Elections. Mile End Institute, Queen Mary University of London. Available at: https://www.qmul.ac.uk/mei/media/mei/tgc-media/filesx2fpublications/161_22- MILEEND_Polling-report_V5_final-WEB.pdf MOPAC (2022) A Problem Profile of Violence, Gangs and Young People. MOPAC Evidence and Insight, September 2022. Available at: https://www.london.gov.uk/sites/default/files/serious_youth_violence_problem_profile .pdf Muir, R., Higgins, A., Halkon, R., Walcott, S., & Jeffrey, B. (2022) A new mode of protection – re-designing policing and public safety for the 21st century. The final report of the strategic review of policing in England and Wales. The police PERCEPTIONS OF SERIOUS YOUTH VIOLENCE IN LONDON – FINAL REPORT UNIVERSITY OF WEST LONDON / ARCS LTD 92 foundation. Available at: https://www.Policingreview.Org.Uk/wp- content/uploads/srpew_final_report.Pdf Naseem, S. (2021) stop and search undermining confidence in policing by black communities. Available at: https://www.Policeconduct.Gov.Uk/news/our-podcasts- and-blogs/stop-and-search-undermining-confidence-policing-black-communities Norman G. (2010) Likert scales, levels of measurement and the "laws" of statistics. Advances in Health Sciences Education: Theory and Practice. 2010 Dec; 15(5):625- 32. doi: 10.1007/s10459-010-9222-y. Epub 2010 Feb 10. PMID: 20146096. Paoli, L. (ed.) (2014). The Oxford Handbook of Organized Crime. Oxford: Oxford University Press. http://site.ebrary.com/id/10958785. Public Health England (2019) A whole-system multi-agency approach to serious violence: A resource for local system leaders in England. London: Public Health England. Available at: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attach ment_data/file/862794/multi-agency_approach_to_serious_violence_prevention.pdf Ramshaw, N., Charleton, B. and Dawson, P. (2018) Youth Voice Survey 2018. MOPAC Evidence and Insight. Available at: https://www.london.gov.uk/sites/default/files/youth_voice_survey_report_2018_final. pdf Van der Kolk, B. (2014) The Body Keeps the Score – Mind, Brain and Body in the Transformation of Trauma. London: Penguin. Youth Violence Commission (2018) The Youth Violence Commission - Interim Report, July 2018. Available at: https://www.yvcommission.com/_files/ugd/ad2256_d4b4f677734a4a4b86cb5833cfce e53f.pdf ©RedQuadrant Benjamin Taylor 079 3131 7230 benjamin.taylor@redquadrant.com 1 Restart Interim Evaluation Drive Partnership Report from RedQuadrant September 2022 Claire Bethel Ellie Hutchinson Neil Reeder 2 Contents 1 Summary .............................................................................................................................. 4 Aims and objectives ...............................................................................................................4 Background ............................................................................................................................4 Context ..................................................................................................................................5 Methodology .........................................................................................................................8 Benefits of the programme ...................................................................................................8 Challenges for the pilot .......................................................................................................10 Conditions needed for programme to bed in: ....................................................................15 Quantitative analysis/value for money ...............................................................................15 Conclusion ...........................................................................................................................16 2 List of recommendations ................................................................................................... 17 3 Methodology ..................................................................................................................... 20 Acknowledgements ............................................................................................................... 21 Use of language in this report ............................................................................................... 21 4 Context .............................................................................................................................. 22 5 Background ........................................................................................................................ 24 6 Description of Restart ........................................................................................................ 26 Aims and objectives .............................................................................................................26 7 Findings and analysis ......................................................................................................... 31 Number of referrals .............................................................................................................31 Risk Level and abuse profile of service users and adult victim-survivors ...........................34 Accommodation pathway ...................................................................................................36 Safe and Together training ..................................................................................................36 Demographic information ...................................................................................................37 Benefits and challenges: ......................................................................................................40 8 Findings: benefits ............................................................................................................... 41 Holistic family approach tailored to individual needs .........................................................41 Positive culture change and practice in frontline staff .......................................................43 Improving safety and outcomes for families .......................................................................47 Advantages of the accommodation pathway .....................................................................51 Flexibility of the intervention ..............................................................................................51 9 Findings: challenges ........................................................................................................... 53 3 Set-up and mobilisation .......................................................................................................53 Culture and attitudes including training ..............................................................................54 Capacity of services to deliver Restart ................................................................................57 Referrals ...............................................................................................................................60 Engagement of families .......................................................................................................61 Complexity and clarity of purpose.......................................................................................63 Limitations of the accommodation pathway ......................................................................69 Funding uncertainty and sustainability ...............................................................................69 Interagency and partnership working .................................................................................71 Changing systems ................................................................................................................71 10 Accommodation pathway ................................................................................................. 74 Context ................................................................................................................................74 Take-up of the accommodation pathway ...........................................................................75 Challenges ............................................................................................................................79 11 Conditions needed for programme to bed in.................................................................... 85 12 Costs and benefits of the Restart programme .................................................................. 87 Costs of the Restart programme .........................................................................................87 Benefit of the Restart programme ......................................................................................89 Break-even point for Cost benefit ratio of the Restart programme ...................................92 13 Conclusion ......................................................................................................................... 93 14 Future work ....................................................................................................................... 98 Appendix 1: ........................................................................................................................... 99 List of interviewees and focus groups .................................................................................. 99 Appendix 2: Case studies .................................................................................................. 100 Glossary and list of abbreviations: ...................................................................................... 106 4 1 Summary 1.1 RedQuadrant was commissioned by the Drive Partnership in January 2022 to carry out an evaluation of the Restart pilot. Restart is a partnership-led multi-agency approach to keeping families safe at home through early engagement and intervention with those causing harm through domestic abuse. The project operates in five London boroughs (Camden, Croydon, Havering, Sutton and Westminster) through a partnership between The Mayor’s Office of Policing and Crime (MOPAC), Drive, Respect and the Domestic Abuse Housing Alliance (DAHA), with Cranstoun as the delivery partner. Funded by MOPAC and the Home Office, it was initially funded for a period of 12 months; this has been extended until March 2023. Aims and objectives 1.2 The project is an innovative attempt to achieve systemic change in the way that families experiencing domestic abuse are dealt with by local authorities. It aims to identify and respond to patterns of domestic abuse (of standard to medium risk) at an earlier stage for families engaged with Children’s Social Care, improving safety, housing and long-term outcomes for adult and child victim-survivors. Case managers (or Violence Prevention Practitioners) assess those who use abuse and initiate behaviour change whilst victim-survivors are offered support through Partner Support Workers. Safe and Together training, delivered by Respect, is an integral part of the programme, intended to improve the response of Children’s Social Care professionals to domestic abuse. This comprises of a four-day ‘core’ training programme as well as other types of training including a one-day overview training. The project delivers housing responses by offering alternative, initially short-term but possible long-term accommodation to the person using abuse as appropriate to ensure that the family is safe and able to stay in their home. Background 1.3 The project followed an earlier project, the Domestic Abuse Early Intervention and Accommodation Trial, set up during the Covid-19 national pandemic, when domestic abuse services were stretched due to increased demand, and victim- survivors found it difficult to seek help due to the lockdowns. At that time, hotel accommodation was readily available, and the participating councils were able to place perpetrators of domestic abuse in hotels to provide a short-term breathing space for families experiencing domestic abuse. This is no longer the case, and the context of this project is, therefore, significantly different, with the majority of the accommodation available for Restart service users now being sourced from the Private Rented Sector although short-term options still include hostels, bed & 5 breakfast accommodation and hotels. In setting-up Restart, several changes were made to reflect the evaluation of the earlier work. Context 1.4 Restart changes the way in which domestic abuse is dealt with so that the person who has used abuse is held accountable for the abuse and may be offered accommodation away from the family home. The intervention is victim-led and, should the victim-survivor wish to, they and the rest of the family may remain in their home where it is safe to do so; see diagram below: 1.5 The Restart pilot builds on an earlier trial which started during the Covid-19 pandemic; this was known as the Domestic Abuse Early Intervention and Accommodation Trial which ran from September 2020 to August 2021. The trial was delivered by Cranstoun alongside Children’s Social Care and Housing teams from the ten local authorities. The intervention offered temporary accommodation and pathways into longer-term accommodation if the survivor does not wish the perpetrator to return or if he is homeless. It also offers short- term behaviour change support for those identified by Children’s Social Care as being at risk of perpetrating domestic abuse, or where abusive behaviours have recently started. The primary outcome intended was behavioural change with 6 reductions in abuse and harm to enable a safe return to the family home or longer-term separation. 1.6 The Restart pilot started in August 2021 but did not start receiving referrals until October/November 2021. During the evaluation, we have reviewed data in relation to 76 referrals received between October 2021 and 1st July 2022. Service users who complete the short term one to one work with Restart are encouraged to then move to a longer-term Domestic Violence Perpetrator Programme (DVPP) – so far, 15 people have progressed to this stage. Cases have not been evenly distributed across the five councils, with Camden referring the most, and Havering the least, as shown in the table below (which uses Restart Client Data). Table 1: Number of Restart cases by local authority Cases (up to Mar 22) Cases (Apr 22 to 1 July) All cases % Camden 18 7 25 33% Croydon 4 15 19 25% Havering 6 1 7 9% Sutton 9 3 12 16% Westminster 7 6 13 17% Total 44 32 76 1.7 On the basis of Initial Needs assessment data, we calculate that 24% of the cohort of service users were identified as having an accommodation need (9 cases out of 38 completed assessments). 1.8 We note that 11 cases (18% of the 76 cases referred to Restart) have been referred to the accommodation pathway (compared to an anticipated referral rate of 25%). Two of these cases are still in progress. 1.9 The main current needs identified for service users (on the basis of 38 completed assessments out of the 76 referrals) were children, families and parenting (47%), followed by housing (24%), and mental health and psychological wellbeing (24%). 7 The level of risk of abuse identified for victim-survivors (of the 24 completed assessments and excluding ‘don’t know’) was 67% standard, 27% moderate, and 7% high-risk based on the Severity of Abuse Grid (SOAG). 1.10 Higher levels of risk were observed for service users (relating to 37 cases with completed assessments – note that data on SOAG ratings was reported to us separately from needs data), namely 38% standard, 38% moderate, and 24% high risk (proportions exclude ‘don’t know’ answers, which were especially high in relation to sexual abuse). The table below (based on assessments for service users) shows the abuse types and behaviours demonstrated by the service user. 1.1 256 Children’s Social Care Practitioners registered for the four-day and e-learning modules of Safe and Together CORE training, of whom 60 completed which is an overall completion rate of 23%; 63% partially completed the course or are currently in the process of completing the course. As practitioners continue their course, the overall completed rate may increase. The low rate of completion is 8 attributable to a number of factors including the pressures faced by social workers and the difficulties experienced by all staff during the Covid-19 pandemic. In addition, there is a one-day Overview training on Safe & Together. 379 registered to attended this, of whom 268 attended (attendance rate of 71%). Methodology 1.2 This evaluation took a mixed-method approach, which included: • desk-based research, • interviews and focus groups with a total of 25 practitioners and strategic leads, • a survey developed with input from Community Researchers with lived experience of domestic abuse, • observation of meetings and, • a quantitative analysis which comprised an indicative review of the impact of the scheme of risk to the victim-survivors, a corresponding illustration on the level of wellbeing gain and reduction in public expenditure required to address the issue, and indicative cost-benefit analysis summarising these effects. 1.3 Although 26 practitioners responded to the survey, only two service users and no victim-survivors returned the survey despite several requests to frontline practitioners to disseminate it. This report therefore does not reflect the voice of victim-survivors or the children and young people in families since it has not been feasible to obtain their views. Listening to their voice and reflecting their views will be a high priority in the next stage of the evaluation. Benefits of the programme 1.4 Benefits and potential benefits identified as a result of our evaluation have been found in the following areas: • Holistic family approach tailored to individual needs The intervention was endorsed by stakeholders as being relationship-based, holistic and family-centred. The comprehensive assessment carried out by the Case Managers forms a firm foundation for delivery of the behaviour change work with some support for the victim-survivor being provided by the Partner Support Worker and ongoing support from Children’s Social Care. The ability to offer one-to-one support (up to 12 weeks) rather than group work was welcomed, particularly but not solely for those for whom English is not their first language. 9 • Positive culture change and practice in frontline staff There was evidence of early signs of culture change being brought about within both Children’s Social Care and housing, with practitioners gaining in competence and confidence in their dealings with families experiencing domestic abuse and with those using abuse in particular. The Safe and Together core training and subsequent implementation activities underpin this success. Positive practice changes have been brought about and there is evidence of a less adversarial way of providing support to families emerging, with social workers building positive and trusting relations with families and social workers being able to offer support to families without having to impose conditions on them. The changes described in social work practice went beyond the cases referred to Restart, demonstrating an increase in the capacity and capability of staff to deal with those using abuse and to change the way they record this information. The project provides an opportunity for effective multi-agency working leading to improvements in the way that different parts of the local authority work both together and with other agencies. There was evidence of a team approach being developed in delivering the intervention. • Improving safety and outcomes for families There was little quantitative evidence of impact in terms of reduced risk for victim-survivors and families or long-term behaviour change because few people had completed the intervention and it is difficult to measure change between the start and finish of the intervention. We would not expect this to be easy to undertake, particularly as change may occur over a long period and this is a short intervention, but this will be essential to assess impact in a quantitative way. Adding greater numbers of completions will assist the process of increasing robustness, however this is a necessary but not sufficient condition for measuring change effectively. Interviews and case studies showed that Restart is thought to have the potential to bring about systemic change and improve both short-term and longer-term outcomes for families, with increased accountability of those perpetrating the abuse. The four-week period of the intervention is being used effectively to build a relationship with the service user and to encourage them to participate in a longer-term behavioural change programme. The successful completion of the four-week intervention with those who use harm and the onward referral of 15 to longer term DVPPs, were seen as early signs of success. In addition, 11 people were 10 considered unsuitable for group work and were instead referred for a 12-week intervention conducted on a one-to-one basis. • Advantages of the accommodation approach The accommodation pathway was seen and evidenced through case studies as having the potential to achieve more victim-centred services by minimising the disruption commonly experienced by those fleeing domestic abuse by providing options for those using harm. The change to the housing narrative was also found to be beneficial for families, providing much-needed respite at a time of crisis. Even in the 85% of cases where the accommodation pathway is not used but seen as an option, its availability is changing the perspective of housing officers and social workers beyond the referrals themselves. They are beginning to gain experience in its application and to work with colleagues in other teams, though it is too early to evaluate the impact. • Flexibility of the intervention The flexibility in implementing Restart was recognised and welcomed, particularly the ability to work on a one-to-one basis over a longer period for those who need it and are unable to participate in a group-based DVPP. Challenges for the pilot 1.5 A number of challenges have been encountered during the lifetime of the pilot in the following areas: • Set-up and mobilisation Restart made the transition from its predecessor programme during the Covid- 19 pandemic. This and other factors slowed down the setting-up of the intervention. Additionally, there was insufficient time built into the mobilisation period to allow for the complexity and challenges of implementing a project on this scale. Further time would have been helpful to establish the new relationships needed for Restart to flourish, to clarify roles and responsibilities and establish ways of working for frontline staff. • Culture and attitudes including training The challenges presented in achieving the cultural changes envisaged in implementing Restart were significant. Victim-blaming attitudes amongst local authority practitioners persist and interviewees encountered colleagues who believed that investing in perpetrators through Restart takes scarce resources 11 away from victim-survivors. Social workers’ lack of expertise and confidence in dealing with perpetrators themselves was also a challenge, particularly in boroughs with little previous experience of working directly with, or commissioning services for perpetrators which was seen by some as problematic. The strength of leadership and support provided by senior and middle management was seen a key determinant of how well the programme is being implemented locally: where this was not in evidence, frontline practitioners were more likely to be resistant to delivering the intervention. Although the Safe and Together training was found to be helpful and was strongly endorsed by those who had taken part, lower numbers attended than had registered. This was felt to be due to the pressure faced by social workers (see below) as well as the difficulties experienced by all staff during the Covid- 19 pandemic. • Capacity of services to deliver Restart Burnout and workforce retention issues were cited frequently as issues that limited the capacity of staff to take on new challenges, however worthy they consider them to be. High staff turnover has made it difficult to embed the new ways of working required. The pilot has taken place during a period of considerable and growing uncertainty with local authorities facing pressures including recovery from the pandemic, rising demand, fiscal pressures, poor morale and a high staff turnover, increasing complexity of cases and, recently, the fall-out from two serious case reviews which have had a national impact1. The Restart pilot is therefore being delivered during a difficult time with social workers coming under considerable pressure on a daily basis. Several interviewees raised the difficulty of social workers wanting to, and coming under pressure from management to close cases, even where there were ongoing problems in the family and the underlying issues had yet to be addressed. It was recognised that this can lead to a vicious cycle with the same families being referred repeatedly to Children’s Social Care (CSC) since the underlying reasons for the initial referral remained unresolved. It was also recognised that leaving cases open may be more resource-intensive, at least in the short term, since it is an acknowledgement that further work needs doing. Restart is not, however, intended to be a way of ‘offloading’ the case following referral. 1 See National Review into the murders of Arthur Lobinjo-Hughes and Star Hobson, The Child Safeguarding Practice Review Panel, 2022 12 • Referrals Whilst the number of referrals is seen as one indication of how the project is being implemented, there has been considerable variation in the numbers made to date with one local authority referring more than the others. Reasons identified for this include the complexity (or perceived complexity) and the time taken to make a referral, a lack of understanding about the intervention and what it can deliver, anxiety about engaging with perpetrators for the first time and a lack of commitment at senior and middle management level which impacts on staff at the social work ‘front door’. • Engagement of families The lack of engagement of either the person using or the person experiencing harm makes it difficult for the referral to proceed effectively. How they are approached and engaged by practitioners is key; skilling up staff to do so is covered in the training and the Restart manual (aimed at case managers primarily). However, as engaging with Restart is voluntary and based on the consent of the service user, there is no way of addressing the harm if they are unwilling to engage (equivalent to the ‘disrupt’ only pathway in interventions such as the Drive Project for perpetrators unwilling to engage)2 or if it is unsafe to attempt contact with the perpetrator. There is also a risk that perpetrators may appear to participate in the intervention who have no genuine intention of changing their behaviour. There were also some cases where services were unable to engage effectively with the victim-survivor which made it difficult to deliver a holistic and victim-led response. • Complexity and clarity of purpose The complexity of the intervention has deterred some practitioners from engaging with it fully despite efforts to provide clarification through written materials and work on implementation. Many practitioners did not understand fully the roles and responsibilities of members of the team and how they fit with locally commissioned programmes. There was also a lack of clarity about the aims and objectives of the intervention leading to poor commitment to the project. Presentation of Restart as being about early intervention when this is not always the case has caused some confusion. This may be due to the different ways in which ‘early intervention’ is interpreted – for some practitioners, it meant that families are supported when they first 2 Drive is an intervention that works with high-risk/serial perpetrators to prevent their abusive behaviour and protect victims, see http://driveproject.org.uk/ 13 experience domestic abuse, others saw it as being more about families experiencing domestic abuse at a lower level of severity or frequency. It was generally recognised that it was not the intention to achieve permanent behavioural change through the four-week intervention but to carry out a comprehensive assessment with initial challenge and motivation work with a view to securing participation in a longer-term programme and addressing some of the family’s underlying needs. • Limitations of the accommodation pathway The lower than expected take-up of the accommodation pathway (around 15% of referrals) is likely to reflect changes made since the original pilot which took place during the time of the Covid-19 lockdowns. Then there was a high demand for emergency accommodation and an over-supply of hotel rooms which made it easy to spot purchase hotel rooms and use them to accommodate those who were referred. Unlike the current pilot, longer term solutions were not available – the current pilot has direct links with the Local Authority housing teams who are able to offer routes to a broader offer which includes longer term accommodation, mainly in the private rented sector. Some boroughs have had low take-up of the accommodation pathway during the pilot so far though this is increasing. Attitudinal challenges were an issue since some frontline practitioners find the idea of providing those who use harm with accommodation uncomfortable, as it may be seen to reward abusive behaviour. It was also recognised that it would not be practicable to offer separate accommodation to the large number of standard to medium risk perpetrators of domestic abuse given its scarcity and high cost. Whilst one of the possible intended outcomes is that victim-survivors can remain in the family home, this may not be sufficient to allay her fear about remaining in the family home but may prevent her from seeking emergency accommodation. Some victim-survivors may choose to move home even if the perpetrator has left the home for a number of reasons; their wishes are the primary concern in reaching decisions on housing options. A further challenge is that the expectations of those offered accommodation through the project cannot always be met. There are several possible reasons for this including the lack of housing stock (particularly acute in many parts of London), the attitudes of housing staff which may make it difficult to prioritise perpetrators and their preference for self-contained social housing rather than private rented properties. Interviewees considered that while the main aims of Restart do not include addressing long-term housing needs, it was recognised 14 that the support provided including the detailed assessment can help to address such needs. The limited options available for service users is a constraint, leading to some turning down what is offered to them. There are a number of structural factors – mainly legal, policy and regulatory issues – surrounding the take-up of the accommodation pathway which have been encountered on the project. These include the difficulties for those with no recourse to public funds and the cap on housing benefit for perpetrators under 35 in private rented accommodation which makes it difficult to secure accommodation appropriate for those who wish to have ongoing contact with their children since this allows only entitlement to shared accommodation which may be unsuitable for children to stay in. There is no obvious solution to some of these but further consideration is needed to look for longer-term solutions. • Funding uncertainty and sustainability Short term funding of the pilot has caused uncertainly for those participating. The extension of the pilot was welcomed by all but the short-term nature of this due to constraints on funders made it difficult initially for managers and frontline practitioners to commit to it and this remains problematic. A pilot of this kind requires time to embed fully given the need to develop confidence and build capacity in so many staff. It is also difficult to assess impact during such a short timeframe when some of the objectives of Restart can only be achieved in the long term. • Interagency and partnership working Whilst Restart provides opportunities for multi-agency working and for bringing together a large number of partners to work together as a team, this relies on strong communication and a shared vision which has not always been evident. • Changing systems There was general recognition that Restart is about systems change rather than making minor changes to processes and the changes in place were seen as important, though challenging, first steps. There was general agreement that the changes needed would take a long time to achieve. Whilst different stakeholders had a different view of what success looked like, there was some consensus that this was as much about winning the hearts and minds of practitioners as achieving a high referral rate and strategic leads building a 15 case for the positive impact of the changes being piloted. There is a need for a comprehensive communications strategy to address these issues. Conditions needed for programme to bed in: 1.6 A number of conditions were identified which are essential for the intervention to be successful including strategic support, commitment and leadership, sufficient time for set-up and mobilisation and to build the necessary partnerships for Restart to flourish. Quantitative analysis/value for money 1.7 Firstly, it is important to highlight that it is not possible to fully conclude value for money at this early stage in the development of the new model, when business as usual delivery has not yet been established. More time is needed for data points that will inform a clearer picture for value for money to emerge. Therefore, all findings and messages here are caveated as emerging learnings rather than conclusions. Our interim assessment has focused on the service delivery aspect of the programme, rather than the provision of training. The reason for this is that, although the costs of the training can be calculated, it is difficult to measure the benefits accruing as a result of the training at this stage. 1.8 Out of a total budget of £615,000 or so during 2021-22, after allocating overheads we estimate that some £205,000 was spent on training, and £199,000 on service delivery, with a further £114,000 spent on preliminary work up to October 2021. With a caseload of 44 families referred between (late) October 2021 and 31st March 2022, and after excluding one-off start-up costs, this implies a cost per case of the order of £4,530. Assessment of the unit cost of training is to be undertaken in the second phase of the evaluation. Having reviewed data available from Home Office research on the wider social costs of domestic abuse (reduced wellbeing and productivity for victim-survivors, public sector costs for the NHS and criminal justice system), we further estimate that if the Restart programme reduces the wider social costs by a level of 9.8% or more, it has achieved a social value that is equal to its intervention costs. But if the intervention is compared against reductions purely in terms of reduced costs for public services, it would need to achieve an impact of the order of 94% which would be a considerable challenge. By contrast, a DVPP programme would require much lower intervention impacts, since their unit costs are lower (around 1/3rd as much, according to the Safe Lives report ‘A Safe Fund: costing domestic abuse provision for the whole family’).3 3 Safe Lives, A Safe Fund: costing domestic abuse provision for the whole family, see https://safelives.org.uk/node/1837 16 1.9 Our next step in relation to service delivery quantitative analysis will be to review the impact of the programme as a precursor to other programmes, as instigating (1) a higher proportion of service users to access those more sustained programmes and (2) a higher proportion of service users to sustain improvement in those programmes. Such an assessment cannot be made at this time, unfortunately, as the medium-term outcomes are not available to us. We expect more data to be available in the next phase of the evaluation. 1.10 Our next step in relation to the training aspects of the programme will be to review the measuring system for training benefits with Restart, and put in place any key metrics identified by that discussion. This will then enable us to provide an indicative quantified benefits assessment of that aspect. Conclusion 1.11 In conclusion, although it is still early days, there were many positive aspects identified about this complex, innovative and ambitious project. It has potential to bring about much-needed change to the systems that determine the services that families experiencing domestic abuse receive. There is some evidence that it is leading to a less adversarial, more victim-centred system with those who use abuse more likely to be held accountable for their behaviour and given the opportunity and support to change. 1.12 The accommodation pathway is a potential game-changer though there are many barriers to be tackled if it is to be made more generally available and to formulate a practical alternative to enable families to stay in their homes safely. We found some consensus that Restart has the potential to keep families safe and improve outcomes for victim-survivors and their children. Measuring impact and ensuring that the voice of the victim-survivors who have experienced the programme and their children should be a major objective of further work during the remaining nine months for which the programme is funded. 17 2 List of recommendations Recommendation 1 Time and capacity for setting-up Restart: We recommend that if Restart is scaled up and rolled out to other local authorities, sufficient time should be built in from the start to allow for proper preparation, the partnership to form and the vision to be developed collaboratively. Recommendation 2 Safe and Together training: We recommend that consideration should be given to ways in which the reach of the Safe and Together training can be extended to encourage a greater take-up and completion of the training by social work staff including senior social care staff. Frontline practitioners in other parts of the local authority such as housing and other agencies such as the police and NHS should be encouraged to attend the one-day overview training. Recommendation 3 Referral to Restart: We recommend that the central team should consider ways of encouraging referrals if these remain lower than expected, looking at the reasons identified and examining whether the referral process could be simplified. The wider dissemination of positive findings from the project and successful case studies might help to allay any concerns and publicise what can be achieved, supported by Restart champions. Recommendation 4 Suitability for referral: The next phase of the evaluation should try to establish through monitoring outcomes which families are most suitable for the intervention and who is most likely to benefit from referral. For example, it would be useful to know whether outcomes may be better in families wishing to stay together or in those who are separating. Recommendation 5 Support for children and young people: We recommend that consideration is given to providing specific support for children and young people as an integral part of the intervention, perhaps by having a dedicated practitioner as a resource for local authorities as there is for victim-survivors/partners. This could be done centrally as a resource for all the boroughs. Recommendation 6 Stronger monitoring of outcomes: various monitoring and evaluation tools already exist within the Safe and Together Training, the four- week early engagement and the in-borough DVPPs. However, there is a need to strengthen (and potentially link up) monitoring processes so that we gain a better picture and understanding as to how outcomes, particularly those which measure risk for families and behavioural change in perpetrators, are changing over time as a result of different intervention arrangements. 18 Recommendation 7 Communications: Where Restart is being set up, we recommend that there should be a clear communication plan in place before the project begins, with dedicated resources, and throughout the project so that staff are kept informed of progress. The accessible guidance already developed should be more widely disseminated and included in induction training for new social workers. Communications should clarify: • the aims and purpose of the intervention, the remit, roles and responsibilities of the key agencies and stakeholders; • that the intervention is not always ‘early intervention’ and that the intervention period can be extended beyond four weeks should the need arise, particularly where it is difficult to engage with the person who uses harm and/or the victim-survivor; • that a team approach is essential to deliver Restart; thus it will be unusual for a case to be closed once a referral has been made until the intervention is complete; • that Restart is intended to be one element of a suite of domestic abuse interventions for families which are available in local authorities, with clarity as to how these fit into local domestic abuse strategies. Communication materials may need adapting in each of the five boroughs to show how it dovetails with local systems. It may be helpful for the central team to work with each local authority (and any new ones if it is rolled out) to produce a Theory of Change to ensure that the aims and objectives are clear and realistic during the remainder of the pilot, and that there is a shared vision and understanding of what success looks like. Recommendation 8 Sustainability: We recommend that, if possible, the funding period for Restart should be extended to a total of three years to allow time for it to bed in, for practitioners to develop confidence in applying these new skills and changes in practice and to test its longer-term impact. Recommendation 9 Partnership arrangements: We recommend that where suitable, opportunities for intra-borough meetings are arranged locally so that practice can be shared at a local level. This will bring about better communication and shared learning between the partners and provide an opportunity to develop and deliver a shared vision. Recommendation 10 Changing systems: We recommend that any communications about Restart make it clear that it is aiming for long-term systemic change which 19 can only be achieved through commitment at all levels within a local authority and other agencies. Recommendation 11 Accommodation pathway: We recommend that: • the limitations of the pathway should be made clear from the outset to avoid raising expectations (or raising unrealistic expectations for the families). This should be set out clearly in literature on the project including training materials; • further work is carried out to assess some of the practical and policy issues involved in implementing the housing pathway such as (1) enabling those who use abuse to maintain contact with their children when safe to do so and (2) considering the options available for those who have no recourse to public funds who use the accommodation pathway but will have a more limited range of options available; • it is necessary to establish, build and maintain relationships at an early stage with the largest local providers of privately registered social housing since these providers may be housing the victim-survivors and perpetrators being supported and may make housing transitions easier; • the next stage of the project should examine some of the barriers and enablers for those using the housing pathway and what difference this makes to the family’s outcomes. The assumption that around 15% of families may use the housing pathway needs to be monitored carefully to see whether this is borne out since this has considerable implications for the future delivery of the project. Recommendation 12 Conditions for scaling up Restart: We recommend that the conditions above and others which the central team could readily identify are mapped and scored in each area which is considering implementing Restart to show the state of readiness for Restart. This could follow the model of a maturity matrix.4 Recommendation 13 Further evaluation: We recommend that a further evaluation is carried out to improve the quantitative and qualitative findings from this project and to inform future work in this area. 4 See, for example, an example of the Early Intervention Foundation’s maturity matrix https://www.eif.org.uk/resource/eif-maturity-matrix-speech-language-communication-early-years 20 3 Methodology 3.1 We employed a mixed-method approach to the evaluation. Our work strands comprised: • Desk-based research, • Interviews and focus groups with 25 practitioners and strategic leads; • A survey developed with input from Community Researchers with lived experience of domestic abuse, observation of meetings, and • Quantitative analysis. 3.2 Desk-based research was conducted following the award of the contract, and a range of documents analysed and reviewed relating to the pilot and the background. 3.3 We then recruited seven victim-survivors who were not involved with this pilot who were identified by the partner organisations as Community Researchers. With their help, we developed a questionnaire to be used as the basis for a survey of adult victim-survivors and for services. This survey was issued through Cranstoun to Partner Support Workers and Case Managers to distribute as well as through the Restart Programme Lead. 21 people from services and 2 perpetrators responded to the survey. 3.4 We then hosted a number of interviews and focus groups with practitioners, trainers and strategic leads to explore their perspectives on what was working, what could be improved and any learning they wished to share. We spoke with 21 people through the focus groups, two people in a shared interview, interviewed four strategic leads and two individual practitioners. We were unable to speak with strategic leads from two of the boroughs, however, these boroughs were represented in the survey. 3.5 We also observed four meetings with practitioners and service leads, which provided an insight into partnership working and shared approaches to working with perpetrators in particular. Anonymised quotations from interviewees and focus group attendees have been included with their permission throughout this report. 3.6 We produced an Emerging Findings report in May 22’ summarising our preliminary findings and recommendations which were used as the basis for a discussion with the Restart team and the funders, three months after this evaluation started. Feedback was received and considered during the remainder of the evaluation period. 21 3.7 Our quantitative analysis had two phases: • An initial phase using comparator study adapted for the Restart cohort – reflecting impact and programme costs; • The second phase drew on data on actual outcomes in terms of the risk of continuing to present harm to their family, as well as actual data on unit costs (see below). 3.8 In terms of assessing value for money, our general approach has been to: • Calculate unit costs, defined as expenditure (2021/22) divided by the number of clients (2021/22), taking into account (a) the share of resource between training provision and work with clients, and (b) noting that much of the overhead costs would not be incurred in dissemination of the programme; • Assess the current social costs in terms of wellbeing loss and cost to public services that result from a reduction in domestic abuse; • Compare the benefits of the programme against the current social costs, as a way of showing the relative improvement required for social benefits to match the cost of the programme. 3.9 A list of those interviewed for this report is set out in Appendix 1. Acknowledgements 3.10 We would like to thank all those who took part in the interviews and focus groups we carried out for this project and to those who provided the information needed to complete this report. 3.11 We would particularly like to thank the seven victim-survivors we worked with who co-developed the survey questions. Thank you for sharing your time, insights and stories with us. Use of language in this report 3.12 In this report, we have tried to use gender-neutral language where possible since domestic abuse can happen to anyone. However, whilst perpetrators may be of any gender, we recognise the gender-based nature of domestic abuse and the fact that the majority of victim-survivors are female and the majority of those who use abuse are male. We have therefore used ‘he/his’ for the most part in discussing those who use harm and ‘she/her’ to refer to victim-survivors throughout our report though this should not be taken to suggest that the intervention works only with male perpetrators and female victim-survivors. The terms ‘service user’ and ‘perpetrator’ are both used interchangeable throughout this report. 22 4 Context 4.1 The pilot takes place during a time of considerable change for domestic abuse policy and practice generally and approaches to perpetrator interventions specifically. Domestic abuse as a major component of violence against women and girls is a high priority at national level and in London. This is reflected in the recently-published Tackling Domestic Abuse Plan5, the Violence Against Women and Girls National Statement of Expectations6 and the Mayor’s Violence Against Women and Girls Strategy7 and the Mayor’s recent campaign which has focused on the role that men play in ending violence8. Recent legislative changes9 have also been made which place a new legal duty on councils to fund support for survivors in safe accommodation, and a guarantee that all survivors will be in priority need for housing, keeping a secure tenancy in social housing if they need to escape an abuser. Children are also recognised as victims of domestic abuse in their own right for the first time. 4.2 These developments follow calls for a national domestic abuse perpetrator strategy for England and Wales10, including a call for additional investment in quality assured perpetrator programmes that address the whole range of risk from perpetrators. 4.3 The Restart pilot and its predecessor programmes have taken place during the various phases of the Covid-19 pandemic and indeed this was partly what gave rise to the pilot. The pandemic is recognised to have had a major impact on domestic abuse: in particular, there was an increase in the level of offences flagged as domestic abuse during the Covid-19 pandemic and an increase in demand for 5 HM Government, Tackling Government Domestic Abuse Plan, March 2022. See https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1064432/E 02735263_Tackling_Domestic_Abuse_CP_639_Elay__002_.pdf 6 Violence Against Women and Girls National Statement of Expectations: guidance on commissioning services to support victims and survivors of violence against women and girls, Home Office, March 2022. This sets out the need for a clear focus on perpetrators in order to keep victims and survivors safe and requires local authorities to carry out needs assessments and ensure there are services in place which manage the risk posed by perpetrators.https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/fil e/1064571/National_Statement_of_Expectations_2022_Final.pdf 7 See https://www.london.gov.uk/publications/tackling-violence-against-women-and-girls 8 See https://www.london.gov.uk/press-releases/mayoral/mayor-campaign-to-help-end-violence-against- women 9 https://www.womensaid.org.uk/what-we-do/campaigning-and-influencing/campaign-with-us/domestic-abuse- bill/; Domestic Abuse Act 2021; https://www.legislation.gov.uk/ukpga/2021/17/contents/enacted 10 A Domestic Abuse Perpetrator Strategy for England and Wales, Call to Action, see http://driveproject.org.uk/wp-content/uploads/2020/01/Call-to-Action-Final.pdf; January 2020. 23 domestic abuse services during the Covid-19 lockdowns11, as well as providing an opportunity for housing innovations as a result of the reduced demand for hotel accommodation by paying guests. These empty rooms could then be used to accommodate perpetrators of domestic abuse. However, by the time the Restart pilot was launched, this was no longer the case and hotels had reopened. 4.4 Other factors that may impact on demand for domestic abuse services are the current cost of living crisis which will place greater financial pressures on families and the ongoing conflict in Ukraine which may place more pressure on housing services. 11 See Domestic abuse during the Coronavirus (Covid-19) pandemic, England and Wales: November 2020 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/domesticabuseduringthecor onaviruscovid19pandemicenglandandwales/november2020#main-points 24 5 Background 5.1 RedQuadrant was commissioned in January 2022 by the Drive Partnership central team to undertake an evaluation of the Restart pilot. The Restart pilot is a partnership multi-agency approach to keeping families safe at home through earlier intervention with those causing harm. One of the issues experienced by families experiencing domestic abuse is the need for separate housing, frequently at a time of crisis. Traditionally, there has been a ’separate and isolate’ approach whereby victim-survivors are either offered a place of safety at a refuge, found alternative housing away from the perpetrator or adaptations made to their home through ‘sanctuary schemes’ to keep them safe through security measures. 5.2 Where such separation is not feasible or is not acceptable to the family, the children may be removed into the care of the local authority. The adverse effects of this approach are well-known with the victim-survivor and her children frequently being subjected to the upheaval of having to move to a different area (with the location unknown to the perpetrator), leaving family and friends, and sometimes necessitating a change of school. Restart is a radical attempt to introduce a systemic change whereby the person who has used abuse is held accountable for the abuse and can be offered accommodation away from the family home, allowing the rest of the family to remain in their home, with support being offered in parallel to the victim-survivor. The approach is victim-survivor led, allowing the victim-survivor and the children to stay in the family home if they wish and where it is safe to do so. 5.3 The Restart pilot builds on an earlier trial which started during the Covid-19 pandemic; this was known as the Domestic Abuse Early Intervention and Accommodation Trial which ran from September 2020 to August 2021. The trial was delivered by Cranstoun alongside Children’s Social Care and Housing teams from the ten local authorities. The intervention offered temporary accommodation and intensive behaviour change support for those identified by Children’s Social Care as being at risk. The primary outcome intended was behavioural change with reductions in abuse and harm to enable a safe return to the family home or longer- term separation. 5.4 The initial six months of this trial was evaluated by the Social Care Institute for Excellence (SCIE). At that stage, 27 cases had been referred into the programme, and seven had successfully completed the programme. They found that the trial had successfully reduced abuse, that removing the perpetrator from the family led to a reduction in abuse and that this also provided emotional space for them to engage in the intervention. They also found that this initial programme acted as a springboard for perpetrators to engage in longer term behaviour change 25 programmes. Partner support running alongside this was deemed very important and that social workers’ involvement with the pilot had changed their practice and approach to domestic abuse. They also found that support delivered via the telephone was less successful than other forms of engagement. 5.5 Eight recommendations were made including needing a greater flexibility in the accommodation offer, the need for smooth transitions into longer behaviour change programmes, pathways for victim-survivor support, training for social workers, the importance of a strong identity for the programme, reflections on working with services users who don’t have English as a first language, linking the intervention to other programmes around substance misuse, and the need for evidence gathering with children, young people and victim-survivors. 5.6 Recommendations that were implemented include the introduction of Partner Support Workers, smooth transitions into longer behaviour change programmes (such as the Men and Masculinities programme12), the development of a distinct brand, reflections on working with those where English is not the first language, and training for social workers. 5.7 A twelve-month evaluation was carried out by SafeLives, building on the findings and recommendations of the SCIE evaluation. This found that the intervention continued to reduce risk around coercive control and physical abuse (but was less clear around the incidence of sexual abuse) and that it increased victim-survivors’ understanding of abuse. It was found to have a positive impact on social workers’ attitudes and practice, and that the housing pathway had some impact but could be used more flexibly, with an awareness that there was potential for the housing offer to be used to perpetrate abuse. The report also noted that what worked less well was communication, the impact of the four-week intervention and working with those with additional support needs. The recommendations made included recommendations for project delivery, as well as reflections for the sector. This included the need to improve communications, adapting intervention materials, developing a clear identity, developing a whole system, multi-agency approach, and for multi-year funding to be released to support complex interventions. Recommendations that were then taken forward were: the development of a clear identity, improved communications, the adaption of intervention materials and a focus on a multi-agency approach. 12 The Men & Masculinities Programme is intended to provide a safe space for men who have engaged in abusive, harmful and damaging behaviour within their relationships. Run by Cranstoun, it is a 24-week programme which focuses on behaviour change and is Respect accredited, see https://cranstoun.org/help-and- advice/domestic-abuse/men-and-masculinities/ 26 6 Description of Restart 6.1 Restart is a pilot project delivered through a partnership between the Mayor’s Office of Policing and Crime (MOPAC), the Drive Partnership, Respect, the Domestic Abuse Housing Alliance (DAHA) with Cranstoun as the delivery partner (they were appointed in August 2021). The project started in August 2021 though the first cases were not received until October/ November 2021. It was initially funded for a period of 12 months, but later extended until March 2023. It operates in five boroughs across London: Camden, Croydon, Havering, Sutton and Westminster. The project has been funded by MOPAC and the Home Office. Aims and objectives 6.2 Restart aims to ‘identify and respond to patterns of domestic abuse at an earlier stage for families engaged with Children’s Social Care (CSC), improving safety, housing, and long-term outcomes for adult and child victim-survivors’. It is intended to achieve systemic change by bringing together child protection, domestic abuse, community safety and housing services with a view to: ▪ Improving safety ▪ Increasing housing stability ▪ Improving long-term outcomes for victim-survivors. 6.3 The specific aims of the pilot are to: ▪ Identify and respond to patterns of domestic abuse at an earlier stage for families, improving safety and long-term outcomes for child and adult victim-survivors ▪ Build motivation and facilitate access to behaviour change interventions for the abusive parent ▪ Facilitate access to alternative accommodation for the abusive parent, alongside specialist support for the whole family ▪ Reduce repeat and serial incidents of domestic abuse in a survivor-centred approach, alongside specialist support for the whole family ▪ Prevent escalation of CSC intervention and the risk of adult and child victim-survivors needing to move/flee to safer accommodation ▪ Provide training and capacity building for CSC practitioners to enable CSC to effectively hold abuse parents to account ▪ Increase safety and housing stability of adult and child victim-survivors by holding perpetrators to account 27 ▪ Shift thinking/approaches in relation to preventing family homelessness by working with housing teams on accommodation pathways. 6.4 Children’s Social Care workers are encouraged to refer families to Restart when they have identified emerging patterns or behaviours of domestic abuse. Safe and Together training is provided as an integral part of the Restart approach and practitioners are encouraged to attend. Safe and Together is a US-derived model for tackling domestic abuse, it is licensed for use in the UK13 and has been extensively rolled out in some parts of the country such as Edinburgh. The training is provided by the Safe and Together Implementation team who undergo training, certification and coaching and are co-located within the local authorities to support practice development. They carry out case consultations with social care practitioners in relation to individual cases, complete case audits and provide briefings to spread awareness of the programme. 6.5 Case Managers and Partner Support Workers are provided by the delivery partner, Cranstoun, they work with perpetrators and partners, or ex-partners, over an initial four-week period to assess family needs. This includes, but is not limited to, housing. DAHA have a strategic role in providing oversight of the accommodation aspects of the project. Their lead (based in Standing Together who co-ordinate the Alliance) links with all the Strategic Housing leads in the five boroughs to review learnings, they also meet regularly with the Accommodation Support Worker based in Cranstoun to provide guidance and help to troubleshoot any problems which arise 6.6 A Practitioners Intervention Manual is available (currently under revision), aimed at Case Managers and Partner Support Workers in Cranstoun, to provide practitioners with comprehensive guidance on the assessment process, working with people who use harm, behaviour change interventions and accommodation pathways. 13 The three main principles of Safe & Together are (1) keeping child safe and together with non-offending parent (2) partnering with non-offending parent as default position (3) intervening with perpetrator to reduce risk and harm to child. It asks questions such as: What if child protection systems could spend less time and money removing children form the adult survivors and more time partnering with survivors to keep children safely in their own homes and communities? What if domestic violence survivors knew that professionals and systems would not blame them for what the perpetrator is doing? See https://safetogetherin.wpengine.com/the-sti-model/model-overview/# 28 6.7 The diagram below outlines the different aspects of the Restart programme and how they interlink. 6.8 Within the project, the key roles and the agencies they are based in are as follows: • Restart Programme Lead, Drive Partnership has oversight of the programme, implementation, roll-out, and monitoring and evaluation tools. • Data Analyst, Drive Partnership informs model development and performance management via monitoring and evaluation. • Practice Adviser, Drive Partnership works with the Cranstoun team and borough housing and CSC Leads to improve best practice across the Restart programme. • Service Manager, Cranstoun has an overview of the project’s case work delivery from case management to partner support workers. • Partner Support Workers, Cranstoun work with victim-survivors to help support awareness of domestic abuse and systems and support agencies who can help. 29 • Case Managers, Cranstoun work with perpetrators in the four-week timeframe to establish relationships, undertake assessments and begin work to initiate behaviour change. • The Accommodation Practitioner, Cranstoun works with perpetrators to explore the housing options that are available to them and coordinates the Restart Housing Panels. • Safe and Together Project Manager, Respect is responsible for co-developing training manual and resources, and for leading the team of implementation leads. • Training and Resources Manager, Respect works with the Safe & Together Implementation Leads and Project Manager to co-ordinate training events for general workforce development perpetrator training, and Safe and Together specific training, respectively. • Safe and Together14 implementation leads, Respect work with social work teams to embed the Safe and Together approach in responding to domestic abuse. There are three Safe and Together implementation leads who are co- located within the local authorities. • Strategic Housing Lead, DAHA provides strategic guidance and oversight on the accommodation support model and housing pathways, and works to build learning from the pilot to improve its immediate delivery and to identify wider systems change work to be done at a practice and policy level. 6.9 More detail is set out in the internal Restart: Early Intervention and Accommodation Project, Practitioners Intervention Manual. 14 Safe and Together is a training package for social workers aiming to shift perspectives on safety and blame in relation to mothering within domestic abuse. This programme was developed in the US and has been used around the world to promote a child-centred, trauma-informed approach to social work responses to domestic abuse. 30 6.10 The following diagram illustrates the relationship between the partners and their roles and responsibilities: 6.11 An overview of the Restart model is shown in the diagram below: 31 7 Findings and analysis 7.1 The pilot started at the beginning of August 2021, and has been running for nearly a year, however, did not start taking on clients until late October/November. Whilst it builds on its predecessor trial, it is still at an early stage, partly because it took longer than anticipated to set it up and to become fully operational. Much of the information we derived came from interviews and focus groups – whilst the intention had been to obtain views from victim-survivors, people who use harm (service users) and practitioners through a survey, there were far fewer responses than we had hoped for, particularly from victim-survivors and service users. We received responses from twenty-six practitioners, two service users and no victim- survivors. In addition, we had hoped to gather some responses from children and young people whose families were receiving the intervention through trusted adults such as children’s social workers or youth workers but this did not prove feasible. We were also unable to secure interviews with strategic leads from all local authorities. 7.2 In addition to the information discussed above, we have also drawn on data provided to us by the Drive Partnership. The data was originally collected by the Cranstoun team via information provided within the referral from CSC, as well as information from the Case Managers and PSW assessments which were recorded in the Case Management System (CMS). Anonymised data was then downloaded by Cranstoun and provided to the Drive Partnership and in turn to RedQuadrant for the purposes of this evaluation. Number of referrals 7.3 Table 2 below shows the number of referrals by borough. In total, there have been 76 referrals to the project until 1st July 22. 7.4 Of these referrals, the data available to us indicates that: 3 were deemed unsuitable for the Restart programme, and a further 17 either dropped out or did not engage and a further 2 had cases closed for reasons not given. 18 referrals were made to other services, including DVPPs, and a further 20 cases are at various stages of being assessed. As of 1st July 2022, 12 were receiving support, with a further 4 having been discharged from a course. 7.5 The table below shows distribution of referrals by borough. 32 Table 2: number of referrals by borough Cases (up to Mar 2022) Cases (Apr 2022 to 1 July) All cases % of all cases as total number of cases Camden 18 7 25 33% Croydon 4 15 19 25% Havering 6 1 7 9% Sutton 9 3 12 16% Westminster 7 6 13 17% Total 44 32 76 7.6 Cases have not been evenly distributed across the five boroughs, with Camden referring the most, and Havering the least. There has been a marked increase in cases in Croydon. 7.7 The following information is based on the 38 service users on whom needs information was recorded. This relates to two service users in the Assessment phase, 12 whose cases are closed, 12 where service users dropped out, and 12 currently receiving support. The data recorded during the assessment on service users includes the number of Adverse Childhood Experiences (ACEs)15, which are summarised below. 15 Whilst the use of ACEs has been criticised as a tool for understanding childhood trauma, it is used as part of the reporting dashboard and is therefore used here. For more information see https://blogs.ed.ac.uk/CRFRresilience/2019/05/15/aces-a-social-justice-perspective/; see also Adverse Childhood Experiences: what we know, what we don’t know and what would happen next, Early Intervention Foundation (February 2020) https://www.justice-ni.gov.uk/sites/default/files/publications/justice/eif-adverse- childhood-experiences-report-2020.pdf 33 Table 3: Adverse Childhood Experiences (ACEs) among service users Experience of ACEs Number (out of 38) % Parental Separation 10 26% Verbal Abuse 4 11% Domestic Abuse 4 11% Direct Physical Abuse 3 8% Drug Abuse 2 5% Alcohol Abuse 1 3% Mental Illness 1 3% 7.8 Assessments are also completed in relation to the current needs of service users, and these are summarised below. Table 4: Current needs of service users Need Number (out of 38) % Children, family or parenting 18 47% Housing 9 24% Mental health and psychological wellbeing 9 24% Work, training or education 6 16% Substance misuse or other addiction 5 13% Social and community support 5 13% Finance & debts 4 11% Purposeful activity 3 8% Physical health 2 5% Immigration needs 1 3% 34 7.9 This shows housing to be as high a need as mental health and psychological wellbeing, with children, family or parenting being the highest need identified. It should be recognised that these needs (for example, housing, finance and debts) can change during the period of engagement. Risk Level and abuse profile of service users and adult victim-survivors 7.10 We analysed the data to look at the level of risk identified by Case Managers for service users and by Partner Support Workers for victim-survivors. The following table shows the risk level as identified for 37 service users (excluding one ‘don’t knows’) by the Case Managers. Of these 37 service users, 3 are in the assessment phase, 6 have been referred to other services, 10 dropped out, 12 are being given support, 4 have been discharged from their course and 2 have received closure for other reasons. Table 5: Service user risk level Level of risk Number of service users Proportion Standard 14 38% Moderate 14 38% High risk 9 24% Total 37 100% 7.11 Of the above proportions, the level of service users regarded as ‘high risk’ (24%, some 9 out of 37) is perhaps surprising for the cohort, given that Restart is not intended for such cases. Otherwise, the data indicates an even balance between moderate and standard levels of risk. 7.12 The graph below shows the proportions of abuse profile split out by risk level. 35 7.13 Turning to victim-survivors, this shows a different pattern. We analysed 15 of the cases associated with the Restart service users who have had an assessment completed. These cases are derived from information recorded by the PSW which includes an initial DASH16 assessment and is focused on the support available prior to the assessments – there is therefore a lower number for SOAG assessments compared to the number of service users. Table 6: Adult victim-survivors’ risk level (derived from assessment by PSW) Risk level Number of victim-survivors Proportion excluding ‘'Don't know’ Proportion including 'Don’t know’ Don’t know 9 38% Standard 10 67% 42% Moderate 4 27% 17% High risk 1 7% 4% Total 24 7.14 The graph below shows the profile of abuse experienced by victim-survivors. Though at first sight it appears to suggest a perception of lower risk among victim- survivors, there are many reasons for not identifying or disclosing risk in the same way and so we would advise great caution in drawing conclusions from this data. 16 DASH stands for domestic abuse, stalking and ‘honour’-based violence. The DASH score is widely used to help identify, assess and manage risk in domestic abuse. See https://www.dashriskchecklist.co.uk/ 36 Accommodation pathway 7.15 So far, a total of 11 referrals have been referred to the housing pathway, of which two were undergoing assessment as of end-July 2022. The following shows the breakdown of these referrals by borough (it should be noted that the two cases under assessment as of end July 2022 may not actually use the housing pathway): Table 7: referrals to the accommodation pathway Camden 4 Havering 1 Sutton 4 Croydon 1 Westminster 1 Total 11 7.16 Further analysis of the accommodation pathway is set out in section 10. Safe and Together training 7.17 Safe and Together training is an integral part of the Restart approach and is being rolled out across the five local authorities, overseen by Safe and Together Implementation Leads (see paragraph 6.8). 256 practitioners have registered for the Safe and Together core training, though the completion rate for the four-day core training (which until now has comprised of a number of online e-learning modules) is low at 23%. However, in addition a further 63% partially completed the course or are currently in the process of completion (in addition to which, 14% did not attend). As practitioners continue their course, the overall completed rate may 37 increase. There are many reasons for the lower rate of completion including the pressures faced by social workers and the difficulties experienced by all staff during the Covid-19 pandemic. Table 8: Safe and Together Core training completion rates Did not attend (no.) Partial (no.) Complete (no.) Total Did not attend (%) Partial (%) Complete (%) Camden 5 30 8 43 12% 70% 19% Croydon 5 32 11 48 10% 67% 23% Havering 5 28 5 38 13% 74% 13% Sutton 14 42 16 72 19% 58% 22% Westminster 6 29 20 55 11% 53% 36% Total 35 161 60 256 14% 63% 23% 7.18 There is also a one-day Overview training on Safe & Together. 379 registered to attended this, of whom 268 attended (attendance rate of 71%) across the five sites. 7.19 In addition, there has also been other training provided. A range of bite sized trainings were held to address knowledge and training gaps identified including five webinar sessions and three rounds of half day works shops. 580 practitioners across all five sites attended this training. Webinars were delivered on: • Responding to counter allegations/ Respect Toolkit • Working with perpetrators • Confidence in discussing perpetrator programmes with families • Unconscious biases and working with perpetrators • The invisible perpetrator and busting myths around information sharing Additionally, three rounds of half-day workshops on: • Working with perpetrators • Responding to counter allegations/ Respect Toolkit Demographic information 7.20 Demographic information is recorded on each service user, the following data is based on the 76 referrals. Key aspects include: 38 • 97% are male, 3% are female (the question was ‘Not set’ in 8 cases, and these have been excluded from the above) • 99% are heterosexual, and 1% is identified as ‘don’t know’ (either the Case Manager did not know or the service user was unsure) (the question was ‘Not set’ in 8 cases, and these have been excluded from the above) • The nationalities with the highest number of service users are shown below: Table 9: Highest attending nationalities Nationality Number of service users English 27 Polish 4 Sri Lankan 3 Jamaican 3 Spanish 2 Portuguese 2 Pakistani 2 Irish 2 Bangladeshi 2 Afghan 2 • These figures imply that the nationality of the service user was English in at least 36% of cases (out of 76 participants, nationality data was not obtained in 11 cases), with the proportion being 42% if missing data is excluded. • The six main ethnic groups are shown below. 56% of service users were White, 18% Black, 13% Asian, and 12% Mixed or any other ethnic background. This indicates that the programme is meeting the needs of a large number of ethnic groups – the proportion is approximately in line with the overall rate for London of 40% from an ethnic minority17. 17 https://www.trustforlondon.org.uk/data/geography-population/ accessed 24th August 2022 39 Table 10: Top 5 ethnicities of service users Number % White: British 22 29% White: Other 16 21% Not Stated/Other 13 18% Black / Black British: Caribbean 8 11% Asian / Asian British 4 5% Total 76 7.21 The table below shows the age of service users assessed at the point of cases opening (note that two service users did not provide data on their date of birth, and so are excluded). Table 11: Age of service users Age range Number % 16 to 19 years 2 3% 20 to 24 years 3 4% 25 to 34 years 24 32% 35 to 44 years 28 38% 45 to 54 years 15 20% 55 to 59 years 1 1% 60 to 74 years 1 1% Total 74 100% 40 Benefits and challenges: 7.22 We have identified several actual or at least potential benefits and some key challenges from our qualitative and quantitative analysis. Further work is planned to measure the extent to which these impact on outcomes for families affected by domestic abuse and to elicit the views of participants since their voice is lacking from this report (see section 13). 41 8 Findings: benefits 8.1 We have identified several possible benefits of Restart through our discussions with interviewees and through case studies. Some of these are benefits that have been realised during the project, others were pointed out as potential benefits which might be realised once some of the initial problems have been overcome and the project has had longer to bed in. Holistic family approach tailored to individual needs 8.2 The Restart approach is intended to be relationship-based, holistic and family- centred in its approach; this was touched on during many of the interviews with interviewees seeing this positively and feeling that this has to some extent been achieved. The assessment carried out by Case Managers is seen as comprehensive and is intended to address the needs of all members of the family including those who use abuse and, through information obtained from the PSW and CSC, the victim-survivor and children as well (guidance on the assessment of those who use abuse is clearly set out in the manual). One practitioner told us: ‘It’s much more on the radar, this idea of, I guess, having a more assertive approach to working with the kind of whole family so…sort of proactively actually thinking about and talking about work with each member of the family, rather than just thinking of these cases that kind of keep coming around.’ Practitioner, focus group. 8.3 Restart was seen by the majority of interviewees as a positive example of a people- led, family-centred programme. The fluidity and adaptability of it around the needs of individuals was seen by many as a strength by both practitioners and strategic leads: ‘I think there’s been a lot of fluidity in the project and it’s organically developed and going in the direction that it’s meant to go in. And the fact that it’s allowed for that flexibility has been a huge benefit. You know, it hasn’t been rigid... So, the fluidity that we’ve been allowed, and the organic development, like I said, has benefited project hugely as a pilot. It’s a pilot, and I understand that with pilots, anything can happen. Things can change…and we have to allow for that.’ Delivery partner, focus group. 8.4 It is hoped that the Restart approach will help to address the cycle of abuse though this is difficult to prove given that it would require a much longer-term cohort study to follow-up people who use harm in subsequent relationships. Several interviewees referred to the fact that, under the current system, there is 42 frequently a cycle of abuse with the same families referred repeatedly to CSC since the underlying reasons for the initial referral remained unresolved. They hoped that Restart would lead to this cycle being broken resulting in fewer serial perpetrators. [normal practice is where] ‘Mum’s going to separate from abusive perpetrator, we’re happy we safeguarded this one family, but now he’s going to start dating the woman around the corner. So, we ain’t actually solved nothing, we’ve just moved the problem on and we’re waiting for another victim to come up basically, because we’ve not done any positive intervention or change at all.’ Frontline practitioner, children’s social care, focus group. 8.5 We found that there were higher rates of referrals in boroughs where strategic leads were committed to people-led services, openness to innovation, and acknowledging families’ ‘fears of ‘state abuse’, with commitment to changing those systems18. One strategic lead told us: [we’ve been] ‘hearing very clearly back from particularly women who are both victims of domestic abuse and also feel... like the victims of state abuse because in fact they’ve been put through the system we have to protect children, which is experienced as coercive’. Strategic lead. Whilst not an impact of Restart, this people-led, open approach to learning from public sector failings in terms of outcomes for families engaged in social work, has enabled Restart to bed in more successfully in this borough. 8.6 The Restart approach was also felt to be more likely to be adaptable to the needs of service users than other programmes. Several practitioners reflected on the benefits of being able to offer one-to-one support to the perpetrator rather than only group work. This was felt particularly important for a number of reasons: establishing relationships with services users, and engagement with minority ethnic families or families who have minimal English. One focus group attendee noted: ‘those with language difficulties, ...cannot go to groups [it’s] just not viable. So that has...made a huge difference to the successes of the project.’ Delivery partner, focus group. 18 For example, the London Borough of Camden has just been awarded an Excellent rating by Ofsted for its children’s services. https://news.camden.gov.uk/camdens-childrens-services-among-countrys-best-after- receiving-outstanding-rating-by-ofsted/#:~:text=rating%20by%20Ofsted- ,Camden's%20children's%20services%20among%20country's,receiving%20'outstanding'%20rating%20by%20Ofs ted&text=Camden%20Council's%20children's%20services%20have,best%20performers%20in%20the%20country 43 8.7 The advantage of the Restart approach is, then, that it meets the needs of both the victim-survivor and the person using harm. The victim-survivor receives ongoing support both through CSC and through the Partner Support Worker while the person using harm will work with the Case Manager and hopefully then be referred to a longer term DVPP. Positive culture change and practice in frontline staff 8.8 Several interviewees told us that there had been an improvement in the competence and confidence of frontline staff in dealing with domestic abuse in families generally and in those who use abuse specifically. This seems to be due to the joint working with the project staff and the Safe and Together training as well as input from the Implementation Leads, which includes case consultations and auditing of cases; for example: ‘I’ve been able to go in and sort of establish and audit cases and sort of find where people’s needs are, look to meet those and then also reintroduce the intervention project as something that can be offered as a solution for some of those problems.’ Implementation Lead, focus group. ‘I think a real success has been the confidence level of practitioners, this is from discussions and from the management feedback that they’re seeing that practitioners are more able to go have conversations with perpetrators, where they weren’t, and more confident to have conversation about harms’. Implementation Lead, focus group. 8.9 This was also clear from the practitioner survey results: ‘Really helpful to have perp resources and training, I feel like this has been really missing in my career. I’ve been on loads of trainings around V/S & safeguarding but the approach to perps feels like a real gap.’ ‘Since undertaking the training I have not had an opportunity to work directly with survivors, victims or perpetrators; however, I believe that the training has given me an opportunity to think about how I would be responding however not necessarily put it into practice.’ ‘I am able to change my thinking, the programme is rolling out and, still new to [Borough], I believe my colleagues are attending training and the thought processes and protocols in [Borough] are changing’. 8.10 The potential for the training to impact on other agencies was raised in several interviews, for example: 44 ‘What we see here, in [borough] sadly, is police officers referring the victim of the last incident, rather than looking at the wider history…What I'm hoping that the Restart programme will achieve and [what] the training will achieve is people looking at the bigger picture, not just going on, you know, the victim’s reaction to a particular incident, [but services to] look at the bigger picture of...what’s going on, look at the history, look at the coercive control element.’ Community safety officer, focus group. 8.11 Consequently, it is possible that the provision of training to other agencies could lead to a more holistic approach. It is recognised that in some cases of domestic abuse, it is not always straightforward to identify the victim-survivor and co- partner violence is a recognised type of abuse. In one case brought to our attention, the person identified as a victim-survivor was in reality the perpetrator and this is not uncommon. The potential for the training, including the additional workforce development training delivered as part of Restart to upskill workers around counter-allegations and identifying primary perpetrators, could be investigated further during the remainder of the pilot. 8.12 There have undoubtedly been some positive practice changes arising as a direct result of the project even though these may be difficult to measure, for example: ‘Overall, the Council has been kind of encouraged, forced to think more widely and differently about the whole family. So, in those cases, we’ve had some really good, wide-ranging, thought-provoking, system challenging conversation. So, we’ve done things we wouldn’t normally do.’ Strategic lead, interview. 8.13 However, the Implementation Leads make it clear that Restart generally and they specifically are not there to provide all the answers themselves but to help practitioners to build up their own expertise and capacity to deal with complex situations involving perpetrators. This is also seen as one way of finding longer term solutions, for example: ‘Social workers, rather than thinking, OK, maybe I can do this piece of work with this father, for example, rather than thinking who can I refer him to and then I won’t have to do anything about it again… Practitioners are realising that if they don’t manage to engage someone, if they’re not successful in terms of being able to…. there’s still value in terms of you having tried and how you record that and how you document that which is such a key part of the model.’ Implementation Lead, focus group. 45 The point about recording information differently from in the past was recognised as being important in several interviews. 8.14 The new approach engendered by Restart and the Safe and Together training was also seen as a less adversarial and ‘abusive’ way to provide support to victim- survivors and their families. One senior manager described victim-survivors feeling like ‘the victims of state abuse’. Interviewees, particularly those from CSC told us that Restart, with its support for the whole family, was better than the traditional approach which was described as being to ‘separate and isolate’ families which was seen as potentially damaging to victim-survivors. One of the disadvantages of this approach was seen to be the risk of having children removed if the couple do not wish to separate. As a consequence, some felt that Restart has potential to lead to better relationships between families and social workers because it was ‘less coercive’ and as a way to build positive and trusting relationships with families. One frontline practitioner in middle management told us: ‘Historically we’ve kind of taken the position of, if you don’t leave him, we’re going to child protection, you’re not safeguarding your children and almost mimicking the behaviour of perpetrators in terms of controlling behaviour, abusive language. And I think that’s really important to try and build that basis of a relationship to enable parents to engage with you to trust you. And to actually have that belief that you’re actually here to support them and do the right thing rather than almost trick them into another sense of insecurity.’ Frontline practitioner, focus group. 8.15 There was a wealth of evidence from interviewees and survey respondents about the fact that practitioners are starting to think differently as a result of the training. The Safe and Together training was introduced during the current project to train 400 social workers on how to hold abusive parents to account. The training has been well-received by interviewees who felt that, despite being a considerable commitment in terms of time, it was worthwhile and had led to a change in their practice. A number of survey respondents commented on this: ‘I am able to change my thinking, the programme is rolling out and still new...I believe my colleagues are attending training and the thought processes and protocols... are changing.’ ‘Since being on the training, it has helped to have open discussions with social workers and colleagues about using this approach when working with domestic abuse. It has given me the confidence to challenge other professionals, improve my documentation and work with perpetrators.’ 46 8.16 Several focus group attendees mentioned the Safe and Together training as one of the key successes of the programme, commenting’ ‘there's a lot of awareness being raised around domestic abuse in the various boroughs and in various ways with the various leads and the Safe and Together model and that training. And, you know, that kind of systems change with social workers and educating social workers…that's been hugely helpful.’ Delivery partner, focus group. 8.17 Within our survey, 75% of respondents also felt the Restart model had changed how they thought about domestic abuse. The below table outlines the impacts that practitioners reported due to Restart. Table 12: Impacts reported by practitioners Changes to how I think about and respond to domestic abuse 75% Changes to how I think about the impact on children and young people 70% Changes to how my colleagues think about the impact on children and young people 55% Changes to how other services think about domestic abuse 55% Changes to how services respond to domestic abuse 55% This was further supported by Safe and Together implementation leads, one of whom commented: ‘I am noticing a difference in how survivors are being treated and the type of services that are being made available and [a] movement away from the ‘failure to protect’ language.’ Implementation Lead, focus group. 8.18 There was some evidence that this was starting to change practice even if it did not lead to direct referrals to Restart, for example: ‘This [Restart] is quite new, in terms of having something and just the fact they’re talking about it…What it means is that those conversations are happening, and people are thinking about it with all of their cases, and recording and documenting differently, which is the real change, which I think is so important, because it’s that, that has massive wide-reaching effects.’ Implementation Lead, focus group. ‘I’ve been doing some work with early health practitioners who have really struggled with having these sorts of discussions with survivors. I think a real 47 success has been the confidence level of practitioners, this is sort of from discussions and from the management feedback has been that they’re seeing that practitioners are more able to go and have conversations with perpetrators where they weren’t, and more confident to have conversations about harms.’ Implementation Lead, focus group. 8.19 Interviewees spoke positively and indeed enthusiastically about the effect that the project has had on their relationships with other agencies (and, in some cases, with other parts of their own agencies). Multi-agency and multidisciplinary working were recognised by many as essential elements of this project and the view was expressed that this would have some impact on their practice in working with other agencies. Some spoke enthusiastically of working as a team to deliver Restart. As an example, this approach was seen to impact on a case whereby a victim-survivor had been wrongly identified as a perpetrator. A focus group attendee said: ‘Because we work in that multi-agency, multiple partners [way]...we all get together quite regularly. We had a case where we had a female perpetrator, and in discussing this matter…with our partners and then with the borough, and then with various practitioners, we were able to identify that…she was the victim.’ Implementation Lead, focus group. 8.20 Linked to this is an improved partnership between agencies and different parts of the local authority. This was cited by several interviewees as a positive development since it brings together Children’s Services, housing and the agencies involved in Restart, for example: ‘We’ve done a lot of work recently to kind of strengthen the way that housing and social services work together...It’s been good to kind of get to know colleagues in Social Services and get to work with them a bit more closely.’ Strategic lead, interview. We were also told by a senior manager that the project provides an opportunity to link more closely with other statutory partners such as the NHS and police. Improving safety and outcomes for families 8.21 Nearly all interviewees felt that the Restart approach has much to commend it and has potential for achieving systemic change and improving outcomes for families by bringing about increased accountability of those perpetrating the abuse. This was in spite of the reservations expressed about the delivery and complexity of the programme (see section 9). For example: 48 ‘I think it has potential in really unmasking a lot of the actions and harms that perpetrators do, and really changing this idea of ‘let’s close the case’, because the perpetrator doesn’t live in the home.’ Implementation Lead, focus group. ‘If we can deliver the programme in the way that Cranstoun intend us to, I think there’s huge potential there…. but will it work every time? Probably not. But will it help some families? Absolutely.’ Frontline housing staff, focus group. There was also general support for the programme which seems to have an instinctive appeal, as one interviewee put it: ‘There’s nobody I have spoken to about Restart that does not have that Wow!’ Front line practitioner, focus group. 8.22 Although many felt that Restart had the potential to keep families safer, there was little quantitative evidence of this having happened so far. There was, however, anecdotal evidence including case studies which were highly positive, particularly in securing the commitment of service users to longer term behavioural change. In case study 3 (see Appendix 2), for example, the perpetrator has committed to a behavioural change programme – whilst the victim-survivor does not feel safe enough, even with a non-molestation order in place, to remain in her home, the commitment to the behaviour change work is a positive outcome. Overall, 15 of the 76 referrals have resulted in referral to a behavioural change programme (20% of the total) as follows: • 11 have been referred to Men & Masculinities group programmes (run by Cranstoun) • 1 has been referred to DVIP • 3 have been referred to Rise Mutual’s DVPP It should be noted that not all of the 76 cases will have reached the stage at which they might have been referred to a DVPP since this will include a number who do not engage at all and some who are still within the four-week early engagement and assessment period. 8.23 All participants and survey responses who spoke positively about the programme felt that the shifting housing narrative and practice brought about as a result of the accommodation pathway is beneficial, particularly when thinking about longer term trauma and disruption caused to the family by the abuse. One survey 49 respondent told us ‘Looking at the perpetrator moving out will be an improvement as less disruption for the family’. 8.24 Two people who used harm responded to our survey; they both positively responded to the programme, stating that they felt they had ‘much less conflict since project involvement’ and that ‘I am able to have a better relationship with her’. However, without comparable data from victim-survivors, we are unable to speak more about their partners’ sense of safety or to identify objectively any reduction in risk levels. It is hoped that further evaluation will elicit a greater number of responses (see section 13). 8.25 The most helpful aspects of Restart for these two men are outlined below. Table 13: Most helpful aspects of Restart The emotional support I’ve received (2 agreed) Strengthening my relationships with my partner/ex-partner (2 agreed) Positive impact on my children (2 agreed) Helping me to understand the impact I have on other people and to change the way I react (2 agreed) Helping me to understand my housing options and rights (2 agreed) The practical support I’ve received (1 agreed) Helping me to find somewhere to live (1 agreed) 8.26 One respondent to the survey told us Restart has helped by ‘learning to accept that I can separate/co-parent respectfully, feel supported by people involved did not realise I could accept support.’ 8.27 There was recognition that those who use abuse would want to change through the intervention and that this was, for some, an opportunity to do so, for example: ‘Especially if they’ve got kids, whether they’re with their partner, they want to build a relationship, even if it is just for the child’s sake. If they’re not together, just to have that mutual respect for each other to help raise their child, and obviously if they’re together, they want to make their relationship better for each other and for their child.’ Case Manager, interview. 50 8.28 Some of the case studies showed that because specialist perpetrator workers were working alongside social work, situations that were dangerous were flagged and escalated, and in some cases, re-opened. One case study tells us: ‘The Case Manager has held two consultations meeting with the social worker and updated the social worker throughout the intervention – this case has highlighted concerns with the process within CSC. The Case Manager articulated these concerns, particularly around potential collusion with the father by the Social Worker. Furthermore, when the Social Worker explained that the case was closing due to no further safeguarding concerns, the Case Manager escalated our concerns over this. Due to escalation in risk, the case was handed over to the Service Manager to liaise with CSC.’ 8.29 There were several examples given of service users in whom the four-week programme had led to behaviour change through the intervention of the Case Manager, or at least a willingness to consider the impact their behaviour is having, and to accept referral on to a DVPP (either group or an individual programme delivered by Cranstoun or another agency). This applies whether or not the service user was still living with the family. ‘And for the guys I've worked with, a lot of them have said that it's helped them communicate with their partners better and to even build a relationship where they're probably wasn't a relationship before and just kind of know how to deal with conflict and kind of learn more stuff about themselves and why they've acted like that and what they want to change, what they kind of want to see for themselves in the future.’ Case manager, interview. 8.30 During our observations at case management meetings, it was also clear that the four weeks allocated to behaviour change is often, in practice, best used as a period of relationship building between service users and case managers or partner support workers. As one focus group attendee notes; ‘Because if we're asking perpetrators to move out their home for four weeks, and then possibly longer-term solutions, are they just digging their heels and saying, I'm not moving out of this family home? I'm willing to work with you, but I'm not moving. And that comes down to how you sell it, doesn't it?... you are asking families to make big changes in their structure, in their dynamic in their parenting...it's a big thing.’ Frontline practitioner, focus group. 51 8.31 The work of the Case managers (or VPPs) is evidently crucial to the impact of the programme – their expertise in dealing with perpetrators is valuable not only for its potential in changing an individual’s behaviour but also as a source of advice for the other practitioners dealing with the case. The relationship with the social worker is key to achieving the best outcome for the family, particularly since the social worker can help to help them engage where there are problems in doing so, for example: ‘If I feel I’m struggling with that engagement, I’ll just revert back to the social worker so they can they be aware that they’re not engaging. Can you help with the engagement side of things and just kind of leave it with them until I get some feedback.’ Case manager, interview. Advantages of the accommodation pathway 8.32 The perceived benefits of the accommodation pathway are set out in a separate section, see section 10. Flexibility of the intervention 8.33 The flexibility allowed by the intervention was welcomed. Whilst intended to be a four-week intervention by the Cranstoun team, there is flexibility in this which is necessary since it may take some time to engage with or even contact the service- user: ‘It hasn’t been rigid: ‘it must be four weeks and then they’re out the door’... The fluidity that we’ve been allowed…has benefitted the project hugely.’ Delivery Partner manager, focus group. 8.34 One case study which illustrated the flexibility in applying the intervention involved a service user who engaged fully with the assessment process but had concerns about starting group work. One-to-one sessions were therefore arranged as an alternative which enabled the Case Manager to start behaviour work and to tackle his denial and partner-blaming stance. This flexibility was also evident in our interview with a Case Manager who said that it might take more than four weeks to complete the assessment and that, if the service user then finds it difficult to engage in a group activity, they will offer a 12-week intervention on a one-to-one basis which works well. This is particularly effective in those who do not speak English well enough to join the group work and who require an interpreter. 8.35 During our fieldwork, we found differing responses from practitioners to the project, with some practitioners thriving with the fluidity and flexibility afforded by 52 the pilot, and others finding the ‘learning by doing’ approach more challenging. As two focus group attendees noted: [they keep]’saying it’s a pilot, you know, we’re here to learn, we’re here to grow. We’re here to listen to ideas, which I think this is how we need to think of it, there is no straight way to go through housing.’ Children’s Social Care, focus group. ‘Whenever you start something new, it can be a little bit messy, maybe?’ Children’s Social Care manager, focus group. 53 9 Findings: challenges 9.1 A number of challenges were identified during the interviews which are summarised below: Set-up and mobilisation 9.2 Although mobilisation was slow in some boroughs, we were told by several interviewees and within the surveys that it felt rushed. It should be noted that the programme started in August 2021 and that the impact of Covid-19 was still being felt at that time so this may have had some impact. For example, there were still few opportunities for face-to-face meetings. In addition, the funding requirements (i.e: with funding initially approved for only a year) meant that there was pressure to mobilise quickly, which meant that mobilisation and delivery of the intervention were happening at the same time. Insufficient time was built in to the mobilisation period to allow for the complexity and challenges of implementing a project on this scale. We were informed that Cranstoun were being required to pick up referrals, for example, before their service manager or accommodation support worker were in post. ‘I think the way the pilot had to start made it so that it was difficult to clearly communicate to all of the partners exactly what we are and how we work together.…We didn’t have the time to figure out the communication plan, the recruitment, what our name is and then communicate with the local authorities, we had to do all of those things at exactly the same time which I think just creates more problems down the line…It’s a problem of funding restrictions...pressures…time limits that make it so you can’t properly mobilise and communicate and do all of those things.’ Representative of partner agency, focus group. A survey respondent told us that: ‘The project was still getting up and running while clients [were] coming into the service.’ 9.3 One interviewee attributed the difficulties to the short-term nature of the funding which inevitably placed pressure on the team to deliver quickly: ‘And when the funder is saying do it within this time period and get these results, then you’ve got a lot of pressure on you to do that, and it leads to problems down the line. Which means you can’t deliver the results you want. So, you know, there’s not as many referrals right now is probably a symptom 54 of the problem which is how it got started.’ Representative of partner agency, focus group. 9.4 Communications during the setting up of the project were drawn to our attention as an issue (see also paragraph 9.49): ‘All the kinds of communication stuff that you would choose to do at the very beginning, they had to do six months down the line when they finally got everybody in post and a fully functioning team.’ Representative of partner agency, focus group. Recommendation 1 Time and capacity for setting-up Restart: We recommend that if Restart is scaled up and rolled out to other local authorities, sufficient time should be built in from the start to allow for proper preparation, the partnership to form and the vision to be developed collaboratively. Culture and attitudes including training 9.5 Although the emerging culture changes resulting from Restart were identified as a benefit (see paragraph 8.8 et seq), it was recognised that the changes in culture and attitudes that are needed for it to succeed will take a considerable time to achieve and embed throughout services. The change in attitudes was alluded to by several respondents as a challenge, with recognition that this is the start of a complex process and that the project should run for a minimum of three years to allow time for this to happen. 9.6 Some noted that attitudes such as victim-blaming attitudes persist and deter victim-survivors from working with social work teams. One survey respondent noted ‘There is a culture in among social workers to blame the victim for not leaving the abusive relationship.’ A number of focus group participants also noted the challenges of victim-blaming cultures and practices, for example: ‘Some social workers that don’t engage appropriately, still have a victim- blaming approach rather than a victim-centred approach at times….I think culture-wise, in general, it’s not just police officers [who victim blame], it’s all professionals, then mistakes are made right across the board. But sadly, mostly that what I see here, it’s the police. And that’s the position that we’re in. But it’s a learning need.’ Front line practitioner, focus group. 9.7 Similarly, some participants raised that for some practitioners, investment in perpetrators may be seen to take resources from victim-survivors, which would be 55 particularly problematic if the programme was scaled up whilst there exists an “acute lack of funding” for victim-survivor services:19 ‘I get why people feel apprehensive about housing…perpetrators but they are people, they deserve a roof over their head.’ Case manager, focus group. 9.8 One of the ongoing challenges for any perpetrator programme is that there is no way to ensure abuse is going to end through the intervention, and without longitudinal assessments of the perpetrators involved, it is unknown whether behaviour change is sustained beyond the intervention and the oversight from services. As some focus group attendees noted; ‘You can put people in programmes, do they really change? Who knows? They might go back a week later, and, you know, assault somebody.’ Frontline practitioner, focus group. 9.9 The lack of experience of working with perpetrators was raised in several interviews and this was more evident in some of the boroughs where perpetrator programmes had not been commissioned previously. The Restart manual and the Safe & Together training both address this issue, but putting it into practice is nevertheless challenging for practitioners whose main business has previously been to support victim-survivors and their children who have never worked with perpetrators directly before. Even where there is strong commitment to doing so, the challenges involved should not be underestimated. Views from practitioners included: ‘I think the problem is that many of us haven’t worked previously with perpetrators per se. We need to be more involved with fathers from the first instance at Early Help and speaking with perpetrators, challenging their beliefs and making them accountable for their actions. All too often we expect the victim to take the safeguarding responsibility without addressing the cause' and ‘What I’ve learned not to do is to use...words like abuse, perpetrator, it’s all about changing, making different parenting choices...because a lot of them, it’s to do with contact issues if they’re not together...so, what do you want in the future of your children? How do you want your children to see you?’ Frontline practitioner, focus group. 9.10 It was also strongly emphasised that different local authorities engaged differently with Restart. One focus group attendee, for instance, spoke of the importance of 19 https://domesticabusecommissioner.uk/early-findings-from-our-mapping-show-a-huge-discrepancy-of- services-across-england-and-wales-and-an-acute-lack-of-funding-that-prevents-services-being-able-to-meet- demand/ 56 strategic management support as well as middle management support, in clarifying the referral process for frontline workers, and in explaining how Restart differs from other perpetrator intervention programmes and how housing stock should be allocated. 9.11 We noted, as a minor issue, that the issue of the language used within the Safe and Together training materials was raised by a minority of interviewees. Since these are American and this is a licensed programme, there were felt to be some differences in the language which some people found off-putting; terms such as ‘batterer intervention’, for example, are not commonly recognised in the UK. We recognise the difficulties in amending any training materials given that this is a licensed and trademarked training programme.20 Whilst it may be possible to learn from other areas of the UK who have embedded Safe and Together in how to navigate this challenge (most notably Edinburgh City Council), the Scottish approach to domestic abuse defines it as a gendered violence between partners or ex-partners, rather than gender-neutral or as inter-family abuse. These differences mean that whilst it is important to learn from other areas, resources and training for practitioners should reflect the English context. The definition of domestic abuse used in the UK and incorporated into the Domestic Abuse Act 202121 is broader than the definitions used in the US or Scotland and now includes the impact on children who see, hear or witness domestic abuse. 9.12 Whilst feedback on the training for those who had completed it was highly positive, some interviewees expressed the view that agencies other than CSC (such as police and NHS practitioners) would benefit from participation and that embarking on the training as a joint exercise would help to foster even closer interagency relationships and understanding. We were told that Housing Officers might also benefit from the training, but it was thought to be too detailed and not targeted enough for them to benefit as the training currently stands. The one-day overview training on Safe and Together22 would be suitable for Housing Officers and other agencies. 9.13 Similarly, interviewees in CSC considered that all social workers and senior managers should participate in the training since it would give them more of an 20 See https://safeandtogetherinstitute.com/about-us/about-the-model/ 21 See https://www.legislation.gov.uk/ukpga/2021/17/section/3/enacted 22 The one-day overview training provides participants with an introduction to, and overview of, the Safe & Together™ Model. The training provides participants with information about creating a domestic violence- informed child welfare system, the principles and components of the Safe & Together™ Model and information about the framework behind competency-building in child welfare around domestic violence. (Taken from Safe and Together website.) 57 understanding of the importance of working with perpetrators and the systemic changes which this could help to achieve. One focus group attendee commented: ‘I think the Safe and Together training should be compulsory for social workers. There shouldn’t be like a choice, they should just have to do it.’ Delivery partner, focus group. 9.14 Whilst the low completion rate (for the Core training) was disappointing, those who had attended spoke highly of it and felt it was beginning to make a significant impact on their professional practice. It is difficult to measure the impact of the training but it was recognised that this is an essential step in achieving the systemic changes needed if Restart is to succeed. The 63% who partially completed the course may also have gained sufficient knowledge to enable them to start to change the way that they work with families and may be able to complete the training at a later stage. It should be noted that the core training is now moving from online to four days face-to-face training although whether this will increase the completion rate is uncertain. 9.15 This undoubtedly reflects the pressure on social workers and the broader workforce pressures which are impacting on recruitment and retention of staff. Many focus group attendees described the difficulty experienced in their carving out time to complete the important Core training. However good the training is, it will not benefit people who are too busy to complete it. Looking at other areas where Safe and Training has been rolled out, such as across Edinburgh City Council23, may be helpful. Capacity of services to deliver Restart 9.16 The pilot has taken place during a period of considerable and growing uncertainty with local authorities facing pressures. These include recovery from the pandemic, rising demand (particularly in CSC), fiscal pressures, poor morale and a high staff turnover, increasing complexity of cases and, recently, the fall-out from two serious case reviews which have raised questions about the quality of Children’s 23 See https://www.edinburgh.gov.uk/domestic-abuse/safe-together- edinburgh#:~:text=The%20Safe%20and%20Together%20model,abuse%20and%20concerns%20about%20childre n. 58 Social Care24 25 26. The Restart pilot is therefore being delivered during a difficult time with social workers coming under considerable pressure on a daily basis. This has manifested itself in a number of ways and may help to account for the poor response rate for the survey, the difficulty for some in completing the Safe & Together training and the reluctance expressed by some interviewees to take on what they perceived to be additional work and new ways of working. In particular, several interviewees raised the difficulty caused by social workers wanting to, and coming under pressure from management to close cases, even where there were ongoing problems in the family and the underlying issues had yet to be addressed, for example: ‘Social workers desperately want to close cases and to refer on, that’s what they’re desperate to do, because they’re you know, drowning in cases and work…. But, of course, no-one’s done any work with the person causing the harm, there’s not been any kind of meaningful partnership work with the family.’ Implementation lead, focus group. 9.17 Other focus group attendees noted that time constraints and information overload has also limited social workers ability to take on board the Restart programme, despite the efforts of others within the Boroughs, with one commenting: ‘[we]continue to raise the profile and put it on people's radars .. we're conscious just how busy social workers and teams are...they get bombarded with lots of bits of information...we'll do the team meetings and other things and people come to team meetings about Restart across the service, but then it's very easy for that just to people to note it, and then kind of go off and forget about it.’ Representative of delivery partner, focus group; similarly ’social workers are inundated, they get 1000s of emails, do they actually read [the newsletter]? I sometimes think that they don't.’ Delivery partner, focus group. 9.18 It was recognised that this can lead to a vicious cycle with the same families being referred repeatedly to CSC since the underlying reasons for the initial referral 24 https://theconversation.com/child-protection-in-england-here-is-what-social-work-experts-know-must- change-in-the-system-182072 25 https://www.gov.uk/government/publications/childrens-social-care-2022-recovering-from-the-covid-19- pandemic/childrens-social-care-2022-recovering-from-the-covid-19-pandemic 26 See National Review into the murders of Arthur Lobinjo-Hughes and Star Hobson, The Child Safeguarding Practice Review Panel, 2022 https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1078488/A LH_SH_National_Review_26-5-22.pdf 59 remained unresolved. There were several examples provided of the project practitioners (in particular, the Safe and Together Implementation Leads) intervening to try and prevent case closure. In one case study, for example (Case study 1 in Appendix 2), CSC wished to close the case despite concerns about continuing safeguarding concerns. 9.19 This is understandable given the pressures that social workers are under, and there was a view that CSC in some local authorities has a target-driven culture with insufficient time available for training. It was also recognised that leaving the cases open may be more resource-intensive, at least in the short term, since it is an acknowledgement that further work needs doing. Restart is not intended to be a way of ‘offloading’ the case following referral, and this needs to be emphasised. 9.20 During data gathering, many focus group attendees and interviewees also flagged issues around workforce retention, burnout and exhaustion and how difficult it is for social work teams to engage with new approaches. This was seen as partly related to having to manage multiple and complex cases, but also (by social workers) to the levels of paperwork required to refer service users onto Restart although the referral process has since been simplified. The normalisation of burnout and social workers exiting the profession may then present as a barrier to bedding in new approaches due to the lack of time and space to reflect on practice and the lack of support in dealing with complex cases. As one strategic lead told us, ‘You’ve also got to take into account the turnover of social workers.’ Strategic lead, interview. One focus group attendee told us: ‘we’re conscious just how busy social workers and teams are...they get bombarded with lots of bits of information.’ CSC manager, focus group. 9.21 The high turnover in staff has also proved problematic since incoming staff are not aware of the intervention nor of the roles of individual workers such as the Implementation Leads, leading to a sense of confusion around who does what (Focus group attendees). Restart can assist to an extent, by increasing the physical presence of both case managers and Safe and Together leads being co-located. However, this remains a key issue for the extension of the evaluation to consider. Recommendation 2 Safe and Together training: We recommend that consideration should be given to ways in which the reach of the Safe and Together training can be extended to encourage a greater take-up and completion of the training by social work staff including senior social care staff. Frontline practitioners in other parts of the local authority such as 60 housing and other agencies such as the police and NHS should be encouraged to attend the one-day overview training. Referrals 9.22 The number of referrals was lower than anticipated initially, although the level was significantly increasing by the time we wrote this report – there were 76 referrals up until 1st July 2022. The number of cases broken down by borough show considerable variation between the five boroughs. 9.23 There seem to be several reasons for this, including the amount of work and perceived complexity in making a referral, but also a lack of interest in the intervention as an option by staff at the social work ‘front door’ and a low level of commitment at middle management level. (This may also explain their reluctance to engage victim-survivors in completion of the survey.) We were told: ‘Actually, making a referral involves a huge amount of work. There’s a lot of paperwork that you have to fill in. And that’s probably a bit like, do I really want to do this?’ Social worker, focus group. There also seem to be some misconceptions about the nature of the intervention and perhaps an overemphasis on the housing element which may deter some practitioners from making referrals to what is in reality a broader intervention with several components. 9.24 In some authorities, we did identify a reluctance to refer cases which might benefit from referral. The appointment of an IDVA in CSC in one had made a difference and was encouraging practitioners to refer to Restart. 9.25 The lack of a commissioned external DVPP to refer on to may be an obstacle – three boroughs have only started to commission a DVPP relatively recently. In Havering, for example, there had been no DVPP commissioned for many years until recently when Cranstoun were commissioned to deliver their Men and Masculinities programme. This had proved successful and formed a firm basis on which Restart could then be based. 9.26 At present, referrals can only come from Children’s Social Care and this was flagged by some focus group participants as a barrier to referral rates: ’The only…way is a referral is children’s social services, it would be really nice to... educate some of the police officers.’ Representative of delivery partner, focus group; similarly: 61 ‘There’s only one referral partner at the moment, which is Children’s Social Care, which obviously narrows down the field significantly.’ Delivery partner, focus group. 9.27 Whilst there is as yet little evidence to identify or demonstrate which families are more likely to benefit from the intervention, some interviewees considered that it was more likely to help those who wish to stay together. It was felt that the provision of alternative accommodation with the behaviour modification aspects of Restart, even on a short-term basis, may help to provide a respite for families and provide an opportunity for reflection during the four-week intervention. This should be addressed during the next phase of the evaluation. Recommendation 3 Referral to Restart: We recommend that the central team should consider ways of encouraging referrals if these remain lower than expected, looking at the reasons identified and examining whether the referral process could be simplified. The wider dissemination of positive findings from the project and successful case studies might help to allay any concerns and publicise what can be achieved, supported by Restart champions. Recommendation 4 Suitability for referral: The next phase of the evaluation should try to establish through monitoring outcomes which families are most suitable for the intervention and who is most likely to benefit from referral. For example, it would be useful to know whether outcomes may be better in families wishing to stay together or in those who are separating and how the abuse profile changes throughout the different intervention stages. Engagement of families 9.28 If the perpetrator does not wish to engage, the intervention cannot proceed as it is based on consent. The training available on engaging with those who use abuse (including in the Restart Early Intervention and Accommodation Manual and the Safe & Together training) may increase the participation of service-users by skilling up frontline practitioners who have no previous experience of this approach. There is also an incentive to engage for those who wish to maintain contact with their children, to resume the relationship with their partner (where appropriate) and for those who wish to become better parents. 62 9.29 The risk of ‘disguised compliance’ was, however, seen as a risk by some since service users could show willing by participating in the four-week intervention but have no real intention of changing their behaviour.27 This possibility increases the need for professionals to risk assess and communicate with the other professionals involved throughout (it also shows the importance of longer-term monitoring on behaviour change). This is one of the areas the Safe & Together programme can help to address, with its emphasis on understanding the victim-survivors’ lived experience and the need for any plan to be victim led and informed by understandings of the perpetrators’ pattern of abuse. 9.30 If victim-survivors do not wish to engage with Restart, it is difficult to proceed or at least less likely to achieve a satisfactory outcome as there is a need to ensure any work undertaken is led by victim-survivors’ wishes and safety considerations. In at least two of the case studies that we saw, it was difficult to engage with the victim- survivor despite the best efforts of the Partner Support Worker. Engaging both the victim-survivor and the perpetrator rather than one or the other and showing how they and the children may all benefit is therefore preferable. The Partner Support Worker has a key role to play in this. 9.31 Challenges also exist around the perception and lived experience of families and their willingness to engage with statutory services, particularly for racially minoritised families. Research tells us that for many minoritised groups, there is a real fear that engaging with any services may result in the removal of their children. This fear was highlighted by several focus group attendees around social work practice and culture: ‘[in social work we had] that culture of you do this, or we’re taking your kids. And unfortunately, all you’re doing is you’re coercive controlling a victim into making a decision they’re not ready to make [regarding separation].’ Frontline practitioner, focus group. 9.32 Another challenge identified by some participants was the lack of specific support for children and young people. We were told: 27 Some professionals noted the risk of service users exhibiting ‘disguised compliance’ by agreeing to participate in the four-week intervention but have no real intention of changing their behaviour. It was recognised by professionals within the sector who work with service users that this is always a possibility and thus the need to risk assess and communicate with the other professionals involved throughout. Similarly, it was noted victim-survivors can also sometimes be labelled as showing ‘disguised compliance’ when for example CSC conditions are imposed on them. This is one of the areas the Safe & Together can help to address which emphasis understanding the victim-survivors lived experience and the need for any plan to be victim led and informed by understandings of the perpetrators pattern of abuse. 63 ‘In my ideal world, I’d have another layer. And I would have children’s counselling, children’s intervention, working with children’s understanding, and also educating them about healthy relationships, because their example has been unhealthy.’ Frontline practitioner, focus group. ‘There’s a huge gap with domestic abuse provision for children generally, and certainly within this project.’ Frontline practitioner, focus group. 9.33 One of the advantages of this would be that it could help to reduce the risk of intergenerational transmission of domestic abuse by supporting children who have experienced trauma as a result of the abuse. Unfortunately, it was not feasible to capture the voices of children and young people during our evaluation (see section 7) but it is hoped that the next phase of the evaluation will include an effective means of eliciting feedback and assessing the impact of Restart on their wellbeing (see section 13). The voice of children, young people, victim-survivors and (where feasible) those who use abuse should be sought actively and heard in delivering Restart. Recommendation 5 Support for children and young people: We recommend that consideration is given to providing specific support for children and young people as an integral part of the intervention, perhaps by having a dedicated practitioner as a resource for local authorities as there is for victim-survivors/partners. This could be done centrally as a resource for all the boroughs. Recommendation 6 Stronger monitoring of outcomes: various monitoring and evaluation tools already exist with the Safe and Together Training, the four-week early engagement and the in-borough DVPPs. However, there is a need to strengthen (and potentially link up) monitoring processes so that we gain a better picture and understanding as to how outcomes, particularly those which measure risk for families and behavioural change in perpetrators, are changing over time as a result of different intervention arrangements. Complexity and clarity of purpose 9.34 The complexity of Restart was raised by several interviewees as a challenge which deterred them and their colleagues from engaging fully. They found it difficult to understand the different layers of the interventions and how the components fit together, making buy-in from them and their colleagues difficult. Some practitioners found it difficult to understand who was responsible for what and the boundaries between the different components of the intervention, particularly as boroughs offered different longer-term behaviour change programmes to follow 64 on from Restart (e.g.: DVIP, RISE, or Men and Masculinities), as well as programmes that focused on varying levels of risk. ‘I was very lost because there are so many layers to Restart and the programme, getting my head around it and not being involved in conversations from the start, was a bit of a tricky place to be.’ Frontline practitioner, focus group. ‘Maybe there’s been some confusion because it’s a model that comprises so many different aspects. There’s Safe and Together, there’s accommodation, there’s support for the person using violence and abuse. I think sometimes boroughs have just been overwhelmed and not really understood the model fully. And that’s feedback we’ve had in the community of practice groups as well.’ Representative of partner agency, focus group. 9.35 Adaptations made to the operation of the programme after it started appears to have added to the confusion felt by some front-line practitioners: ‘I think the actual project has gone through too many changes since it’s come in. And it’s been difficult for us as receiving people who are receiving the service to keep up with those changes, and to kind of understand what is happening now.’ Frontline practitioner, focus group However, this may also be due to the project building on the recommendations of previous evaluations around having a distinct identity as well as growing organically and adapting in response to emerging evidence and needs. 9.36 Implementation Leads explained to us that they spend some time explaining their role and the different components of Restart and Safe and Together to colleagues, for instance: ‘I have to spend a lot of time explaining, you know, where I sit that I’m not doing an intervention, I don’t work with families, I’m there to support social workers working with the whole family.’ Implementation lead, focus group. In boroughs where there is a high turnover of social care staff, this may cause some frustration since incoming staff will know nothing about the programme and will have to have this built into their induction. 9.37 Since the pilot is operating in five different boroughs, there is the added complexity of each borough having its own systems and different approaches. One of the boroughs has had considerable financial difficulties and the project has therefore been implemented during a period of considerable internal upheaval 65 with high staff turnover and workforce capacity issues (see paragraph 9.16 above). In the same borough, there appeared to be a lack of collaborative working which is essential to the successful delivery of Restart. The Implementation Leads and other project staff have to be flexible in their approach and to integrate their own ways of working with the local authority’s system. 9.38 There was some evidence of a lack of clarity about the aims and objectives of Restart leading to poor commitment and buy-in from frontline practitioners. Communications were regarded as unclear by some interviewees and there was confusion evident amongst staff as to the difference which Restart could make: ‘I was kind of led to believe that it was sort of taking the perpetrator out of the home, working with them, giving them that breathing-space, and then potentially, for them to go back. Obviously, since then, on our side is we’ve been working with clients who can’t go back, but don’t necessarily need a homeless application. So yeah, it’s been a little bit shaky. Like it wasn’t explained to me properly, the outcomes of it.’ Frontline practitioner, focus group. ‘I’m constantly confused, because when we’re discussing the layers, for the introduction to things, I have to sort of think back and think right, we’ve had that conversation and we’re moving on to another conversation.’ Frontline practitioner, focus group. 9.39 The way in which the project is presented may be a factor in this, since some assumed that the project would be focused on accommodation; for example: ‘I think that’s because Restart was seen primarily about housing and accommodation. And I think that’s always been the kind of tagline, this sort of innovative housing pathway where perpetrators are housed rather than victims. Actually, there’s so much more to it.’ Implementation lead, focus group. 9.40 Whilst the project and its predecessor were intended to achieve early or at least earlier intervention for families, we were informed that this is not necessarily the case. An analysis of the Severity of Abuse Grid (SOAG) data (see Table 5 in Findings and Analysis which shows those clients who received SOAG assessments) illustrates that some of the families referred were experiencing a high level of harm on referral (according to their classification) and that the abuse may have been going on for some time. 66 9.41 Some referrals made to Restart may therefore include some where the abuse is long-term and higher risk despite the fact that all the cases referred remain below the MARAC threshold; for example: ‘We’ve taken on some…higher risk, not high risk necessarily, but higher risk referrals. And a lot of them are not early on in the process……I was told it was early intervention. So, you know, and I said, what does that mean? And I was told it was, you know, the first few arguments, the first police call-out the first time there’d been a sign of physical abuse of, you know, all the firsts. It’s not like that in practice. And it’s very difficult for it to be like that in practice.’ Delivery partner, focus group. 9.42 In areas where Drive is operating, this is perhaps less likely to be an issue since the most serious cases would be referred to Drive as a more appropriate intervention for high risk, high harm cases. In Sutton, for example, where Drive is operating, the Restart Case Manager has referred cases to Drive which were considered too high risk for Restart (or the associated DVPP which was available). In Westminster, where there is no high-risk intervention programme in place to refer people to, this was seen as more difficult since ‘social workers [are] kind of scratching their heads not knowing what to do with their clients’ Delivery partner, focus group. 9.43 Different views were expressed about what constitutes early intervention, some taking the view that the fact that the family had been referred to Children’s Services meant that it was already a high-risk case and therefore too late for early intervention: ‘One is that I think it’s interesting to me that the referrer is Children’s Social Care, because it feels like in most circumstances, by the point at which social care is involved, and they have a social worker, it’s gotten to the stage where it’s high risk and requires high risk intervention, whereas this is advertised as an early intervention approach where that risk of domestic abuse isn’t high.’ Representative of partner agency, focus group. Whilst this may be true of some families, it should be recognised that many who are referred will not be at high risk or require an intervention for those at high risk. For some families, domestic abuse may be one factor of many which has led to the referral. 9.44 It is feasible that, as the intervention and the training become embedded in the five councils and referrals increase, intervention may be initiated earlier, with practitioners more confident and better able to identify domestic abuse at an 67 earlier stage and to respond more proactively. This would need to be tested out in future work. 9.45 Some interviewees identified the risk that practitioners would refer to Restart with the intention of stepping away from the case and closing it. As one senior manager pointed out: ‘Once the intervention took place, Children’s Social Care were then stepping back and closing the case, which kind of fundamentally misunderstands the purpose of the project and, in a sense, this is not a demand management mechanism.’ Strategic lead, interview. It needs to be clear to practitioners that their continuing engagement is key to the success of the intervention, that the intervention requires a team approach and that this is expected to pay dividends in the longer term. 9.46 The challenge of achieving behavioural change in those who use abuse within a four-week intervention was raised frequently, with many interviewees recognising that this is not feasible due to the complexity of cases and the difficulty in engaging with victim-survivors and the person who uses abuse. In practice, the intervention is frequently delivered over longer than four weeks (which we also cited as a benefit, see paragraph 8.40), particularly where it has been difficult to make contact: ‘So, I was told it was a four-week intervention. And I believed it would be a strictly four-week intervention. Obviously, it never is in practice’. Delivery partner, focus group. 9.47 Developing or revising a shared Theory of Change with all partners and practitioners in each local authority in order to develop a common understanding of change and success may be helpful. It may also be helpful to develop a staged, strategic plan for future programmes outlining how to engage with different local authorities depending on which conditions are in place, as participants have already started to identify the conditions needed to enable the programme to bed in.28 We have outlined these in section 10. 28 A useful reference point is the Tri-Ethnic Centres programming handbook on Community Readiness https://tec.colostate.edu/wp-content/uploads/2018/04/CR_Handbook_8-3-15.pdf and Victoria Health’s guide to encountering resistance; Strategies to respond to resistance to gender equality initiatives https://www.vichealth.vic.gov.au/-/media/ResourceCentre/PublicationsandResources/PVAW/Encountering- Resistance-Gender-Equality.pdf as tools to map “readiness” of local authority areas and what stages each of the boroughs are at. 68 9.48 The limitations of the intervention and also the fact that it is not strictly limited to a four-week period need to be made clearer in communications and in any future measurement of impact to help to manage expectations and build trust with practitioners. The flexibility allowed was welcomed given the time taken to engage with the victim-survivor and those using abuse. 9.49 It should also be recognised that Restart is a short-term intervention and the initial 4-week assessment period cannot be expected to achieve long term behaviour change, particularly where there are difficulties in engaging with the service user. It should be presented as the start of a longer-term process of engagement, in particular with the service user (though the victim-survivor will also frequently continue to be supported by a Partner Support Worker during Restart and during the longer-term DVPP). 9.50 Communications was raised by a few interviewees as an issue where further work was needed to inform staff and bring them up to date with developments. This was addressed differently in each borough, with some including information about the project in regular bulletins. But some interviewees said that they needed and had requested a flow chart to illustrate the structure of the project and had found this difficult to obtain. This was a particular issue when the project started (see 9.4): ‘One of the things we sought to do from the start is to get that buy in from our director and also lead…housing as well.’ CSC manager, focus group. 9.51 Whilst it is clear that Restart is meeting need, given the range of domestic abuse cases which present to CSC, it is important to have a range of interventions which can be applied as one component of an overall domestic abuse strategy. This is the case in one local authority (Camden) where Restart has formed one component of a much bigger strategic piece of work. Some of the participating local authorities have little specific provision for domestic abuse perpetrators which may make it more difficult to find longer term solutions at the end of the four-week intervention period. Restart should not be expected to fill this gap. Recommendation 7 Communications Where Restart is being set up, we recommend that there should be a clear communication plan in place before the project begins, with dedicated resources, and throughout the project so that staff are kept informed of progress. The accessible guidance already developed should be more widely disseminated and included in induction training for new social workers. Communications should clarify: 69 • the aims and purpose of the intervention, the remit, roles and responsibilities of the key agencies and stakeholders; • that the intervention is not always ‘early intervention’ and that the intervention period can be extended beyond four weeks should the need arise, particularly where it is difficult to engage with the person who uses harm and/or the victim-survivor; • that a team approach is essential to deliver Restart; thus it will be unusual for a case to be closed once a referral has been made until the intervention is complete; • that Restart is intended to be one element of a suite of domestic abuse interventions for families which are available in local authorities, with clarity as to how these fit into local domestic abuse strategies. Communication materials may need adapting in each of the five boroughs to show how it dovetails with local systems. It may be helpful for the central team to work with each local authority (and any new ones if it is rolled out) to produce a Theory of Change to ensure that the aims and objectives are clear and realistic during the remainder of the pilot, and that there is a shared vision and understanding of what success looks like. Limitations of the accommodation pathway 9.52 The accommodation pathway is an integral and perhaps the most novel element of the Restart approach. See section 10 for an analysis of the benefits and challenges of the accommodation pathway. Funding uncertainty and sustainability 9.53 It was clear from several interviews that delays in securing an extension to the pilot impacted on delivery since there was some doubt about whether Restart would continue. This was due to the funding cycle which led to funding only being available one year at a time which deterred some from engaging fully in the programme, for example: ‘the idea of...eight months felt like it was a bit of a kind of token effort really, but then with the new funding, that has felt like a real opportunity now, where we’ve been able to really build on what we’ve already...started.’ Implementation lead, focus group. 70 ‘I think one of the challenges is how short a timeframe it is…and you’re asking areas to really buy in and to utilise these things and push it and …when you don’t know if you can trust a pilot to stick around, how much do we want to invest in that relationship and putting it within a referral pathway if it’s going to disappear in five months?’ Representative of central team. ‘I agree the sustainability stuff is huge in a project like this because I think it affects the buy-in from the boroughs as well in terms of income. What is the point getting on board with this new model when actually you’re going to be gone in a few months? and I think that’s some of the feedback that implementation leads had early on.’ Representative of partner organisation, focus group. 9.54 The short-term nature of the funding inevitably led to some cynicism about the project and doubt as to whether it was worthwhile for practitioners to invest their time and energy into it. Although the project has been extended to March 2023, this is insufficient time for a project of this complexity to become embedded in everyday ways of working and for frontline practitioners to develop confidence in its delivery and to test out the changes in practice they learn about in the training. It is also a short time to measure the impact on families since some of the measures which the intervention aims to influence are long-term in nature. 9.55 Similarly, the short-term nature of the funding has also meant that it has been harder to tell if the programme works, as one strategic lead told us; ‘do something for five years, you’ll probably really be able to see the benefit of it’. Strategic lead, interview. Recommendation 8 Sustainability: We recommend that, if possible, the funding period for Restart should be extended to a total of three years to allow time for it to bed in, for practitioners to develop confidence in applying these new skills and changes in practice and to test its longer-term impact. 71 Interagency and partnership working 9.56 Close partnership working is essential for Restart to work effectively. Although we identified this as a potential strength of the programme (see paragraph 8.19 and 8.20), there were also issues identified in the way in which agencies worked together. This was not just different agencies but also different parts of the same organisation which need to work closely together to deliver Restart. For example, there was seen to be a gap between Children’s and Adult Social Care, and between Social Care and Housing departments in some local authorities. Although ASC was seen to have a more limited role in relation to supporting victim-survivors (representative of partner organisation) than CSC, it was acknowledged that there could be closer collaboration which could benefit the victim-survivor. 9.57 Similarly, in one borough, communications between the different parts of the agency were said to be poor, with little routine contact or regular meetings between housing and CSC even though collaboration was essential to the success of the project. Although the Community of Practice meetings are a helpful way of sharing the learning, these are not intended to address local issues or help to forge local relationships which are needed for the scheme to bed in. Recommendation 9 Partnership arrangements: We recommend that where suitable, opportunities for intra-borough meetings are arranged locally so that practice can be shared at a local level. This will bring about better communication and shared learning between the partners and provide an opportunity to develop and deliver a shared vision. Changing systems 9.58 The Restart pilot is taking place in the context of a significant time of change in domestic abuse policy and services (the new Domestic Abuse Act 2021 was enacted during the course of the project) and greater efforts being made at national level to move away from victim-blaming approaches and to make perpetrators accountable for their actions29. We also noted a change in the language used in relation to domestic abuse which some interviewees cited, with practitioners at Sutton Council, for example, moving away from talking about perpetrators and victim-survivors to people who use abuse and people who experience abuse. 29 See Tackling Domestic Abuse Plan: HM Government, March 2022, in particular, Pursuing Perpetrators. 72 9.59 Several focus group attendees and interviewees told us that ‘I think what we’re working towards is thinking...this is a systems change model.’ Representative of partner organisation, focus group. ‘Like the success of the programme to me would be...a change, a visible change in the way that perpetrators and victims of domestic abuse are responded to, from statutory services, predominantly social care and housing...trickled through across the local authorities and into other statutory services as well...the training is available wider than just social care and, and housing…less of the of the onus on victim survivors, less of the language around failure to protect, more around failure to appropriately parent and more around accountability placed on perpetrators of the impact that their behaviour is having on their children and on their family rather than the impact that the perpetrator behaviour is having and what she can do to stop that because it’s not her responsibility.’ Representative of Strategic organisation. 9.60 If Restart is going to be embedded or rolled out, it was recognised by interviewees including senior managers that whole systems changes are needed to change the way that perpetrators are seen and engaged with by Councils and other services. This will inevitably take some time and is difficult to achieve given the pressures on busy frontline practitioners and the pressure from both funders and staff to see early results. Within the survey and focus groups, a number of participants flagged that it was ‘too early to tell’ what impact the programme had on either services or service users. ‘Sometimes I think we are I feel quite impatient on the project. And I want to see results immediately.’ Delivery partner, focus group. ‘I think probably our children's services needed some more intervention training, at the beginning of the promotion of Restart, and maybe it would have streamlined through the services better.’ Frontline practitioner, focus group. 9.61 Changing complex systems which have been in place for some considerable time is not going to be achieved in the short term and is undoubtedly complex. It requires time, patience, commitment and imagination and this can be difficult at a time when services are under pressure in terms of the fiscal environment and workforce pressures. Restart is attempting to change more than one system as well as professional practice established over many years and embedded in a legislative structure which may not be conducive to the changes underway. Many 73 interviewees acknowledged and indeed welcomed the systemic changes underway and were keen to play their part in making these happen. Recommendation 10 Changing systems: We recommend that any communications about Restart make it clear that it is aiming for long-term systemic change which can only be achieved through commitment at all levels within a local authority and other agencies. 74 10 Accommodation pathway Context 10.1 The accommodation pathway sets Restart apart from many other interventions for domestic abuse perpetrators. Our findings on this are therefore presented in a separate chapter to look at how the aims and objectives have been met so far and at some of the challenges in doing so. The purpose of the accommodation aspect of the Restart Project is to address the housing needs of families who are being supported through Restart. In particular, the purpose is to provide housing pathways and support for perpetrators to safely leave the family home, either temporarily or permanently, so that survivors of domestic abuse, including children, can have ‘space for action’, and safely remain within the family home. 10.2 An objective of the project is to test out the learning as far as offering accommodation is concerned, add to the evidence base and inform future practice. Housing panels meet in each local authority on a weekly basis to consider referrals within five days of an assessment that identifies a housing need. This is an integral component of the intervention and are important in reviewing the options from the family’s point of view. In the authorities which had had no candidates for the accommodation pathway at the time our initial interviews were held30, mock housing panels were used to simulate and test out what the issues would be in following the pathway and to build confidence in the staff who would implement it should the need arise. 10.3 Each local authority committed to providing two placements specifically for Restart clients. This can consist of short term immediate or emergency accommodation for up to four weeks, longer-term temporary accommodation or longer-term accommodation (between six months and two years). In addition, the intention was that short term placements would also be found for service users for up to four weeks with help being provided to finding longer term provision where needed. 10.4 The offer of accommodation is intended to keep the victim-survivor and family safe, enabling them to stay in their own homes by removing the person causing the harm, thereby minimising disruption to their lives. The accommodation offered may provide a short-term respite for the family, with the perpetrator being able to return at a later stage, or it may be a stepping-stone to his moving away more permanently. The accommodation aspect may also support families where the perpetrator may have already been removed from the family home or have 30 This has now changed – every participating council has now made at least one referral. 75 insecure housing which could impact on the safety and housing security of the survivor, and the perpetrator’s ability to engage with behaviour change and other offers of support.31 There is an obvious risk involved that is not present with traditional approaches such as refuges in that the perpetrator knows where the victim-survivor lives. 10.5 It is assumed that in instances where the victim-survivor would remain at high risk if she stayed in her own home, she would not be referred to Restart since this is not intended to be used in cases where there is a high risk of serious harm or homicide towards the victim-survivor. It is recognised, however, that not all victim-survivors will wish to remain in their homes, and this is determined on a case-by-case basis depending on their expressed needs and wishes. The (draft) Restart accommodation manual makes it clear that housing options provided to the perpetrator will also include housing support being provided to the survivor to maintain their home or to leave it if they would prefer to do so. One of the case studies provided illustrated that not all families will feel safe staying in their home when the perpetrator knows where they live. Victim-survivors who do wish to leave the family home (but do not wish to or are unable to seek refuge) could still be offered support through Restart. Take-up of the accommodation pathway 10.6 So far (until 1st July 2022), a total of 11 housing referrals have been made, of which two were undergoing assessment. Table 7 in paragraph 7.15 provides a breakdown by borough. It is therefore early days in terms of assessing the outcomes of the accommodation pathway. However, we sought views from people delivering Restart to see whether they felt this was beginning to make the difference or had potential to do so once it had been implemented fully. It was initially thought that around 30 perpetrators would seek hotel accommodation and around 10 would require support to access longer term accommodation32, equating to around 25% of participating families, but the intention was that this should be tested and reviewed through the pilot. 10.7 In practice, this has turned out to be around 14%. Some interviewees said that they and their colleagues found this approach difficult in that it appears on the surface to reward and even incentivise abusive behaviour and could be seen as colluding with those who perpetrate harm, as one frontline worker lead put it: ‘why would we reward someone for their bad behaviour?’ (Community of Practice meeting 17th May) We noted that take-up of the short-term accommodation 31 Taken from Restart manual. 32 See accommodation guidance page 3. 76 support was higher during its predecessor programme (which took place during the lockdowns due to the Covid-19 pandemic). The take-up during the Restart pilot has been fairly low so far, with considerable variation between local authorities. Some of those we interviewed had therefore had little direct experience of the new pathway. One Housing Needs Officer spoke of her disappointment that she hadn’t been allocated a case at the time she was interviewed but this may have changed since the time our interviews were conducted. 10.8 Several explanations were given for this being lower than the expected level which was in any case an estimate. There is £57,000 funding available for the emergency accommodation. The options for accommodation are more limited since the end of the lockdowns during the pandemic. It was recognised that the high rate of accommodation provision during the previous programme was due to the impact of the lockdowns, with large numbers of families under strain and needing to be separated very rapidly. This is no longer the case, which may account for the housing needs being lower than anticipated. The Restart Pilot Dashboard shows the housing needs of service users; the latest data suggests that this is 17%, the second highest need after Children, family or parenting at 22%33. 10.9 We have included two case studies provided at Appendix 2 to illustrate how this has been delivered. Benefits 10.10 Many interviewees, particularly those in CSC but also in housing, saw this as a significant step forward towards achieving victim-centred services. It was seen to have potential to minimise the disruption commonly experienced by those fleeing domestic abuse and having to seek alternative accommodation, including in refuges, frequently in other areas. We were told: ‘So, I think kind of reframing it in terms of the perpetrator being the one to move. I think it’s a bit of a step change really and although it’s come from the idea of assisting perpetrators move home, giving them housing assistance can be controversial – if that does kind of reframe the balance of sometimes the victim-survivor not being the one, that they can retain the home, then that’s probably a positive stat.’ Strategic lead, interview. ‘For us, it’s about focusing on enabling survivors to be able to stay safe within their own home and put the responsibility of leaving on perpetrators…recognising that perpetrators will find it difficult to be 33 See Restart Pilot National Dashboard Quarter 4. 77 responsible for their behaviours and address them when their basic housing needs aren’t met.’ Partner organisation, focus group. 10.11 The offer of temporary accommodation also provides a solution, at least on a temporary basis, for perpetrators whose shame of using harm may deter them from moving in with family and friends. As one of the service managers told us: ‘Dads who were sofa-surfing and wanting to return, we thought that there was more give in saying to them actually, we could offer you something outside if you don’t want to return if we can find you somewhere to live. And also, the shame probably in going to your friends or family or saying that you cannot return home because you’ve perpetrated abuse against your partner is quite shameful, I suppose.’ Frontline housing manager, focus group. 10.12 From the case studies provided to us, it was also clear that someone who perpetrates abuse can use his homelessness or potential homelessness as a form of coercive control over the victim-survivor, coercing her to let him stay in the family home. Providing accommodation, initially on a temporary basis, at the same time as the four-week intervention to address his behaviour, can help to overcome this. This is a highly significant potential advantage from the point of view of the victim-survivor and needs testing further. 10.13 Even where the accommodation pathway was not utilised, it was suggested that the fact that alternative accommodation might be available changed the nature of the conversation with families in a positive way. There was also evidence that it was starting to change practice and attitudes despite the complexities, for example: ‘And normally we wouldn’t do anything for him whatsoever. And why use up our priority resource when we struggle to house victims? So, we’ve taken a completely different view. We said, like these experts and specialists are saying to us, the victim-survivors will be both safer and happier and so will their child if he leaves the household and we have to facilitate that because he has the ability to do that under his own resources.’ Strategic Lead, Interview. 10.14 Although it was seen as early days, some interviewees did feel that the intervention was showing signs of early success, for example: ‘I think we have created a shift in the way people think about the housing needs of the family and the fact that we should be considering how the perpetrator can leave the family home. And I think even thinking about that 78 was an option at all is coming a long way. Because I think in many areas, they still have a very blinkered view that the survivor should be the one to leave. They should be the one to go into refuge.’ Representative of partner organisation, focus group. 10.15 Given that the intention has been to learn from the intervention as to what approaches are likely to work, this appears to have been successful: ‘Additionally, the next thing that I think is positive is that we're really doing the work of trying to figure out how you make that work in practice, and what exactly you can offer perpetrators within the remit of housing legislation and what local authorities can practically offer and what the impact of that is on the family, we are actually testing out what does and doesn't work. So, I feel like, even if we don't successfully house a lot of perpetrators because I don't even think that's the end goal, we only may house a few because there are only maybe a few that need it. I think we have gained a lot of learning about what does and doesn't work and why.’ Representative of partner organisation, focus group. 10.16 It was also seen as important by some interviewees not to present the accommodation option too early on in the intervention since, as one Case Manager pointed out to us, she didn’t want perpetrators to think that the main point of the engagement is to get accommodation. This could lead to them disengaging as soon as they had been rehoused. 10.17 Considerable effort has been made to look at the different options open to local authorities and how some of the potential regulatory stumbling blocks could be overcome. It was, however, noted that some of the difficulties were less challenging than anticipated, for example: ‘One thing that has been interesting to me is that I think we assumed that in the vast majority of cases, the tenancy would be in the perpetrator’s name or it would be a joint tenancy and it would be hard to extract the perpetrator from that situation. But actually, in a fair number of the cases either it was a social tenancy in the survivor’s name and the perpetrator had no legal rights to the property or it’s a PRS tenancy in survivor’s name and again, the perpetrator had no legal rights to the property.’ Representative of partner organisation, interview. 79 Challenges 10.18 One of the main challenges for the project is to change the mindset of housing practitioners who may baulk at placing people who use harm above people on the housing list who have been waiting for accommodation for some time. Although this is new, it was pointed out that this is not unprecedented in that ex-offenders, for example, are offered accommodation on leaving prison. In order to achieve the systemic change that is needed to deliver Restart, the housing and homelessness workforce need to understand the rationale behind this, recognising the intention to keep the victim-survivor and her children safe and to improve outcomes for the whole family. However, we found in several interviews that the traditional approach may be entrenched in some staff; for example: ‘There is a waiting list for privately-rented accommodation as well. So, for these people to be sort of leapfrogged over everyone else, there has been a little bit of disdain, until it’s been explained that that is to protect the victim.’ Frontline practitioner, focus group, and ‘The days of saying to perpetrators…you’re not homeless, just go back and stop beating your wife needs to change….They can’t only be only the victim can apply as homeless because that’s not fair on the victim. She should have options and choices but someone’s got to lead the way on that because it’s culture. It’s very deeply engrained culture within homeless services which isn’t homeless services’ fault.’ Strategic housing lead, interview. 10.19 In addition, it was also recognised that it would not be feasible to offer all perpetrators of domestic abuse alternative accommodation and that it would therefore be difficult to prioritise who should be eligible: ‘And if I’m going to stand up in front of all my equivalents across London, they’ll say, oh well, let’s just accept every single perpetrator. There will be so many to rehouse, they’re going to laugh me off the stage effectively. We’ve got to have strict parameters, especially in the London context…..Giving him his own Council tenancy is really unpalatable culturally, he’s the perpetrator, he’s in the wrong, why should we help him?...that’s culture and I think it’s understandable to some extent.’ Strategic housing lead, interview. It should be noted, however, that service users would not be placed in social tenancies in this pilot unless they qualify for this in their own right. 10.20 Explanations put forward by interviewees for the low rate of referrals include the culture change in allocating housing to perpetrators of abuse who might otherwise 80 have been expected to remain in their home or to be considered homeless if they left. Offering alternative accommodation, even on a temporary basis, to perpetrators of domestic abuse represents a considerable change in the mindset since it could be seen as a reward for ‘bad behaviour’. 10.21 It was pointed out to us by one local authority, Camden, that they have around 2500 single people on their housing register at any one time; rents are 40% or so above average rents in London. It would be difficult to offer accommodation (whether in the private rented sector or in the supported adult pathway) to many people given those pressures. This has not been a problem so far as few referrals have been made but it was anticipated that it might be difficult should too many referrals be made. Similarly, Westminster have made two tenancies available for the project – any demand above that level would be difficult to meet both financially and politically, particularly because the process for those who use abuse is simpler than for others facing homelessness: ‘We’ve essentially agreed that once a referral comes through, housing won’t come and apply any of those statutory tests that you would do around someone who’s maybe threatened with homelessness. And we will just kind of fast track them to our private rented sector team. And I think speaking honestly, if we were to commit large numbers, it could become quite politically difficult to justify why we were in essence prioritising perpetrators of domestic abuse for a service that other people wouldn’t get.’ Strategic housing lead, interview. 10.22 The lack of buy-in from senior and middle management was also considered to be a factor deterring housing officers from engaging with the project , particularly when housing stock is so low, and in the post-pandemic landscape, hotel options have changed significantly since the beginning of the pilot. One focus group attendee told us: ‘Obviously housing is a huge issue for everybody……we don’t have anything set aside [for this project].’ Frontline practitioner, focus group. As noted previously (see paragraph 10.3), each of the five boroughs had committed to providing up to two tenancies, although this can be difficult to find in practice given capacity constraints. 10.23 Interviewees emphasised that, whilst the accommodation pathway is integral to Restart, the intervention is far broader and this is just one element of it. It is seen as advantageous to have it as an option where needed: 81 ‘It’s great that it’s on offer, just the fact that it’s an option that you can talk to a family about immediately…that it’s just not on the adult survivor and children. So, I think there’s value in it, even when it’s not being used.’ Implementation lead, focus group. 10.24 It was also suggested by some focus group attendees that perpetrators’ expectations of housing may not match what is available to them; the majority would prefer to be offered social housing whereas this is difficult given the high demand and limited supply. They are therefore more likely to be offered temporary accommodation in the Private Rented Sector (PRS). This may have an impact on how the housing pathway is used: ‘What everyone wants [is] self-contained accommodation. And that’s what everyone thinks have the right to, so it’s just about managing those expectations early on.’ Frontline practitioner, focus group. ‘This is just a general housing problem, but people tend to have higher expectations about what we can do than what we can actually do. So, majority of our clients are assisted into privately rented accommodation, because social housing waiting times are so long.’ Frontline practitioner, focus group. The case studies illustrate how important it is to manage expectations at the point of referral. Understanding early on what the accommodation offer is likely to be for perpetrators is key to this. This has been recognised (for example, at the Strategic Steering Committee meeting held on 7th April 2022) – it should be clear that one of ‘the aim[s] is to change the relationship dynamic, not to give a short cut to solve a personal housing need34’. This does need to be made clearer to families by practitioners early on when the intervention is explained to them. 10.25 It is also necessary to establish relationships with local providers of accommodation in the Private Rented Sector at an early stage since their involvement and co-operation will frequently be needed to enable families to move. This was emphasised in two of the case studies: Case one: where the victim-survivor wanted to move location but this required the approval and support of the housing provider which was difficult to obtain; Case three: the family home was a secure tenancy in the perpetrator’s name – even though the perpetrator agreed to sign his tenancy over to the victim- 34 Restart Strategic Steering Committee meeting held on 7th April 2022. 82 survivor, it proved difficult for the accommodation panel to engage the private social housing provider to facilitate this. Due to this, rather than exploring the possibility that the tenancy could be directly signed over to the survivor, lock changes and additional security measures put in place, the victim/survivor has received the support of a family solicitor to seek a tenancy transfer. She has also been supported by the Partner Support Worker with emotional support, benefits applications, and support to gain a non-molestation order and to advocate for housing needs through the housing panels and direct advocacy. The perpetrator has been supported to prepare to access longer term accommodation, and as he is in full-time employment, he was able to access longer-term accommodation. The legal complexities and restrictions were also cited as an issue by some interviewees as a further limitation on councils using the accommodation pathway. We were told that the homelessness and housing legislation have been in place for a long time and emphasise the priority of accommodating victim- survivors, making it more difficult to adapt to rehousing the person who is causing abuse although there are ways of doing so. 10.26 The complexity of removing those who use harm from the family home was cited, including the legal issues resulting from the homelessness legislation. This was seen as a possible impediment to the Councils’ ability to remove them, particularly in instances where they were unwilling to leave and worried about the risk of losing their tenancies if they do so. There is also a risk that perpetrators who take up the accommodation pathway may be seen as intentionally homeless and therefore considered ineligible for rehousing. Since perpetrators may not wish to identify themselves as such, it was pointed out to us that one of the major blockages is that someone who is using abuse may try to regain access to their home and avoid admitting to the abuse, by saying that his partner has ‘kicked him out’. In addition, lawyers may advise perpetrators not to give up their tenancy. In one case that was discussed at a housing workshop meeting and presented in a case study (see case study 3 in Appendix 2), the perpetrator was willing to give up his tenancy leaving it for his partner and children, but was told that the only legal route to do this was for the victim-survivor to obtain a non-molestation order, which he found frustrating. 10.27 Housing legislation around priority need (which is one of the tests a homeless person needs to pass in determining their entitlement to help with housing) was also raised by a number of focus group participants as a complex aspect of the intervention, some citing the limitations of the under-35 cap on housing benefit which means that perpetrators under 35 offered PRS accommodation will only be offered a room within shared accommodation. This makes this option unpopular 83 and problematic, since perpetrators under 35 would be expected to share accommodation which may be unsuitable for fathers wishing to maintain contact with their children. The Restart Accommodation Process guidance points to the importance of avoiding placing a burden on the survivor to facilitate contact in the family home35. This was raised in one of the case studies as a wider issue that would need to be addressed. However, since the under-35 cap for victim-survivors of domestic abuse has only recently been removed, it is unlikely that this will be changed for perpetrators of domestic abuse in the near future. Several focus group attendees and survey respondents also regarded this as a structural barrier to change: ‘There is a need to bypass statutory homelessness assessments which can cause delays.’ Survey respondent 10.28 One further issue raised was the difficulty in providing support for those with no recourse to public funds (due to their immigration status) using the accommodation pathway. Although they are able to use Restart’s emergency housing fund, they will have fewer options available to them in the longer term since they will not be entitled to social housing or benefits. There is no obvious solution to this issue but it is something that the central team are well aware of. One Case Manager pointed out that this was a struggle. 10.29 The next stage of the project should examine in more depth some of the barriers and enablers for people using the housing option and what difference this makes to the family’s outcomes. The impact of any external factors such as the influx of refugees from Ukraine and the cost-of-living crisis should also be looked at with a view to seeing how they impact on the availability and affordability of accommodation options within the programme. 10.30 In terms of some of the structural challenges the project has come across, the central team may wish to consider linking the findings of this work with their own policy and campaigns focused activities. This would include work to remove legislative barriers that may prevent service users from taking up the accommodation offer (no recourse to public funds, priority housing, and shared accommodation for under-35s, for example). Recommendation 11 Accommodation pathway: We recommend that: • the limitations of the pathway should be made clear from the outset to avoid raising expectations or unrealistic expectations for families including the fact 35 Restart Accommodation Support Process, page 2 (Child contact and care) [now superseded, this does not appear in the refreshed guidance]. 84 that Restart does not offer a fast track to social housing for perpetrators of domestic abuse. This should be set out clearly in literature on the project including training materials; • further work is carried out to assess some of the practical issues involved in implementing the housing pathway such as (1) enabling those who use abuse to maintain contact with their children and (2) enabling those who have no recourse to public funds to be further analysed and reviewed at the Strategic Steering Committee. • it is necessary to establish, build and maintain relationships at an early stage with the largest local providers of privately registered social housing since these providers may be housing the victim-survivors and perpetrators being supported and may make housing transitions easier; • the next stage of the project should examine some of the barriers and enablers for those using the housing pathway and what difference this makes to the family’s outcomes. The finding that around 15% of families may use the housing pathway should be built into funding assumptions for the project. 85 11 Conditions needed for programme to bed in. 11.1 From our interviews, we have identified some of the conditions that would seem to be needed in each local authority for the programme to bed in successfully. We hope that this it will be useful in any consideration of rolling out Restart to other areas. As one strategic lead said about the success of implementation in his own borough: ‘I suppose the conditions were right.’ Strategic lead, interview. Our initial findings based on what interviewees have told us include the following conditions for programme success: ▪ Strategic support and leadership for people-led approaches to public service and innovative practice. Support and leadership within the LA are essential for success. Implementing a pilot across five different local authorities with very different priorities, political climates and facing different issues is in itself complex. Securing commitment from senior leadership in every local authority is therefore key to delivering the pilot. ▪ Time to have strategic conversations in the set-up stage including with stakeholders to secure buy-in, secure commitment, build a partnership and formulate a vision and theory of change together. The local authorities that did this were likely to have higher referral rates and higher buy-in from staff and a sense of a shared vision as to what Restart was about. These local authorities also tended to prioritise attendance at our interviews or focus groups. ▪ Established relationships with the major providers of Private Rented Sector accommodation in each borough early on should be a priority in rolling out Restart to other local authorities. ▪ Participatory and consultation spaces for groups who have experienced the social work system to feed in their views to the local authority and influence commissioning and strategic development. Areas that recognised the trauma that the current system can cause for families were more likely to have high level buy-in and support from frontline staff. ▪ Commitment from senior and middle management to exploring what can be achieved through Restart and providing support for their staff, ensuring that they have the capacity to deliver. Building relationships within the local authority and with partner agencies is key to successful implementation. It was suggested by focus group participants that in local authorities where there was pressure to take on more cases and close cases rapidly due to workforce pressures, they were less likely to promote Restart or encourage referrals. There needs to be buy-in to the delivery of Restart at all levels and a ‘golden thread’ in terms of the commitment and the messaging from senior leads down to all front-line practitioners. 86 ▪ Space and time for social work practitioners to reflect on practice and long- term engagement with families. This was reflected in the ability of case workers to see the benefit of longer-term engagement with families, as well as a willingness to work with perpetrators directly, and to take a relationship-based approach to their work. ▪ Having a range of interventions for domestic abuse available covering the full range of risk to include interventions for those who use abuse as an integral part of the Violence Against Women and Girls strategy. In those areas where services were more widely available, there seemed to be more of a commitment to achieving the systems changes needed for Restart to succeed. ▪ The Safe and Together implementation leads are an essential part of the programme and help to embed Restart in the five local authorities. Similarly, having champions to promote the use of Restart36 has proved helpful where take-up has been low. Recommendation 12 Conditions for scaling up Restart: We recommend that the conditions above and others which the central team could readily identify are mapped and scored in each area which is considering implementing Restart to show the state of readiness for Restart. This could follow the model of a maturity matrix.37 36 There are, for example, five or six in Havering. 37 See, for example, an example of the Early Intervention Foundation’s maturity matrix https://www.eif.org.uk/resource/eif-maturity-matrix-speech-language-communication-early-years 87 12 Costs and benefits of the Restart programme 12.1 In this section, we consider the costs of the Restart programme, before assessing the impact of the programme on outcomes, the benefits of that impact, and the ratio between costs and benefits. It is important to highlight that it is not possible to fully conclude value for money at this early stage in the development of the new model, when business as usual delivery has not yet been established. More time is needed for data points that will inform a clearer picture for value for money to emerge. Therefore, all findings and messages here are caveated as emerging learnings rather than conclusions. Our interim assessment has focused on the service delivery aspect of the programme, rather than the provision of training. The reason for this is that, although the costs of the training can be calculated, it is difficult to measure the benefits accruing as a result of the training at this stage. 12.2 Our assessment relates to the financial year April 2021 to March 2022 to match our available financial data. This covers a pre-implementation phase from August 2021 to the end of October 2021, prior to implementation starting from November 202138. Costs of the Restart programme 12.3 We draw on data on costs during the financial year 2021/22 provided for us by the accounting team at SafeLives. Costs have been accrued by the Drive Partnership (£165,000), Respect (£206,000), Cranstoun (£208,000) and DAHA (£37,000), amounting to a total of £616,000. 12.4 These costs include both training and client-based work. We have derived an indicative assessment of unit costs for client-based work by: • Clarifying which aspects of costs relate to training, which to client work, and which to preliminary phase work – we have excluded costs on preliminary activities from the analysis; • Assessing the relative proportions of direct costs for training (46%) versus client work (54%) over the year 2021/22, and then allocating overheads (costs of the Drive Partnership during the implementation phase) in line with these proportions; 38 This is a slight simplification as the case opening date for the first service user for the implementation phase was 28th October 88 • We then divide costs for client work of £199,460 by 44 service users, to obtain a unit cost of £4,533 in relation to service users. Note that because a variety of activities were undertaken by Restart staff, our assessment of unit costs on training will not be available until the extension phase of the evaluation. 12.5 Table 14 below summarises our calculations. Note that we have focused on the specific service delivery functions of the Restart programme, and so we have excluded costs relating to paying for places on a DVPP, and to evaluation costs. We have further made a distinction between one-off pre-implementation costs, and ongoing operational and overheads costs. Lastly, it should be noted that the estimate of unit costs of service provision if DVPP is included is based on an altered allocation of overheads (59%: 41%) between service provision and training after including DVPP spend among service provision direct costs. Table 14: Indicative assessment of costs of the Restart programme Training and engagement Service delivery Total Preliminary phase (Aug 21 to Oct 21) - set-up costs Respect, Cranstoun, and DAHA £26,135 £76,100 £102,232 Drive Partnership £12,145 Total preliminary £114,375 Implementation (Nov 21 to Mar 22) Respect, Cranstoun, and DAHA £179,840 £168,940 £348,780 Drive Partnership (excluding DVPP and evaluation) £56,175 Implementation costs including estimated share of overheads £205,490 £199,460 £404,950 Unit costs (excluding preliminary phase) £4,535 Unit costs (including DVPP) £6,640 89 Benefit of the Restart programme 12.6 The Restart programme has both short-term and longer-term consequences for families. When further data is available, we will look to review the effects of Restart and follow-up programmes versus the counterfactual of some undertaking follow-up programmes alone, and some receiving additional support prior to DVPP. 12.7 Such impact data will then be assessed against three types of benefits that occur through the reduction of domestic abuse: • The wellbeing of the victim-survivor is improved due to a reduction or cessation in the level of abuse • A lessening of the loss of earnings and productivity of the victim-survivor as a consequence of reduced trauma • Reductions in public sector costs (in relation to organisations varying from the A&E services of the NHS, to the criminal justice service). 12.8 In reviewing the potential benefits from effective activity to reduce domestic abuse, we have updated analysis from the 2019 Home Office study (Oliver, R. et al)’39. It examines the economic and social costs of domestic abuse and so provides a starting-point for consideration of the issue. 12.9 Limitations of that study include the exclusion of costs relating to children victim- survivors – though we have looked to make an adjustment for this gap (in relation to Children Services costs, though medium-term and long-term effects for children in terms of wellbeing and educational and employment prospects are not included) - as well as the substantial flaw that there is not ‘full consideration of the number of injuries incurred by victims during their abuse’. Together these imply that the assessments should be treated as indicative, particularly since although the Home Office study40 notes (p9) that the estimated costs ‘relate to the number of victims within a particular period (financial year ending 31 March 2017)’, it makes the caveat that effects over a wider time period are included – ‘… Even if the duration of harm exceeds or predates this period, the full costs (including costs outside of this period) to people who were victims during this period are included.’ 39 Home Office, The economic and social costs of domestic abuse, Research Report 107, see https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/918897/ho rr107.pdf 40 Ibid. 90 12.10 We will look to review the figures with a view to providing a more robust assessment in the extension phase of the evaluation. 12.11 Indicative estimates of wellbeing loss, economic productivity loss and increased public sector costs per victim-survivor based on that Home Office study are shown below in Table 15, updated to 2021/22 prices and also including Children’s Services costs. Table 15: Estimate of wellbeing loss, economic productivity loss and increased public sector costs (2021/22) per victim-survivor Wellbeing, productivity and public sector cost per victim-survivor of domestic abuse Wellbeing Physical and emotional harm £27,070 Economic productivity Lost output £8,070 Public sector Health services £1,335 Misc victim services £100 Housing services £315 Police costs £720 Criminal legal £190 Civil legal £80 Other £10 Total £37,890 12.12 As noted earlier, we have looked in addition to make an indicative allowance for costs of children’s services. We make the indicative calculation that domestic abuse leads to requirements for £398m spend on Children in Need and £457m spend on Children Looked After. • In relation to Children Looked After, costs are based on £5.70 billion annual spend on Children Looked After (from 2020/21 local authority revenue out- turns), increased by 3% to 2021/22 prices as per the ONS GDP deflator41 scaled 41 See www.ons.gov.uk/economy/grossdomesticproductgdp/timeseries/l8gg/qna 91 by 35% (new entrants as a proportion of Children Looked After)42, scaled further by 22% (the proportion of cases relating to family stress and dysfunction). • In relation to Children in Need, our indicative estimate is based on 598,000 referrals to Children in Need during the year ending March 202143, scaled by 19% of cited factors behind Children in Need requirements relating to domestic violence, multiplied by £2,980 case study cost per child in need. 12.13 As there were some 590,000 families affected by domestic abuse in 2020 (according to ONS data estimating 8.08 million families with dependent children in England44, and a 7.3% proportion of domestic violence prevalence for females45), this suggests a cost per family of the order of £1,340 (based on £792m costs divided by 590,000) - though it should be noted that this may well be an under- estimate requiring refinement, since the prevalence of domestic violence is higher for young age groups, and younger women are more likely to have dependent children. 12.14 Together, the Home Office figures plus indicative estimates relating to Children’s Services represent a unit wellbeing and productivity loss and increased public sector cost of the order of £39,230 for domestic abuse. However, this figure would be higher if the cohort is more harmful than the average case of domestic abuse in England and Wales; we will consider further analysis on this point in the second phase of this evaluation. 12.15 The key issue is the effect of the Restart programme on that figure of £39,230, taking into account the extent to which the programme can ameliorate the extent of domestic abuse compared to a counterfactual in which no intervention took place. We discuss a rough-and-ready “break-even” point at which benefits exceed the cost of Restart below. 42 https://explore-education-statistics.service.gov.uk/find-statistics/children-looked-after-in-england- including-adoptions/2021 43 https://explore-education-statistics.service.gov.uk/find-statistics/characteristics-of-children-in- need/2021 44 https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/families/datasets/families andhouseholdsfamiliesandhouseholds 45 https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/articles/domesticabusevictimcharact eristicsenglandandwales/yearendingmarch2020 92 Break-even point for Cost benefit ratio of the Restart programme 12.16 We have estimated the cost per case of Restart as some £4,535. This cost relates to activities based in London, and so in order to present a fair comparison against benefits figures that are based on national figures, we either need to scale up the benefits figures to London values (for instance by adjusting the lost output figures to London wages levels) or scale down the costs to national levels. 12.17 We have chosen the latter for simplicity, and so therefore we scale down the Restart costs by an illustrative 15% - to £3,855 per case - to present an estimate of the Restart pilot’s costs if undertaken outside London. 12.18 It then follows that the break-even point at which the reduction in wellbeing, productivity and public sector costs outweighs the costs of Restart is of the order of 9.8% (since £39,230 * 9.8% = £3,855). In other words, if the Restart programme reduces the of adverse effects of its service users by 9.8%, it has achieved a social value that is equal to its intervention costs. 12.19 It also follows from the above analysis that the break-even point at which the reduction in public sector costs outweighs the costs of Restart is of the order of 94% (since £4,085 * 94% = £3,855). This is a high bar, and perhaps reflects the relatively high unit costs of Restart in terms of its four-week programme for service users – almost three times the average cost of a DVPP programme of £1,59046 at 21/22 prices. This is due to the additional and multi-systemic support (including one-to-one support) provided as part of Restart which are not included in DVPPs which are frequently delivered in groups. 12.20 A key question for the extension of the evaluation is the extent to which domestic abuse is reduced in practice when the wider effects on follow-up programmes and consequent behaviour are taken into account. 12.21 It should be noted that the benefits and costs discussion so far has been restricted to service delivery. A further issue to be considered in the extension of the evaluation is undertaking an assessment of the benefits of the training programme on public sector worker behaviour and consequent outcomes for victim-survivors. 46 Based on estimate of cost of £1,500 in the Safe Lives report "A Safe Fund: costing domestic abuse provision for the whole family" (p15), scaled up by 6.1% to reflect the GDP deflator between 2019 and 2021/22 93 13 Conclusion 13.1 We found that there was much to commend Restart and its potential to provide a holistic, victim-centred approach to families who have experienced domestic abuse. We saw signs of genuine commitment to systemic change in the way that local authorities deal with these families and a clear opportunity for more collaborative partnership working between the different agencies. There was strong support for the testing of an innovative approach to enabling families to stay in their own home with those who use abuse encouraged to commit to longer-term behavioural change. We hope that further work will enable the impact to be measured including the impact on those who do follow the accommodation pathway (see Future work and recommendation 12). 13.2 Although Restart has been running in five London boroughs for nearly a year, it is early days to reach a conclusion about its actual impact and whether it has met its aims and objectives. Some staff have still had little direct experience of the intervention and, at the time of speaking to us, were unable to give feedback about their views on its likely impact. Measuring the long-term impact of the intervention has also not been possible within this timeframe; nevertheless, interviewees were generally positive about what has been achieved. 13.3 Our findings are limited in terms of what we know about the impact of Restart on trauma, risk and harm as we were unable to hear directly from victim-survivors or children and young people, and had feedback from only two service users; we therefore did not have a robust means of measuring outcomes. Whilst this limited our findings to some extent, our discussions with 25 participating practitioners and stakeholders provided us with considerable insight into their views about Restart, its potential impact and their perspective on the benefits and challenges over the last year. We have also gained valuable insight into the multiple and complex demands on practitioners’ time, and identified some of the gaps in the intervention (particularly around securing the voice of service users in assessing outcomes), but also the extent of the boroughs’ commitment to implementing the project. 13.4 One advantage of the Restart approach that was emphasised in our field work is that it is seen as less adversarial and as less traumatic for victim-survivors than current practice. It therefore has the potential to lead to improved relationships with CSC staff since victim-survivors feel less threatened and afraid of their children being removed. We saw encouraging signs that the intervention has the potential to stop the ‘revolving door’ associated with domestic abuse, with the 94 same families reappearing on caseloads and the same perpetrators moving on to use harm on other victims. 13.5 Safe and Together training and implementation has been well-received and helped to win the hearts and minds of practitioners, with signs of a resulting shift in culture and practice in social work responses and recognition of the need for systemic change. Whilst it was seen as comprehensive and of high quality, the downside is that it is inevitably time-consuming and a significant commitment for busy staff. This is reflected in the low completion rate. The potential for this being rolled out to other professional groups was welcomed by many interviewees. 13.6 The housing pathway was seen as an important element of the intervention, even where it is not actually followed. However, it is important that this is not seen as an end in itself, particularly since around 15% of service users have pursued this option (in line with the overall 15% in whom housing was identified as a need). Achieving behavioural change and reducing risk for the family in this way was seen as a far more important – and arguably achievable – end-point. Nevertheless, enabling families to stay in their homes where it is safe for them to do so would be an enormous step forward and far less disruptive and traumatic. As one senior manager put it: ‘And if this programme helps even a handful of victim-survivors to stay in the home that they’re settled in whereas they may have felt that they needed to be the ones to leave if it hadn’t been for this programme’s intervention, then I think this programme has been a success.’ Strategic lead, interview. 13.7 An equally important aim which we believe has been partially achieved was to strengthen the evidence base and learn from the project to see if the model presented a workable solution and was scalable. Much has been learned from the project in the first year and we have spoken to many highly committed practitioners who have achieved a great deal, sometimes in challenging circumstances given the pressures facing local authority staff. 13.8 There was a strong consensus amongst interviewees that Restart has the potential to keep families safe and to improve outcomes for families as well as to start the process of behavioural change in people who use abuse. There were many examples and case studies of those using abuse agreeing to participate in longer and effective evidence-based behavioural change DVPPs (such as ‘Men & Masculinities’) and this is undoubtedly critical to the success of the programme: 95 ‘I think it's a great springboard to the longer-term group Intervention programme, and that's where … a lot of work is done.’ Partner organisation, focus group. We found that 15 people who use harm have been referred on to a group programme (20% of the overall number referred). This is in addition to the number who received one-to-one support following the four-week intervention. The next stage of the work (see 13.2) will look at this in more depth. 13.9 However, whilst the programme is beginning to bed in, mobilisation has taken a long time despite the earlier programme which served as a firm foundation. The low rate of referrals in some boroughs is thought to be a reflection of this and, arguably, a lack of commitment at middle management level. This may help to explain the low rates of return for the survey tool – if staff were enthusiastic about participating, they may be more likely to encourage victim-survivors to complete the survey. However, we acknowledge that external factors may also have influenced the ability of boroughs to set up and implement Restart, including the impact of Covid-19 on staff, the cost-of-living crisis and the existing pressure on housing services exacerbated by the need to accommodate refugees. 13.10 The success of the programme is difficult to measure particularly when the voice of the victim-survivor and their families is missing from the debate and when long term outcomes are unknown. Success for the programme, however, appears to mean different things to different boroughs and to different stakeholders. For some, success looks like staff knowing about the programme and putting the Safe and Together training into practice with referral numbers at this stage not being a criterion of success; for others, it means high referral rates and uptake of the accommodation pathway. Echoed throughout our findings was the sense that success at this stage looks more like ‘a good understanding of what Restart is and what it’s trying to achieve and who the different workers are.’ Representative of partner agency, focus group. 13.11 From the funders’ point of view, they hope to see leadership and support for the programme at a senior level so that it is taken up across the authority with a view to achieving longer term change. They also wish to establish an evidence base to show what works in taking a fundamentally different approach to families experiencing domestic abuse and to help decide whether this is a scalable intervention: ‘I see the role [of Restart] is more about encouraging that innovation and adding to the evidence base….you want to galvanise others, to convince them that this is the practice that they should adopt, and that they should 96 then invest in their workforce to influence outcomes…..I think its role is around innovating, allowing that room for good practice to be developed and evidence-based practice and then put it out there and say…we think this is important, this meets our strategic priorities.’ Strategic organisation, interview. 13.12 Others commented that success is: ‘All internal partners, you know, the IDVA, the services, are all working collaboratively for the best outcomes.’ Frontline practitioner, focus group. ‘Our successes would be the practitioners actually understanding that there’s something else on offer something else they can actually refer to access support from, and the word is starting to spread.’ Frontline practitioner, focus group. Although the majority of interviewees were committed to the project, many had concerns about implementation and the complexity involved. Some frontline practitioners expressed their lack of understanding about Restart and how its components fit together, as well as the roles and responsibilities of those involved, with some misunderstanding the aims and objectives exemplified by a desire from social workers to close the case following referral. Restart can only be delivered successfully if a team approach is taken with everyone understanding the objectives for an individual family clearly and agreeing how they are best supported. The complexity (or 'multi-layered') nature of the intervention has undoubtedly acted as a barrier and been a source of some frustration for those working on it who inevitably spend their time explaining its purpose to practitioners. Nevertheless, there was some evidence of a team approach in some boroughs with the project staff working closely with staffing in housing and CSC. 13.13 As with many innovative projects of this kind, sustainability is an issue. Although the funding period has been extended, the long-term systemic change which is the goal here takes time to achieve, particularly given the large number of staff involved in delivery – though perhaps not on a regular basis – and a high turnover. A three-year funding commitment would allow time for proper testing and measurement of long-term outcomes. 13.14 We found a potential gap between the 'promise' and delivery – the project has been presented as a four-week programme focusing on early intervention and accommodation. It was not always appreciated that the four-week programme is intended to be the start of a longer process and should not be seen as the entirety of the Restart programme, or the Safe & Together approach. However, 97 practitioners repeatedly told us that four weeks is not enough time to do the work, and that they did not perceive it to be early intervention given how many victim- survivors present with a moderate or high abuse score. This may have an impact on the way in which practitioners perceive the project, and may prevent them from engaging fully in the programme if they feel it is unlikely to work in the timeframes set. In reality, there is considerable flexibility around the four weeks with each case considered on its merits and important longer-term work initiated during the four-week period including referral to a DVPP. 13.15 Lastly, given the tight financial constraints faced by the public sector, it will be important to assess the extent to which 'delivery’ and impact on outcomes does exist, particularly since the Restart programme (in respect to support for service users) is expensive compared to funding a DVPP without the additional components included in Restart, and quantitative assessments in respect of the value of training are not currently available. 98 14 Future work 14.1 The extension of funding for the Restart project to March 2023 provides an opportunity for further work. This should enable a strengthened quantitative analysis and additional qualitative analysis. 14.2 The additional quantitative analysis will include further work on the outcomes, the identification of a suitable counterfactual and a refinement of our calculations of unit costs. More robust statistics on impact should be feasible based on data on victim-survivors and service users who took part in the four-week intervention early on. The identification of a suitable counterfactual will enable us to compare the outcomes of Restart with ‘business as usual’. A better evaluation of the impact of the intervention would also be feasible based on a follow-up of the cases referred during the early part of the project to assess whether the longer-term impact of Restart for families have been achieved. Further work will be done to examine the numbers of people who use harm who go on to participate in a longer-term intervention. 14.3 It would be helpful to collect more information on the use of the accommodation pathway to address some of the specific outcomes on accommodation. This has not been feasible so far given that the numbers who have followed the pathway so far have been low. The intention would be to find out whether there is a need for a perpetrator accommodation option and to examine the barriers to the current re- housing options. 14.4 Since the survey instrument did not prove to be a successful way of eliciting the views of victim-survivors, those who use abuse or children and young people, further qualitative work should be conducted to ensure that their voice is heard. A good starting-point would be to work with staff to explore how these views would best be obtained and fed into the project. Ways should be found of building such feedback into the ongoing data collection for the project. Recommendation 13 Further evaluation We recommend that a further evaluation is carried out to improve the quantitative and qualitative findings from this project and to inform future work in this area. 99 Appendix 1: List of interviewees and focus groups Deidre Cartwright: Senior Housing Manager, Standing Together (6th July 2022) Ross Harvey, Housing Department, Westminster City Council (4th April 2022) Sade Howard, Violence Prevention Practitioner, Cranstoun (27th June 2022) Shaun Flook, Head of Housing Needs, London Borough of Camden (29th March) Roisin Madden, Director of Children’s Social Care, London Borough of Croydon (16th May) Martin Pratt, Deputy Chief Executive of London Borough of Camden (4th April) Focus Group 1 (25th April): Hannah Candee: Restart Programme Lead Amy Hewitt: Practice Adviser, Drive Partnership Frankie Mgbadike: IDVA, Advance, based at Westminster City Council Rachael Reynolds: Safe & Together Programme Manager Focus Group 2 (25th April): Caroline Edgerton: Housing Needs Officer, Croydon Kakoli Kumar: Service Manager, Assessment Team, Westminster Lexi Alafouzo: Strategic Housing Lead for the Restart Pilot, Standing Together Deidre Cartwright: Senior Housing Manager, Standing Together Focus Group 3 (26th April): Jonathan Lung: Service Leader, Children & Families, Croydon Diane Leavers: Specialist Domestic Abuse Support Worker, Sutton Sarah Strang: Community Safety Officer and MARAC Co-ordinator, Havering Danielle Hemming: Team Manager – Havering Adolescent Safeguarding Team, Havering Focus Group 4 (5th May): Chetana KL Brar-Mander: Service Manager, Cranstoun Sabrina Karim: Accommodation Practitioner, Cranstoun Asher Ansell: Advice and Prevention Team Leader, Encompass Sade Howard: Violence Prevention Practitioner, Cranstoun Safe and Together Implementation Leads: Respect (23rd May): Raanaa Javid Christina Tomprou Rupert Bagenal MOPAC Cathy Hickey: Programme Officer – Violence Against Women and Girls Lisa LeMasson: Programme Manager – Violence Against Women and Girls 100 Appendix 2: Case studies Case Study 1: Mike (42) and Edie (39) Referral information: Mike (42) was referred to Restart by his social worker due to concerns about his use of abuse against his wife, Edie (39). The couple share two daughters, (15 and 12) and a son (2), although it has since been confirmed that Mike is not the father to their son (2). Mike’s domestic abuse towards Edie was raised with professionals, when the eldest daughter saw her school nurse due to self-harming and a referral was made to CAMHS. The daughter disclosed the situation at home and a referral was also made to children’s social care. The younger daughter is also vocalising a desire to self-harm. Since learning that their son is not biologically Mike’s, Edie has left the child in the care of her family in Uganda and returned to the UK and the family home. Both children at home have clearly stated that they would like to remain at home with their father, and do not wish to have contact with their mother. Despite high levels of anger towards his wife, Mike would like the marriage to resume. Edie would like to separate, but the children to remain with her and once in place, have her son returned to her care. During assessment, Mike detailed his use of abuse towards Edie, including controlling behaviour, verbal abuse and threats. Mike felt justified in this due to Edie’s infidelity and subsequently telling him that he was the father of their son. Mike believed that the impact on the children and their mental health was due to their mother’s behaviour. However, despite clear denial, minimisation and partner blaming, Mike does not want the marriage to end and would like Edie to return home and her son to return. Support: Mike did six sessions on the Men & Masculinities group, delivered by Cranstoun. Whilst engaging with group sessions, Edie left the home and was accommodated in social housing. Mike was not able to take ownership of the domestic abuse throughout these sessions and once Edie left, he stopped coming to group as he believed he was no longer a perpetrator. As CSC had closed the case, the case manager was concerned about the situation and offered Mike one-to-one sessions for safety reasons and to do some post-separation work. The case manager explains that Mike was ‘stuck in the story’ of how Edie had an affair and a child which was not his. Fortunately, more recently, after around eight one-to-one sessions since he left the group, Mike is able to understand that his behaviour is abusive and is making very small changes. The case manager believes this change is motivated by the fact that Edie left him and he wishes for her to come back. The Case Manager has held two consultations meeting with the social worker and updated the social worker throughout the intervention. This case has highlighted concerns with the closure process within CSC. The Case Manager articulated these concerns, particularly around potential collusion with the father by the social worker. Furthermore, when the Social Worker explained that the case was closing due to no further safeguarding concerns, the Case Manager escalated our concerns around closure at this point as both daughters were still on the waiting list for CAMHS support, refusing to see their mother and at risk of being further 101 manipulated by their father. The case was handed over to the Cranstoun service manager to liaise with CSC. Lessons: Edie is engaging well with the PSW and is also engaging with the Accommodation Support Worker at Cranstoun. Concerns were raised when she initially left the property to be rehoused but then returned from Uganda without her son, refusing to leave the family home without her two daughters. Edie is engaging with her support plan and attending regular one- to-one meetings with her PSW. DASH risk assessments are being carried out to monitor risk and a referral to MARAC and for an IDVA is currently being completed. Case Study 2: James and Emily Referral Information: The Family Intervention Team referred the family to Restart in December 2021. The perpetrator (*James), and the victim-survivor (*Emily) have one child together, aged 1 years old. Emily has two older children (ages 6 and 10 years old) In October 2021, James attended the address of the older children’s father, as Emily and the children were staying there due to fears of violence from James if they returned home. James arrived at the property at 1am heavily under the influence of alcohol and created a disturbance on the doorstep, insisting he wanted to see his child. Emily attempted to ask him to leave as not to wake up the children, however James dragged her to the floor by pulling her hair and punched her in the face causing severe bruising and mild concussion. James has made threats to take his life when they have separated and attempted to get hit by a bus by standing in the middle of the road holding up traffic. On this occasion, Emily called the police, and James was arrested and remanded in custody and advised not to return to the family home. Before this incident, he was not permanently living in the family home, but had been homeless for over 2 years, sofa surfing between the Emily’s address and her brother’s address. During this period of homelessness, James has also been in the criminal justice system, spending time in prison, before returning to sofa surfing. Emily shared with the Partner Support Worker (PSW) at Cranstoun that James would use his homelessness as a form of coercion and control, often sitting outside her flat until she let him into the property. She felt blamed for his homeless status. Emily has a secure social tenancy in her name, which is not joint with the perpetrator. As James and Emily are not married or civil partners, James has no matrimonial rights to the family home. Accommodation Support: After receiving information about James’s housing needs and his continued homelessness, he was referred by the Cranstoun Case Manager, to the Accommodation Support Worker within the Cranstoun Team in April 2022. The delay in the referral was due to the Accommodation Support Worker coming into post in April 2022. 102 The Accommodation Support Worker (ASW) conducted an initial accommodation needs assessment with James to ascertain his current needs and circumstances (including housing, employment, finance, benefits, additional needs and vulnerabilities, access to behaviour management support), and then based on this assessment, the ASW made recommendations to the Housing Panel for immediate and longer-term housing solutions. Through this initial assessment, the ASW ascertained that in addition to the need for housing, James needed ongoing support with budgeting and managing finances to sustain a tenancy and contribute towards the children’s support. Ongoing housing panels were arranged for the family Social Worker, Accommodation Support Practitioner, Case Manager Partner Support worker (PSW), and LA Housing Operational Lead, where professionals come together to make joint decisions regarding how they will address the family’s housing needs. Following the initial assessment and housing panel discussions, the panel agreed that James would be offered immediate hotel accommodation funded by Restart. During this four-week period, James was also receiving support from the Case Manager to address his abusive behaviour and potentially engage in longer-term behaviour management programme. He was also receiving support from the ASW to manage budget and finance, to access universal credit and housing benefit. After the 4-week period in Restart hotel accommodation, the borough placed James in four weeks of temporary accommodation, while sourcing longer term accommodation. The borough has committed to providing James with access to PRS accommodation, including first month’s rent and deposit grant (to be repaid by James in small increments), and rent at local housing allowance rates, on a two-year contract. James remains in temporary accommodation provided by the borough Council until PRS accommodation can be sourced. In the meantime, he has moved on from the four-week assessment support through Restart and joined the Transform Behaviour Change Programme. In addition to the actions taken to support James to access alternative accommodation, Emily has sought a Non-Molestation Order against James, and contact has been suspended between James and the children. Emily has disclosed that even with the NMO in place (which ends in July), she does not feel safe to remain in the family home and wishes to access alternative accommodation. During this time, she and the children have been staying with a friend. The Partner Support Worker has provided support to Emily alongside the Case Manager support offered to James. The PSW has conducted an initial assessment and provided emotional support and housing advice, feeding into the housing panel discussions. Emily confirmed that although she has a NMO in place, she still does not feel safe to return to her home with the children and wishes to access alternative accommodation, potentially through a management transfer. James has also engaged well with his Case Manager and completed a number of one-to-one sessions both in person and over the phone due to work commitments. These has included: • Safety planning/Time-out 103 • Physical abuse and the impact on victims and children • Emotional abuse and the impact of victims and children James has also received support around: • Budgeting and finances • Mental health • Physical health • Child contact. Lessons: • This case also highlights the need to engage early on the social housing provider. In this instance, the victim-survivor has a social housing tenancy, and wants to move location. This could be supported through a management transfer, however, without the engagement and support of the housing provider, this is not possible. • It should not be assumed that by addressing the perpetrator’s housing needs, it will be possible to prevent a survivor and children from moving to improve their safety. Instead, all responses should be survivor-led and informed. By addressing James’ housing needs, it increased his ability to engage with his Case Manager and book onto longer term support programmes. It has also decreased the risk of James using his homeless status as form of coercion and control against Emily. By also providing Emily with support to access alternative accommodation, this only increases her opportunity for safety and housing security alongside these other actions. Case Study 3: Julie & Greg Referral information: This case was referred to the Restart Team by the Family Intervention Team in February 2022. The perpetrator, *Greg (aged 38) has been using long-term coercive and controlling behaviour, and emotional and economic abuse towards his ex- partner, Julie*, aged 45 years), and towards their 3 children aged 3 years, 4 years, and 14 years. The perpetrator was asked to leave the family home by the police, when they attended an incident at the family home on the 3rd of August 2021, due to a neighbour reporting loud shouting and were concerned. The police noted that Julie was clearly upset while talking to officers, but the rest of the family were calm. The incident was logged as a non-crime domestic. The Police interviewed Julie, they found that she initially stated that everything was okay however, on further discussion she became tearful and disclosed that she is struggling with two children who have autism and that her partner is financially controlling and becomes angry a lot. She stated that he isn't physically violent but shouts at her and the children. Julie informed the police of two separate occasions when Greg, has pushed his two young children 104 causing them to both hit their heads, one has led to them being sick. Greg and Julie did not seek medical attention for this. Julie later disclosed physical and verbal abuse from Greg, she also raised concerns that Greg has increasing anger issues and may be suffering with his own mental health difficulties. It is known that two of the children have autism and it has been reported that Greg may also struggle with this and that this causes additional stress to the family. Following the information reported to the police, a referral was made by the police to CSC, and a section 47 was initiated. The perpetrator left the family home, and didn’t return to the family home, but with nowhere else to stay, was living in the stock room at his place of work. The family home is a secure social tenancy in the perpetrator’s name (with a local privately registered provider). As Greg and Julie are not married, Emily also has no marital rights to the family home. The perpetrator was referred to the ASW to address his immediate and long- term housing needs alongside the needs of Emily via the Partner Support Worker. Accommodation support: The ASW completed an accommodation assessment with the perpetrator and presented his case at the weekly accommodation panel. It was decided that he would be placed in immediate hotel accommodation funded by Restart for a 4-week period. He shared with the ASW that he wished to sign his tenancy over to Julie, particularly as the children were living with her, and the younger are on the autism spectrum. However, it has been difficult for the panel to engage the private social housing provider to facilitate this, and in the absence of this, the victim/survivor has received the support of a family solicitor to seek a tenancy transfer. This first requires that the victim/survivor seek a Non-Molestation Order and an Occupation Order, which she is returning to court to do, and then to apply for a tenancy transfer thereafter. Emily has been supported by the Partner Support Worker with emotional support, benefits applications, and support to gain a non-molestation order and to advocate for housing needs through the housing panels and direct advocacy. In the meantime, the ASW has been supporting the perpetrator to prepare to access longer term accommodation, including submitting all required documents for affordability assessments and access to benefits. The perpetrator works full-time, earns a decent wage to be able to access longer-term accommodation. After two weeks in hotel accommodation, the borough placed him in LA temporary accommodation in preparation for placing him in specific studio accommodation for working people. It is an assured shorthold social tenancy for two years. He will separately need to be nominated for longer term social housing. He has since been placed in the accommodation. The perpetrator has not worked with the Case Manager much to address his abusive behaviours, however this will be readdressed once accommodation is established, as engagement with Case Manager is required for long-term accommodation support. 105 Lessons: • Having the involvement of privately registered social housing providers is vital from the beginning and requires the LAs and Cranstoun to build relationships with their largest providers. This is because these providers may be housing the victim/survivors and perpetrator Restart is supporting and may make housing transitions much easier. In this case, the perpetrator and victim/survivor may have been able to work with the housing provider to sign the tenancy over to the survivor, provide lock changes and additional security measures. • It should not be assumed that the perpetrator will be supported first with behaviour management and secondly with accommodation. In some instances, it may be more effective to support the perpetrator to address immediate housing needs, to be able to focus on behaviour management. As engagement with DA support programmes are a requirement for longer-term accommodation, this will provide an opportunity for support which we will monitor and evaluate. 106 Glossary and list of abbreviations: ASC Adult Social Care CAMHS Child and Adolescent Mental Health Services DASH Domestic Abuse, Stalking and Honour-Based violence risk assessment score Drive Drive is an intervention aimed at high-risk high harm perpetrators of domestic abuse who pose a risk of serious harm to people they are in intimate or family relationships with. It challenges and supports perpetrators to change and works with partner agencies like the police and social services to disrupt abuse. It is run by the Drive Partnership, see http://driveproject.org.uk/stakeholders/ DVPP Domestic Violence Perpetrator Programme IDVA Independent Domestic Violence Advocate (or Adviser) MARAC Multi-Agency Risk Assessment Conference Men & Masculinities The Men & Masculinities Programme is intended to provide a safe space for men who have engaged in abusive, harmful and damaging behaviour within their relationships. Run by Cranstoun, it is a 24-week programme which focuses on behaviour, see https://cranstoun.org/help-and-advice/domestic-abuse/men-and- masculinities/ PRS Private Rented Sector PSW Partner Support Worker S&T Safe and Together SOAG Severity of Abuse Grid/Violence VPP Violence Prevention Practitioner VS Victim-Survivor VSW Victim Support Worker April 2011 An evaluation of the Diamond Initiative: year two findings Foreword The Diamond Initiative, which became operational at the beginning of 2009, was an imaginative criminal justice policy innovation set up by the London Criminal Justice Partnership. It had at its heart two key features. First, in England and Wales short-term prisoners (those with sentences of less than twelve months) receive no statutory supervision on release, so Diamond offered them resettlement help that they would otherwise not have received. Second, studies of the geography of crime have consistently shown that deprived urban areas suffer disproportionately from criminal victimization, and they also contain among their residents a higher than average proportion of registered offenders. Focusing a fresh crime reduction initiative on deprived urban areas – as Diamond did – therefore makes good strategic sense. But, as we pointed out in the Foreword to the Interim Research Report on the Diamond Initiative, crime reduction initiatives do not always work, and it is vital that they are rigorously researched. The London Criminal Justice Partnership is therefore to be congratulated, not only for its courage in setting up this new Initiative, but also for its determination to evaluate the initiative fully. That task has been undertaken by the Strategy, Research and Analysis Unit (SRAU) of the Metropolitan Police, supported by an independent Academic Reference Group (ARG) of which we have been the members. At the time of the Interim Report, no definite conclusions could be drawn about the results of the Diamond Initiative, but there were a number of reasons for cautious optimism. It was of course right, in the interests of accountability and transparency, to state these reasons in the Interim Report. But we live in a complex world where crime issues are of great political interest, and a 24-hour media culture is voracious in its search for stories. So publication of the Interim Report had the effect of increasing the already high political profile of the Diamond Initiative. Under pressure from the media and others to declare Diamond an unqualified success, SRAU researchers and members of the ARG found themselves having resolutely to insist that the research was not yet completed, and that it would be inappropriate to act prematurely on interim results. This Final Report on the results of the research justifies the earlier caution. After a very careful and thorough evaluation, SRAU researchers have concluded that the assessment of the crime- reductive potential of the Diamond Initiative must be less encouraging than appeared to be the case from the data available at the time of the Interim Report. This report therefore carries within it the potential for a reverse danger from that of the Interim Report. Whereas the earlier report encouraged some to think in terms of unqualified success, the more disappointing results now presented may lead to careless talk of unqualified failure. Neither of these reactions is justified. In reality, the results are complex – reflecting the complexity of the real world – and they need careful interpretation to ensure that the right conclusions are drawn from them. In particular, it is important to emphasise that the results presented in this report do not mean either that the Diamond Initiative was inappropriately mounted, or that the principles of what is often called ‘integrated offender management’ are undermined by the research. The Ministry of Justice has recently been speaking about a ‘rehabilitation revolution’. However, those who are familiar with the history of research evaluations into rehabilitative work with offenders – a history that now stretches back for half a century – tend to be uncomfortable with this kind of language. On the basis of this body of research work, several conclusions can be drawn with some certainty, but they lead to a policy stance of gradual reformism rather than revolution. What are these conclusions? There are four. First, helping people to turn their backs on a life of crime is a slow and uncertain process. It is frequently a case of “two steps forward, one step back”, because of the complexity of offenders’ lives (including, very often, matters such as a lack of qualifications and work experience; drug and alcohol use; a tendency to revert to previous patterns of behaviour in a crisis, and so on). But second, and encouragingly, most people who become heavily involved in offending also eventually desist from crime – completely or largely – at some stage in their lives. Despite their criminal records, the evidence is clear that most of them do not wish to continue with a life of crime, and gradually start to turn away from it. Third, the most thorough research evaluations of treatment initiatives, when they show positive results, tend to report fairly small effects (in small-scale ‘demonstration’ projects run by trained specialists, stronger results are sometimes obtained, but it has proved very difficult to replicate such results when a similar programme is ‘rolled out’ on a larger scale). The main implication of these studies is that programmes of work with offenders can accelerate the desistance process, but viewed in the round, they tend to have only a modest impact. This is probably because the work of police, prison and probation officers are only one of many influences on offenders’ lives. Hence, there are several useful tools in the rehabilitation armoury – but no ‘silver bullets’. Fourth, rehabilitation programmes can easily become derailed, especially when they involve complex partnerships between several different agencies; the successful implementation of fresh and apparently promising approaches is a real challenge. The Diamond Initiative put into practice a set of principles for which there is considerable empirical support. If (as is the case) most repeat offenders would, in adulthood, like to move away from crime, it makes obvious sense to try to support their aspirations by providing appropriate help in matters such as housing, employment, and drugs treatment. In developing such support packages, it is also obviously sensible for the police and probation services to work together, in close partnership with other relevant agencies such as housing departments and job centres. Additionally, there are very good strategic reasons to focus on providing such support to people leaving prison after short sentences: we know that such offenders tend to have extensive criminal records, and a high headline reconviction rate, so this is a group for whom the biggest returns can be expected from a successful intervention. Given all this, the Academic Reference Group that oversaw this evaluation were surprised by the principal finding of the SRAU research – that the extent and nature of known reoffending in the experimental group was very similar to that in a matched control group. How is one to make sense of this result? The first possibility is that the project had to move forward too far too quickly, and that the scheme did not have enough time to find its feet. There is indeed some supporting evidence of implementation problems. For example, the police were able to deploy resources to Diamond more rapidly than the probation service and other partners – a finding that is not uncommon with partnership projects – with the result that the ‘prospectus’ offered to early participants failed to deliver fully on its promises of multi-agency support. Moreover, as the report states, the Initiative never fully reconciled some working tensions between the police and other partners. A further (and related) possibility is that the timescales for measuring an impact were too short. We know that it takes time – and often many attempts – for people to change ingrained but self-destructive habits such as smoking, drinking, drug use and overwork. Offending is no different, and the processes of change do not always suit the timetables of organisations operating in a highly politicised environment, and looking for quick results. The Academic Reference Group obviously asked itself whether the reasons for the lack of impact might lie in the research rather than the programme. For example, the main outcome measures, reconviction rates, might have been artificially inflated in the experimental group precisely because they were kept under closer surveillance by official agencies than the control group. Alternatively, the assumption that reconviction rates are closely related to reoffending rates may not hold up in the high-crime areas in which the Diamond Initiative was set. Possibly, too, the process of selecting a control group matched to the experimental one could have been flawed. These are certainly possibilities, but the Group’s overall conclusion was that the evaluation was conducted to a higher than usual standard of rigour, and that the lack of effect, relative to the control group, is real rather than artefactual. However, the report also contains a more positive set of findings that we think it is important to highlight. Diamond’s support was offered on a voluntary basis to offenders serving sentences of imprisonment. They were in no way obliged to agree to the offer and in deciding whether to accept or not they were aware that many of those staffing the Initiative were police officers. Moreover, these were not neophyte offenders; on average, they had been in court and convicted on eleven separate occasions. Of course, some suspicions were expressed. But no fewer than 60% of those who received an offer of support from Diamond accepted it; and those who accepted tended to have a higher risk of reconviction and more social needs than those who did not. There can be no better testimony as to why this initiative was worth attempting, and why it should now be built upon. To adapt a memorable saying of Samuel Beckett: ‘No matter. Try again. Try better.’ What, then, are the main practical implications to draw from this evaluation? Perhaps the most important is that helping people to desist from crime involves a long-term commitment; if the police, probation service and their partners - or indeed the government - expect a return on their investment in the space of a year or less, they will very likely be disappointed. Second, the study has certainly demonstrated the feasibility of involving police officers in rehabilitative work with offenders; and we would expect the long term effects of joint working to be beneficial both for the police and for their partners in the probation service and elsewhere. Third, there is an obvious value in information-sharing across custody and community settings, yet also some serious challenges in making this happen as smoothly as it should. Although the headline results of this report will disappoint staff on the Diamond Initiative, there is nothing here that calls into question the principle of multi-agency teams designed to plan and deliver social provision that is inclusive of and helpful to offenders. There is a great deal in this report that may help the Metropolitan Police, the London Probation Service and their partners make this vision a reality in the future. They are to be congratulated for their investment in this Initiative and this evaluation – and the SRAU are to be congratulated on the rigour with which the evaluation has been completed. Sir Anthony Bottoms Emeritus Wolfson Professor of Criminology, University of Cambridge (Chair) Rob Allen Formerly Director, International Centre for Prison Studies, King’s College, London Roger Bowles Formerly Professor, now External Affiliate, Department of Economics and Related Studies, University of York Brian Francis Professor of Social Statistics, Lancaster University Mike Hough Professor of Criminal Policy, Institute for Criminal Policy Research, Birkbeck College, London Gill McIvor Professor of Criminology, University of Stirling and Co-Director, Scottish Centre for Crime and Justice Research Nerys Thomas Analysis and Science Programme Manager, National Police Improvement Agency Contents Page Chapter 1: Introduction 1 Background to the London Diamond Initiative 1 Setting the scene - the challenge of the prison population in 4 England and Wales Getting to the crux of the problem: what is known about reducing 5 reoffending? The evolution of offender management 10 Chapter 2: Methodology 12 1. Online staff survey 13 2. Staff interviews 14 3. Offender interviews 15 4. PNC analysis 16 5. Economic impact 17 Chapter 3: The process of delivering the Diamond Initiative - 18 insights into implementation 1. Initial set up and how the DI matured over time 18 2. Multi-agency working and meeting offender need 23 3. The challenge of engaging with offenders 27 4. Working across organisational cultures 30 5. The offenders’ experience of Diamond 33 6. Gearing down and looking forward 35 Chapter 4: The impact of the Diamond Initiative - what do we know 38 about reducing reoffending? Evaluating the impact of a local offender management initiative 38 Did the Diamond Initiative reduce reoffending? 42 Close-up on crime - did the Diamond Initiative have any 44 other effect on offending? Checking for impact among Diamond participants 50 Insights into promising aspects of the Diamond Initiative 52 Did better implementation lead to better reoffending outcomes? 53 Economic evaluation 55 Chapter 5: Discussion 58 Lessons from the Diamond Initiative 60 The future of offender management 63 Final thoughts on the evaluation 65 References 68 Appendices 1. Forming the control group - borough and ward level matching 74 2. Top 10 best matched boroughs, for each Diamond borough 76 3. Top 6 best matched wards, for each original Diamond ward 77 4. Evaluation cohort and control group comparison statistics 79 5. Diamond referral outcomes 81 6. Reoffending and previous offending for UK born and non-UK born 82 Sub-sections of the evaluation and control groups 7. Examples of categorised offences committed by the evaluation cohort 83 post-release 8. Proportion of those offered Diamond who accepted by sub-groups 84 9. Forming the participant matched control group 85 10. Diamond participants and participant-matched control group 86 comparison statistics 11. Main and Later cohort, six month and predicted reoffending 88 by referral outcome group 12. Staff survey results (selected questions) 89 13. Conceptual comparison of DI and an IOM approach 91 The Strategy, Research and Analysis Unit - Diamond evaluation - April 2011 Led by Paul Dawson & Prof. Betsy Stanko Supported by Andy Higgins, Ubaid-Ul Rehman, Renata Hope, Ann Walker, Lynne Grossmith, Alison Squire, David Martin, James Bennett, Michael Keenan, Liam Fenn, Colleen Reddin, (MPS Strategy, Research and Analysis Unit) & MVA Consultancy. The Academic Reference Group This group had oversight of the entire DI research proposal and provided guidance and support. Prof. Sir Anthony Bottoms Prof. Brian Francis Prof. Roger Bowles Prof. Gill McIvor Prof. Mike Hough Rob Allen Nerys Thomas Thanks to the staff at the LCJP, in particular the DI central team who have been supportive at all stages of the research Andrew Morley Martin Stevens Mary Alston Diana Binding Chris Hemstead Holly Bright Jackie Ashby Ian Thomson Mark Darby Thanks to all Diamond project teams and Local Authorities for their support of the programme and the research over the two years. Please cite as: Dawson, Stanko, Higgins and Rehman (2011). An evaluation of the Diamond Initiative: year two findings. London. Metropolitan Police Service; London Criminal Justice Partnership. A large print version is available on request. The Diamond Initiative - evaluation results and learning In 2008/09 there was a compelling case for testing a Justice Reinvestment model of offender management within London. Drawing on the ‘million dollar blocks’1 concept and an emerging consensus on the benefits of multi-agency collaboration, the London Criminal Justice Partnership (LCJP) - working with the Metropolitan Police Service (MPS), London Probation and six Local Authorities developed the Diamond Initiative (DI), a multi-agency offender management scheme aimed at breaking the cycle of reoffending among problematic groups of offenders in some of London’s most challenging locations. 1 See Tucker and Cadora, 2003 Although the political and economic context altered during the two-year pilot period, this only served to increase the urgency of the search for effective alternatives to custody and the unsustainable increase in demand on the Criminal Justice System. With Integrated Offender Management (IOM) approaches taking root across the UK and with the Coalition Government’s Rehabilitation Revolution set to catalyse innovation in service delivery, insights into how cost-efficient offender management can contribute to reducing crime and demand on Criminal Justice are keenly sought. An evaluation of the Diamond Initiative: year two findings The Diamond Initiative set out to pilot Justice Reinvestment principles and an integrated approach to managing short-sentence offenders, in some of London’s most challenging areas. This comprehensive evaluation of implementation and impact holds relevant insights for practitioners and policy makers in London and beyond. It describes an initiative, hampered in the early stages by implementation issues, which matured in terms of multi-agency delivery but never fully reconciled ‘working tensions’ between the police and other partners. It experienced some success in engaging with offenders through a voluntary model but may not have adequately aligned the support provided with offenders’ needs. These insights provide context to the finding that no impact was detected in terms of reduced reoffending as a result of the initiative, although indications of more positive outcomes for particular sub-groups warrant further follow-up. Method This report presents the findings of a robust evaluation of the Diamond Initiative carried out by the MPS Strategy, Research and Analysis Unit (SRAU). It sets out to investigate both the implementation and the impact of the DI, drawing on a range of methodologies including: 1) Four waves of on-line staff surveys designed to capture the changing views of practitioners over the course of the scheme. 2) Two waves of qualitative staff-interviews to dig deeper into implementation issues and capture key learning from a representative sample of multi-agency staff. 3) Two phases of qualitative offender- interviews, to reflect the experiences of those participating in the scheme. 4) A quasi-experimental study of reoffending, using data from the Police National Computer (PNC) to compare the offending behaviour of Diamond referrals to a well matched control group, over a one- year period. 5) A cost-benefit analysis of the DI, comparing the economic benefits generated by the scheme, in terms of the cost of crimes averted, against the set-up and running costs of programme. Key Findings Process evaluation Implementation developed and matured over the course of the pilot. Teams experienced initial problems in recruiting and retaining sufficient and suitable staff, securing appropriate accommodation, equipment and IT and adjusting to new ways of working. Although problems persisted in certain areas, overall implementation improved over the course of the pilot. Multi-agency working also took time to develop. In particular, links with local housing and drugs intervention services strengthened during the programme, however questions remain as to whether the network of services brought together by Diamond adequately addressed the challenging and complex needs of the client group. Working across agency boundaries also brought challenges, best conceptualised as ‘working tensions’ between police and other partners. While the police brought dynamism to the initiative, some officers struggled to adapt to a role with a supportive rather than supervisory focus and although police personnel often reflected positively on working in a new way, their inexperience in areas like case management led some to question the appropriateness of the police- led model. Although the referral process used had significant flaws (highlighting issues with the flow of information between prison and the community) teams were relatively successful at engaging with offenders. Most of those offered Diamond assistance agreed to participate, and those who engaged tended to be toward the more problematic end of the offending scale. This indicates strong, need-driven demand for the services offered by Diamond and that the appeal was not substantially undermined by offenders’ initial apprehensions about getting involved with the police. Despite this, project case-loads tended to be lighter than had been anticipated. Practitioners were clear on the benefits of starting the engagement process early, while offenders were still in prison. In line with the desistance literature, the offenders interviewed frequently expressed the desire to change their lives and to stop offending; it was also clear that since prison, many felt their lives had improved and saw Diamond as integral to this transformation. Impact evaluation The reoffending rate of the Diamond referral group was found to be significantly below the (61.1%) national reoffending rate for short term prisoners. This should be understood in terms of pre-existing differences between the Diamond group and the national sample and not as an impact of the initiative; Diamond referrals had less prolific offending histories and lower predicted reoffending rates than the national average. It was also observed that a substantial proportion of Diamond referrals were born outside of the UK (44%); this group were found to commit fewer offences (both historically and during the follow-up year) - and are likely to make up a larger proportion of the Diamond group than the national sample. An Intention to Treat (ITT) design was used to compare the offending behaviour of a cohort of less-than-12-month sentence offenders, referred to Diamond during the first nine-months of the programme, against a well matched control group of similar offenders drawn from the similar, matched parts of London. Over a 12-month follow up period, no significant difference was found between the reoffending rate of the Diamond referral group and the control group; 42.4% of the Diamond referral group and 41.6% of the control group committed a criminal offence (which subsequently resulted in a conviction at court), during the year after they were released from prison. A deeper, more nuanced examination of crime data similarly revealed no differences in terms of the speed of reoffending, the number of offences committed or the seriousness of those offences. A second test for impact, comparing Diamond participants (as opposed to referrals) with a suitable control group, produced similar findings. Although small base sizes and other factors dictate that findings should be treated cautiously, there are some indications of more positive outcomes from Diamond, for instance:  Diamond referrals with a recent history of violent and disorderly offending tended to reoffend less within these offence types than the control group.  Those who achieved ‘outcomes’ related to finance, benefits and debt through Diamond, show promising reoffending results across a range of measures.  A cohort of those referred to the scheme during a slightly later implementation phase (when many of the delivery improvements described above had taken effect) show encouraging early indications of a more positive impact on reoffending. In the light of the above findings it is unsurprising that no economic return on investment was detected in terms of the cost to society of crimes averted. The Diamond evaluation presents substantial learning for offender management within and beyond London and the results are relevant for practitioners, academics and policy makers. The results should lay down an important evidence- based marker in terms of designing and evaluating an offender management initiative. The results of the evaluation do not demonstrate that Diamond has failed as a concept; instead careful consideration is required as to why the results are not as expected. Time should be spent examining the learning generated through the research to ensure that the next phase of offender management is as effective as it can be. Key learning from the Diamond Initiative evaluation The evidence assessed in this evaluation lends itself to a range of key learning points which are likely to be relevant to practitioners and policy makers working across offender management in London and beyond. Implementation learning Recruitment and staffing levels were a consistent challenge. Coordinating recruitment protocols is likely to be crucial when forming interdisciplinary teams. Staff attitudes, motivation and ‘buy-in’ directly affects implementation; seconding reluctant staff to offender management roles should be avoided. Effective staff training, using continuous feedback for monitoring success, should be a core part of any offender management. An effective case management and referral system should be in place at the outset with only minor amendments subsequently made. Despite initial problems, implementation matured over the course of the programme. Any future-roll out should anticipate and make contingency plans to address early implementation issues (e.g. equipment, buildings, staffing, IT systems). The mix of help offered may not have been effectively matched to client need. Provision should be designed based on local analysis of the needs of the client group, but anticipation of standard problems, based on previous research should be built in at the planning stage. Unmet client need is as much a strategic problem as it is a practical problem. Staff reported positively on the practice of engaging with offenders in prison. Given the evident problems with the flow of information between custody and community and the importance of the first weeks and months following release, an effective link-up with local prisons is a key to success. The challenge of how to increase participant numbers was felt throughout the DI; the scale of provision needs to be matched to the likely supply of eligible clients for the defined catchment area. The original focus on key wards was found to be restrictive in practice. Although teams generally worked well across disciplinary boundaries, tensions between the police and other agencies linked to different ‘working cultures’ were never fully reconciled. There is still work required to establish an integrated approach. Pilot initiatives suffer from uncertainty. Staff feel this uncertainty. Strategies and contingencies for the end of an initiative should be considered in programme delivery. Impact learning The process of identifying and engaging with potential participants was severely hampered by the working of the referral system. Improving the flow of information between custody and community is likely to be crucial for delivering effective offender management. 60% of offenders offered the scheme agreed to participate; this indicates substantial demand for resettlement help and support. Moreover, those who accepted support were at greater risk of reoffending and presented more needs than those who refused, therefore the demand for help appears to be need-driven. The voluntary model for offering help appears suited to engaging the most problematic offenders. No difference has been found between the Diamond referral group and the matched control in terms of the proportion who reoffended, the speed at which they reoffended, the number of offences they committed or the seriousness of those offences. The two groups of offenders exhibited highly similar offending careers, which continued to be comparable after prison release, despite the fact that the evaluation group were referred to the Diamond Initiative. Therefore, no indication of a positive impact of the Diamond Initiative on offending has been detected. The role of implementation issues in this should be considered. Any offender management initiative should consider the most robust practical means of assessing impact and use this to drive improvement. The level of reoffending within the Diamond referral group was below that of the national less- than-12-month sentence cohort. This should be understood in terms of differences in the make-up of the referral group compared with the national sample. This suggests the need for regional / local benchmarking in order to provide a framework for assessing the differences and effectiveness of local offender management approaches. Those who reoffend tend to do so quickly after prison release; this indicates a ‘window of opportunity’ for supportive engagement immediately following (or commencing before) prison release. There are tentative indications that those referred to Diamond later in the pilot, might have achieved better reoffending outcomes. Further analysis, over a longer follow-up period would be required to draw firm conclusions; however if this is sustained in the next analysis, the lesson that early implementation issues should be carefully programme managed, will be clear. Some differences in success were observed and are worthy of more research when the numbers of clients enable more robust statistical analysis. In particular, there are tentative indications that the Diamond approach may have had a positive impact on violent / disorderly offending and that help to address debt and personal financial issues may be linked with reduced reoffending. Given that no overall impact on offending has been found it is unsurprising that no economic benefit, in terms the cost of crimes averted, has been detected. Cost efficient offender management must be scaled to the size of the target population and target resources effectively, for instance by using risk assessment tools (such as OGRS), to achieve standardised and potentially impactive selection. 1 Chapter 1: Introduction This report presents the findings of a two-year evaluation of the London Diamond Initiative (DI). It presents insights into both the implementation and the impact of the DI - a multi-agency offender management approach aimed at reducing reoffending among key groups of offenders in some of London’s most challenging areas. The report sets out to capture the learning generated over the course of the scheme and to contribute to the evidence-base available for policy makers and practitioners to develop and deliver effective offender management within London and beyond2. Background to the London Diamond Initiative In terms of background, in 2008/09 there was a compelling case for developing a Justice Reinvestment model of offender management within London. Is there an economic case for a shift of resources from central budgets to needy geographic sites, in order to reduce crime and the consequent demand on the Criminal Justice System? This approach draws on the 'Million Dollar Blocks' concept developed in the USA, where it was shown that large sums of public money were routinely spent on cycling and recycling offenders from certain residential areas into the Criminal Justice System without notable benefit to the community. It has been argued that by channelling some of these funds away from the prison system to organisations working with offenders in these communities - recidivism, the demand on the Criminal Justice System and the harmful effects of crime in these areas could be simultaneously reduced (Tucker & Cadora, 2003). Concurrent with these developments, offender management programmes in the UK (such as the Prolific and other Priority Offender scheme and the Integrated Offender Management approach) are increasingly demonstrating the value of information sharing and multi-agency co-operation in effective and cost-efficient offender management. With these principles in mind, the London Criminal Justice Partnership (LCJP) and the Metropolitan Police Service (MPS) sought to develop the Diamond Initiative as a test- case for Justice Reinvestment principles in the London context, by bringing together the MPS (and in particular harnessing the maturing Safer Neighbourhoods (SN) model), 2 Year one interim findings which can be found at http://lcjb.cjsonline.gov.uk/area23/library/Diamond Initiative/Diamond Initiative.pdf. 2 London Probation Service, Local Authorities and key local agencies, with the aspiration of reducing reoffending in the client group and consequently generating substantial savings within the Criminal Justice System. The twin unique selling points of the Diamond approach were the emphasis on needy geographic areas (e.g. certain London wards) and the focus on the less-than-12-month sentence group of offenders. In line with ‘Million Dollar Blocks’ principles, the LCJP targeted the DI at particular wards within six London boroughs (Croydon, Hackney, Lambeth, Lewisham, Newham and Southwark) which, analysis showed, presented particular challenges for criminal justice agencies in terms of high custody returns. Within these localities, Diamond teams were established and tasked with addressing less-than- 12-month sentence offenders returning to the community as their principal client group. This group - often seen as the perennial thorn in the Criminal Justice System - exhibit high rates of reoffending (and consequently generate high costs to the public purse) and lack a statutory offender management response. Diamond therefore sought to target a ‘problem’ offender group in ‘problem’ locations, encouraging and supporting desistance from crime where it was most needed and consequently maximising reductions in offending and associated financial benefits. In addition to addressing the less-than-12-month group, Diamond teams also sought to test the concept of ‘supported compliance’ by working with those sentenced to Community Payback Orders within the catchment areas. The LCJP’s performance data indicate some success and positive learning from this approach including a 20% improvement in compliance rates over a six month period. While acknowledging the promise of supported compliance this evaluation primarily focuses on the principal less- than-12-month-sentence target group. During the course of the Diamond pilot, the political and economic climate in which it operated altered substantially. In January 2010, for instance, the House of Commons Justice Committee stressed that in a time of recession, funding and resources to all elements of the criminal justice system agencies must be cut, and in light of this, new ways of dealing with offenders and offending behaviour should be examined - many eyes 3 have consequently turned to the Diamond Initiative as a test of a new way of working with offenders in a challenging financial climate. The MPS Strategy, Research and Analysis Unit (SRAU)3 were tasked by the LCJP/MPS to conduct a two-year evaluation of the Diamond Initiative. The evaluation has two main elements: firstly, to explore the process and implementation of the DI; and secondly to examine the impact of the scheme, both on crime and wider financial efficiencies. Promising interim results were published last year, although in terms of crime impact the year-one report concluded that 'the absence of a matched control group and a short follow up time preclude firm conclusions around the impact of the initiative at the present time'. Otherwise, substantial learning was generated in relation to both the less-than-12- month sentence client group and the implementation of offender management initiatives. For example:  The key target group on Diamond were shown to have relatively extensive criminal careers - with an average of 23 convictions and 12 conviction occasions. 3 The SRAU is a member of the Government Social Research (GSR) group, and offers independent evaluation and research to drive improvement. An overview of the Diamond Initiative:  Built on a compelling case for a 'Justice Reinvestment' model in London.  A two year progressive offender management initiative focussed within six London boroughs.  A focus on areas with high concentrations of offenders returning from custody.  Principally targeted at the less-than-12-month sentence release group, with a wider target group including those on Community Payback Orders.  Delivering services through a combination of criminal justice agencies (police, probation and prison) and resources from a wider range of stakeholders (particularly local authorities, health and third sector organisations).  A voluntary, needs based offer of support and assistance for offenders. 4  Multi-agency staff considered working on the Diamond Initiative to be a positive experience and a new way of working due to the emphasis on rehabilitation: in particular this was viewed as a departure from traditional police culture.  Diamond Initiative offenders reflected positively on their experience of the initiative providing examples of benefits such as not committing crime, assistance with their needs and finding stability and structure in their lives. This report builds on the learning from the year-one evaluation and presents conclusions in terms of the implementation process, impact on reoffending and value for money whilst also raising the implications, both methodological and practical for offender management initiatives. Before doing so however, it is necessary to lay the foundations by exploring the context in which the initiative has operated and summarising what is known about effectively reducing reoffending as part of an offender management programme. Setting the scene - the challenge of the prison population in England and Wales The prison population in England and Wales is, per head of population, the highest in Western Europe and the fourth highest in the world. Currently over 85,000 prisoners reside in the 137 prisons in England and Wales (HM Prison Service, 2011). Between June 1995 and June 2010 there has been a rise of 66% in the prison population, with tougher sentences (such as Indeterminate Sentences for Public Protection (IPP)), more serious offences coming before the courts, and legislative and policy changes cited as contributory causes (Ministry of Justice, 2009). An increasing prison population has associated costs. There has been an increase in central spending on prisons and offender management from £2.6 billion in 1996/1997 to £4.6 billion in 2009/2010 and accommodating short term prisoners costs the tax-payer £286 million a year (Audit Commission, 2010). During 2007/2008, the cost of dealing with the re-conviction of offenders was estimated at between £9.5 and £13 billion, most of which was spent dealing with those who had previous short-term sentences (Audit Commission, 2010). 5 Official statistics reveal that most offenders enter prison on short-term sentences. Every year more than 60,000 adults receive a custodial sentence of less than 12 months; this group account for nine per cent of the overall prison population, but make up some sixty- five per cent of all those sentenced to immediate imprisonment. On release the less-than- 12-month sentence group receive no statutory probation supervision, no dedicated provision for welfare and resettlement needs (The Social Exclusion Unit, 2002) and sixty percent reoffend within one year (Ministry of Justice, 2010a). Given this high level of reoffending, there is a growing consensus that prisons alone cannot provide the solution to the economic consequences and challenges of offending outlined above. Innovative work needs to investigate whether support and supervision of short-term sentence offenders in the community, may hold the key to reducing offending behaviour and breaking criminal lifestyles, in turn cutting the prison population and bringing a saving to the economy (Prison Reform Trust, 2010). Getting to the crux of the problem: what is known about reducing re-offending? There is a wealth of research that has sought to examine either the development of desistance or criminal careers (Bottoms et al., 2004; Farrington, 1997; Weaver & McNeill, 2007; see Farrall et al., 2011 for an overview). This research can provide essential insights and direction on bringing about behavioural change within an offender management approach. Criminal careers can be examined in two ways: at a cross-sectional level (e.g. as a snapshot of all offences and all offenders at a given time) or longitudinal level (e.g. following a set group over their lifetime). Whichever technique is adopted, the key observations are similar. Typically, individuals begin offending in early adolescence, peak in late adolescence/young adulthood and then gradually reduce, with a typical cessation before the age of 30 (Bottoms & Shapland, 2011; Laub & Sampson, 2003; Van Mastrigt & Farrington, 2009). While age and other static factors (e.g. criminal history) are clearly important in understanding criminal behaviour and desistance, and have been used to generate 6 accurate predictive models such as OGRS4 - these static factors by their nature, cannot be altered and thus are of less value in terms of practical offender management. Importantly, research highlights how contextual, situational and motivational factors also play an important role in the management of offenders. This is because such issues, typically known as dynamic risk factors, can be actively addressed with the aspiration to bring about behaviour change. Table 1 provides a basic overview of the factors associated with offending, illustrating the range of issues at the heart of the re-offending debate. It has been demonstrated that offenders often have numerous problems / needs and those with a complex needs-set are the ones that are most likely to re-offend (McGuire, 2002; Stewart, 2008). These needs have also been evidence based and used in the development of the Offender Assessment System (OASys) - the risk assessment tool used by the prison and probation service. OASys promotes individualised sentence plans matched to risk and addressed to dynamic (i.e. changeable) needs. Reporting on the Sheffield Longitudinal study5 Bottoms and Shapland (2011) highlight that even within a highly criminal sample, offenders reported an apparently sincere desire to desist from crime. Although this aspiration was often not fully accomplished, instead smaller steps (i.e. reduced frequency of offending / periods of abstinence) were frequently apparent. This indicates that desistance can be a gradual process, often impacted by 'significant obstacles' including external factors (such as financial problems) and psychological factors such as self-efficacy, a lack of empathy and the emotional pull of offending. This raises the important point of the offenders ‘motivation to change’ which is particularly relevant to Diamond given the voluntary nature of the scheme. We know from previous research that offenders who complete an initiative generally achieve better outcomes than those that drop-out or do not start. One explanation of this is those offenders that ‘complete’ have enhanced motivations to change their criminal lifestyle (Day & Howells, 4 The Offender Group Reconviction Scale (OGRS) is a regression based model for predicting likelihood of resanction based on static factors such as age, gender and criminal history. It has been used by probation and prisons in the UK since the late 1990s. 5 The Sheffield Desistance study interviewed offenders born in either 1982,1983,1984 over four points in time convicted on at least two separate occasions (Bottoms and Shapland, 2011). 7 2002; Hollin, 2008). On a therapeutic level there is learning on understanding and improving motivation to change; including the importance of staff training and motivational interviewing to promote life change (see McMurran 2002 for an overview). However, many questions remain to be addressed, for example, what is motivation to change? How can it be measured? Are there acceptable levels before entering behavioural change programmes? Indeed, Weaver and McNeill (2007) stress the value in being realistic when attempting to address criminal behaviour within criminal justice settings, noting that 'the criminal careers of reoffenders can't be switched off like a tap; it takes time to change entrenched behaviours and the problems that underlie them'. In terms of reducing offending behaviour, beside the research into criminal behaviour previously discussed, there is a substantial body of research that demonstrates that the actual design and implementation of an initiative can play a substantial role. The ‘what works’ principles (McGuire, 2000) are an established evidence-based method of working with offenders, that demonstrate that programmes work best in reducing re-offending when certain key principles are met. This has been shown in relation to sex offender, drug treatment, education and accommodation programmes (Aos et al., 2006; Lipsey, 1995; National Institute of Drug Abuse, 2003). The ‘what works principles’ would expect a programme to:  be based on a robust model/theory with a underpinning rationale;  include risk classification;  target and respond directly to offender needs;  ensure treatment is skills-orientated, based on cognitive behavioural techniques; and  have programme integrity. This issue of programme integrity warrants further examination here - initiatives with 'integrity' set out a planned series of activities and are delivered over a specific timeline to selected individuals. A programme with integrity ensures that the intervention or treatment is delivered exactly as intended. This would likely be supported by clear training, guidance documents, manuals, monitoring instruments, and in many cases would involve both pre-assessment work (i.e. enhance offender motivation) and post programme work 8 (i.e. follow up / reviews). How the programme is designed and delivered affects the type of results obtained and initiatives without integrity do substantially worse than those with strong programme integrity (Barnoski, 2004; Coulter, 2010). 9 Table 1: Key research on offender needs and crime A past criminal history Past convictions, custodial sentences and arrests have been found to be predictive of future offending (Copas et al., 1994; Kurlychek et al., 2004; Spivak & Damphousse, 2006). Associated to this are biological/genetic theories of crime (Mednick et al., 1987). Housing Living in suitable accommodation following prison release appears to be a protective factor against reoffending (Social Exclusion Unit, 2002). Conversely, having no fixed abode, living in a high crime area or estate, and frequent address changes are identified as risk factors for reoffending (Morton, 2009; Raynor et al., 2000). Education and employment A lack of education and employment have been identified as important factors in reoffending (Gendreau et al., 1996; Hollin & Palmer, 2006). This is also associated with school exclusion (Raynor et al., 2000), dropout (Farrington, 1990); poor literacy and numeracy skills (Baker, 2002). Employment related programmes have demonstrated positive results (Lipsey, 1995; Sampson & Laub, 1993). Financial management Whist a lack of finances may be the by-product of unemployment, there is evidence to further suggest that poor financial management also contributes to reoffending (Nilsson, 2003). Offenders in the Sheffield desistance study reported a lack of money as the most important obstacle to ‘going straight’ (Bottoms & Shapland, 2011) Relationships Positive relationships appear to play a strong role in desistance from crime - for example within childhood (poor parenting, abuse, family criminality) (Farrington, 2002; Gendreau et al., 1996; Rutter et al., 1998). Adult relationships and receiving family visits when in custody also reduces offending, with clear links to social networks (King et al., 2007; Niven & Stewart, 2005; Sampson & Laub, 1993). A criminal lifestyle / thinking Youth delinquent or adult pro-criminal associates are strongly associated to future offending (Gendreau et al., 1996; Raynor et al., 2000). Associated here are pro-criminal attitudes and recidivism (Baker, 2002; Palmer & Hollin, 2004). Mental health and personality issues Prisoners experience increased levels of mental health problems (Herbst & Gunn, 1991; Sawyer & Lart, 1996). There is some evidence that certain types may be predictive of recidivism (Bonta et al., 1998). In addition impulsivity, risk-taking, lack of self control and a need for excitement have all been identified as important (Baker, 2002; Raynor et al., 2000; Rutter et al., 1998). Substance misuse There is considerable research indicating that drug misuse is a strong predictor for recidivism of offending (Raynor et al., 2000; Spohn & Holleran, 2002), and in particular acquisitive crimes (Oldfield, 1996). The relationship regarding alcohol is less clear, although still relevant (Oldfield, 1996; Rutter et al., 1998). Key learning: Many of the problems faced by offenders can be anticipated, based on previous research, and can be built into provision at the planning stage. 10 The evolution of offender management The recent Green Paper’s proposals (Ministry of Justice, 2010a) go some way to addressing current issues relating to the search for alternatives to custody (such as offender management, Community Payback and restorative justice) and reducing demand on the Criminal Justice System. The localism agenda covers aspects of offender management. The idea is to get local people involved and participating in what will affect them, for example communities are being consulted on what work those on Community Payback Orders should be doing in their communities. The rationale for the increased use of community sentences includes retaining offenders’ contact with social capital, maintaining stable housing, and building bridges to increase the chance of re-integration into the community. By working with offenders in a local area, public services can come together to help prevent and reduce the risks of re-offending (Allen & Stern, 2007). It is also the case that previous offender management initiatives such as the Prolific and Other Priority Offender Scheme (PPO) and Drugs Interventions Programmes (DIP) continue to yield positive findings (see Dawson & Cuppleditch, 2007; Home Office, 2010; Skodbo et al., 2007). These programmes are however becoming increasingly subsumed within a broader Integrated Offender Management (IOM) approach, which is an attempt to bring the efficiencies of multi-agency working, across and beyond local public sector agencies, to wider problem offender groups, as defined by local priorities. The recent Green Paper positions IOM as the principal model for managing and rehabilitating persistent offenders and importantly, envisages a key role for the police in delivering reductions in offending - it states “The contribution of the police is critical, working in partnership to identify, manage and control known offenders. Many of the skills needed for good policing have also proved well suited to help manage offenders into a law abiding and disciplined way of life, and to provide reassurance to victims that cutting crime remains an overriding priority”. (Ministry of Justice, 2010a). This brings us to the Diamond Initiative, a multi-agency model, with a strong police component, which has been conceptualised as London's IOM approach and is the only current IOM initiative with a robust and published evaluation that has sought to examine 11 the impact upon crime. DI is viewed therefore as the first test of a new way of working with offenders under an IOM banner. While the future trajectory of IOM remains uncertain, in light of the ‘Rehabilitation Revolution’ heralded in the Coalition Agreement (Cabinet Office, May 2010) and subsequent Green Paper, likely next steps will include the increased emphasis on service provision from outside of the public sector. ‘Payment by results’ mechanisms are being developed to promote outcomes-based commissioning, and with the first such scheme launched in Peterborough in 2010, the favoured model of voluntary and community sector service delivery, with initial funding and investment risk shouldered by social impact investors, presents an intriguing framework for the future of offender management. One implication of particular note is the need for accurate measurement of outcomes to indicate whether finances should be paid, emphasising the need for initiatives to be able to accurately demonstrate performance, but also the need for robust research and evaluation to drive improvement and increased effectiveness. Summary The growing prison population and high costs of crime emphasise the need to invest in the rehabilitation of offenders. There is a research base which can inform efforts to reduce reoffending and provide lessons about the type of initiative that is likely to have a positive influence on crime. This accords with the new government’s desire to push towards non-custodial offender management and, in effect, the Diamond Initiative has unknowingly found itself as the first high profile test of this new Integrated Offender Management; and one that has the potential to generate significant savings within a group of offenders that place a high demand on the CJS. 12 Chapter 2: Methodology There is a substantial debate within the literature as to what is acceptable in the methodological design of a crime reduction initiative evaluation (See Dawson & Williams, 2009; Harper & Chitty, 2005; Hollin, 2008; Maguire & Raynor, 2006; Sherman et al., 1997). At the heart of the debate is a balance between quality versus pragmatism in evaluation. That is - should researchers be instructed to attempt 'gold standard' Randomised Control Trials or control groups, and if not, how much of a compromise in rigour is acceptable before confidence in the results is diminished? This is further complicated as local programmes would most likely not have access to large sample sizes or the technical expertise to develop accurate control groups. The Diamond Evaluation should be seen in this light, a demonstration that a locally based initiative can yield a robust and technically accomplished evaluation. The MPS Strategy, Research and Analysis Unit have adopted an approach which used a variety of research methods to evaluate the Diamond Initiative (see figure 1 for an overview of the evaluation). The research has spanned the entire two-year life-cycle of the DI and has employed various techniques to capture change over the course of the programme, something many evaluations do not examine; this allows insights into improvements in implementation over time by using a triangulation of methods approach. Overall the evaluation sought to examine the process and impact of the DI. Further detail on the methods used is provided below. 13 Figure 1: evaluation timeline 1. Online staff survey Staff feedback enabled the research team to capture the learning around the planning and delivery the Diamond Initiative at various points over the two-year project. An anonymous online staff survey was conducted four times over the course of the evaluation: twice in the first year (wave 1 in March 2009; wave 2 during December 2009), and twice in 2010 (wave 3 June 2010; wave 4 December 2010). The surveys aimed to capture perceptions and experiences of staff in relation to a number of key aspects of their role such as partnership working, staffing and practical issues over the duration of the programme. Near the end the DI pilot, the final survey incorporated questions which aimed to document staff views on key lessons learned and opinions on the future of offender management on their borough. A consistently strong response rate was achieved throughout the research. Wave 1 achieved a total of 51 completed responses (a 67% response rate); wave 2 achieved 67 responses (a 94% completion rate), wave 3 achieved 64 responses (89% response rate), and wave 4 achieved 66 responses (97% response rate; see table 2 for the numbers of Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09 Jul-09 Aug-09 Sep-09 Oct-09 Nov-09 Dec-09 Jan-10 Feb-10 Mar-10 Apr-10 May-10 Jun-10 Jul-10 Aug-10 Sep-10 Oct-10 Nov-10 Dec-10 Jan-11 Feb-11 Mar-11 First Diamond teams operational Staff survey W2 Staff survey W3 Staff survey W1 First year evaluation report drafted and published (June 2010) Final evaluation report written Staff interviews (Year 1) Staff interviews (Year 2) Offender interviews (Year 1) Offender interviews (Year 1) Ward and borough matching conducted Cohort generated - 6 month reoffending data analysed Control group generated Final PNC data run and analysed Staff survey W4 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09 Jul-09 Aug-09 Sep-09 Oct-09 Nov-09 Dec-09 Jan-10 Feb-10 Mar-10 Apr-10 May-10 Jun-10 Jul-10 Aug-10 Sep-10 Oct-10 Nov-10 Dec-10 Jan-11 Feb-11 Mar-11 First Diamond teams operational Staff survey W2 Staff survey W3 Staff survey W1 First year evaluation report drafted and published (June 2010) Final evaluation report written Staff interviews (Year 1) Staff interviews (Year 2) Offender interviews (Year 1) Offender interviews (Year 1) Ward and borough matching conducted Cohort generated - 6 month reoffending data analysed Control group generated Final PNC data run and analysed Staff survey W4 PNC Analysis Offender Interviews Staff Interviews Staff Survey 14 responses). The research team were unable to survey any members of staff who had left the DI team. Table 2: Staff survey responses by borough (2009-2010) Borough Wave 1 (March 2009) Wave 2 (Dec 2009) Wave 3 (May 2010) Wave 4 (Nov 2010) Croydon 7 11 10 9 Hackney 7 11 8 11 Lambeth 6 9 11 11 Lewisham 11 11 9 9 Newham 10 13 13 13 Southwark 10 11 10 13 Unspecified - 1 3 - Total 51 67 64 66 2. Staff interviews The staff surveys were complemented by interviews, which provided further context to aspects captured in the surveys. MVA, a market research company, commissioned by the SRAU, carried out 36 one-to-one semi-structured interviews with Diamond staff, from across the six boroughs, in each year. Interview schedules were developed to expand on a variety of issues raised in the first waves of staff surveys and to provide further context and insights into staff experiences of setting up and delivering the DI programme. The fieldwork was conducted in August 2009 and again in September and October 2010, with a focus on capturing change and learning over the course of the DI. A good mix of staff was achieved - the sample of staff selected for the interviews represented a range of job functions and included both strategic and operational staff from all six boroughs, along with Central Team members, as detailed in table 3. 15 Table 3: Staff interviewed, by role (2009-2010) Role title Year One (2009) Year Two (2010) Police Sergeant 6 6 Police Constable 8 6 Police Community Support Officer 5 5 Probation Officer 6 6 Local Authority staff 5 7 Strategic Leads 6 6 Total 36 36 3. Offender interviews The evaluation aimed to capture the perceptions and experiences of offenders engaged on the DI. This enables the evaluation to understand what motivates users to be involved in the DI, what their perceptions and experiences are about the DI and how their needs are being met. Offenders were nominated by the teams and interviewed in October 2009, and again in July 2010. Fifty-eight face-to-face semi-structured interviews were carried out with DI offenders across the six DI boroughs over the two years. See table 4 for a breakdown by borough. 16 Table 4: Offenders interviewed by borough (2009-2010) Borough Offenders interviewed (2009) Offenders interviewed (2010) Croydon 6 5 Hackney 3 4 Lambeth 4 5 Lewisham 6 6 Newham 5 5 Southwark 5 4 Total 29 29 The interviewees consisted of 49 males and 9 females. A total of 33 were White and 25 were from Black and minority ethnic groups; reflecting the demography of the parts of London covered, the diversity of both groups, in terms of nationality and country of birth, was notable. The ages of those interviewed in 2009 varied from 21 to 59, with a mean of 38. Those in 2010 varied from 21 years of age to 55, with an average age of 35. The criminal history of the offenders interviewed varied (with an average of 20 previous offences, (range 1 – 72) in year one and an average of 30 previous offences (range 1- 133) in year 2). DI teams arranged and ‘selected’ participant interviews and therefore the evaluation team was only able to interview offenders who were currently engaged with the scheme, not those who had declined, dropped-out or been removed from the DI. 4. PNC6 analysis In order to investigate the impact of the DI on crime, a quasi-experimental approach was used to compare the PNC-recorded offending of a cohort of 368 Diamond referrals, during a one-year period after prison release, against a statistically matched control group of offenders drawn from similar London wards. 6 The Police National Computer (PNC) records information relating to individuals for the use of the police and law enforcement, including details of all criminal convictions received. 17 This approach has allowed for comparisons to be made in terms of the binary ‘proven reoffending rate’ as well as more nuanced comparisons of survival rates, number of offences committed and offence seriousness. In addition, a second control group, matched to Diamond’s active participants has also been generated to explore impact. A detailed discussion of the approach taken, the matching processes used to generate the control groups and the reoffending measures employed is presented in chapter 4 and in appendices 1 and 9. 5. Economic impact MVA Consultancy were contracted to carry out a cost/benefit analysis of the Diamond Initiative. Using PNC data and a model based on Home Office research, the cost of crimes committed by the Diamond referral cohort during the year following prison release was calculated and compared to that of the control group. The costed-benefit of the Diamond Initiative in terms of crimes averted (i.e. the difference between the costed crime of the referral and control groups) was then compared against the cost of delivering the Diamond Initiative. Further methodological detail is provided in chapter 4. 18 Chapter 3: The process of delivering the Diamond Initiative - insights into implementation This section presents the challenges and learning relating to the set-up and continued implementation of the DI and, where appropriate, attempts to illustrate improvements over time. We know that these implementation issues are key to delivering programme integrity and the effects of these on outcomes should not be underestimated. This section is organised around the following topics: 1) Initial set up and how the DI matured over time. 2) Multi-agency working and meeting offender need. 3) The challenge of engaging with offenders. 4) Working across organisational cultures. 5) The offender’s experience of Diamond. 6) Gearing down and looking forward. 1) Initial set up and how the DI matured over time One of the strengths of the methodology employed was the ability to track progress and staff perceptions over the course of the DI. This section will explore the initial set-up and implementation of the DI alongside any improvements identified over time, to examine whether and how the scheme has matured over the two year pilot period. Staffing the Diamond Initiative The majority of the staff members surveyed in wave 4 (69%) had been in their role for over a year, with 85% having been in their position for at least six-months. This means that a large number of the views expressed in the survey were based on solid experience of working on Diamond. Staff expressed the feeling that team building was hindered at the beginning due to different agencies mobilising staff at different speeds. To illustrate, whilst police staff could be deployed rapidly (within a week); probation and local authorities required job descriptions and formal recruitment was necessary. 19 In the first-year report, an important issue arose in that police staff described how they were ‘volunteered’ or ‘posted’ into DI roles, which could impact on motivation and ‘buy-in’ to the Diamond ethos. Results from the second year indicate that this ‘posting’ issue had in the main been rectified. In wave 4, 85% of staff had stated they had joined DI through their own volition. None-the-less there were still cases where police recruitment was problematic; standard police deployment practices, which can tend to see officers as interchangeable resources, are perhaps not best suited to filling roles that require specific personal qualities, motivation and openness to new approaches. Two groups where most staff attrition was noted were police constables and Local Authority staff. Police constables due to the issues described above, and Local Authority staff due to the perception that they did not always have adequate skills to work in the teams. Team staff levels improved throughout the scheme (see figure 2), but the majority still described their teams as not fully staffed. So much so, that in Wave 4 less than 30% of staff perceived their team to be up to strength. This staffing issue in itself may have implications for effective implementation and therefore programme integrity. Figure 2: Is your DI team fully staffed? 7.8 11.9 10.9 28.8 Wave 1 Wave 2 Wave 3 Wave 4 Finally, as reported in the year-one report, there continued to be staff discussions during year two as to the balance of police and probation staff on the DI and who is more appropriate to lead such an offender management scheme. This issue will be revisited later in the chapter, where the split between working cultures will be examined in more detail. % answering ‘Yes’ 20 Training and developmental learning At the beginning of the DI in early 2009, one week training courses were delivered to all six DI teams: over the course of the DI this morphed into a two-day training scheme delivered on demand. Comparing the staff survey responses over the scheme, it is clear that staff became far more confident in their understanding of their DI role (45% wave 1, to 76% wave 4). Similarly, when asked about training, over the course of the scheme, staff were more positive that training enabled them to do their job effectively (see figure 3) - although a minority of staff still believed improvements could be made. Areas highlighted where additional training would be beneficial were predominantly practical in nature, such as housing processes and the law, dealing with victims, mental health, drugs treatment information, and offender communication strategies. Figure 3: To what extent do you agree that you have received sufficient training to do your job on the DI team effectively? 46.0 59.1 56.4 60.6 Wave 1 Wave 2 Wave 3 Wave 4 Key learning: Recruitment and staffing levels were a consistent challenge. Coordinating recruitment protocols is likely to be crucial when forming interdisciplinary teams. Staff attitudes, motivation and ‘buy-in’ directly affects implementation; seconding reluctant staff to offender management roles should be avoided. % who ‘agree’ or ‘strongly agree’ 21 An interesting cultural divide appeared in consideration of the training - between probation and Local Authority resettlement officers, who tended to focus on offender needs, compared to some police officers (and PCSOs), whose focus was more on Justice Reinvestment and a different approach to using criminal justice resources. Several respondents felt that police personnel were coming into the DI with less offender management experience. This gap needs to be adequately addressed in the training provided. ‘The police staff have no experience of this type of work and they are given almost no training. I feel that this is unfair on them and they do flounder for some months at the start’. (DI staff member) Interviews with police officers revealed that some thought they had no need for specific training, as a lot of what they needed to know was seen as ‘common sense’; this is perhaps reflective of cultural differences, for instance while probation officers tend to value their professional training and qualifications the emphasis within the police is perhaps more on experience and on-the-job learning. To illustrate, one police respondent said that motivational interviewing training was ‘fun’ but he has not used it since. None-the-less, nearly all (91%) of staff survey respondents at wave 4 reported that they had learnt something new whilst working on the DI. Reflecting on the contrast with their previous experiences in Safer Neighbourhoods and enforcement roles, many police personnel noted that their attitude towards offenders had changed and they had developed a greater understanding of offending behaviour, were more open-minded and less judgemental. This builds upon the year-one results where staff often highlighted the generally positive experience of being part the DI. This is an important point to focus on, and one that may sow the seeds for enhanced offender engagement in the future careers of those police officers that participated in DI. ‘Offenders do face challenges. Obstacles are put in their way (when they shouldn't be)… I am more open-minded. I am a better officer being involved in this project’. (Police, DI staff member) 22 Accommodation: getting the right space Lack of access to computing facilities, desk space, suitable office accommodation and cars were cited throughout the DI as barriers to smooth running of the scheme - although to a lesser degree in year-two than year-one. Furthermore, co-location of staff in team accommodation was once again identified as a key facilitator for delivering DI services, sharing information, and multi-agency working. This perceived benefit of co-location echoes previous offender management evaluations (Dawson, 2007). However, staff continued to report that there was a lack of adequate space to meet or interview their clients. To illustrate, one team was not allowed to bring offenders into their local authority owned premises as this was ‘against house rules’. To interview clients, the staff would use nearby cafés, which raised questions about client confidentiality and public safety. Other teams based within police stations had to see clients within interview rooms. This was seen by staff as increasing the possibilities of disengagement of offenders with the teams as they may not appreciate going to a police station, engaging with the police or wanting to re-visit any feelings of being in a police interview room. IT systems IT-related issues caused staff frustrations throughout the first year of the implementation, in particular with the LISARRT system - the bespoke system used by teams to monitor referrals from prison and record activities, outcomes and case-notes relating to service users. Indeed, staff often used paper records until LISARRT was fully operational. The second year interviews appear to indicate that staff were far more comfortable with the LISARRT system. However, there remained criticism, such as slow running, time consuming to complete, hard to navigate through and a difficulty understanding some of the terms used in the system. Interviews with the Central DI Strategic Team revealed that they continuously sought to improve LISARRT to meet the needs of the DI teams, (and to a lesser extent, of the evaluation). Whilst this was appreciated by staff, the feeling was that much greater user input, earlier on, could have led to a more rapid development and use of LISARRT. Key learning: Effective staff training, using continuous feedback for monitoring success, should be a core part of any offender management. 23 2) Multi-agency working and meeting offender need As outlined in the introduction, the assessment and the meeting of offender need is essential within offender management to enable desistance from crime. As such, a key question for the Diamond Initiative should be whether offender needs were effectively met, because if not, obtaining a positive impact upon crime will be very difficult. Staff were broadly positive when asked whether they thought resources were matched to offender need during their work with offenders - with a clear improvement being seen over the scheme, albeit with a minor drop in wave 4 (see figure 4). Figure 4: Do the resources aligned to the DI team match the needs of the offenders? 47.1 61.2 75.0 70.0 Wave 1 Wave 2 Wave 3 Wave 4 Key learning: An effective case management and referral system should be in place at the outset with only minor amendments subsequently made. Overall, despite initial problems, implementation matured over the course of the programme. Any future-roll out should anticipate and make contingency plans to address early implementation issues (e.g. equipment, buildings, staffing, IT systems). % answering ‘Yes’ 24 To explore this in more detail, the evaluation obtained OASys data on 147 clients eligible for the DI within the first year of the scheme. Table 5 illustrates the OASys needs identified within the DI group7. The DI group present high levels of need across the board, with education, training and employability, lifestyles and associates, relationships and attitudes being the most frequent needs; demonstrating the complex nature of this client group. The on-line staff survey throughout each wave has asked about staff perceptions of the different kinds of support involved in the DI. Piecing these two data sets together can reveal insights into whether or not offender need was being met. Table 5: Diamond client needs identified In terms of the services involved in Diamond, as demonstrated in Table 6, police and probation were (as expected) perceived to be centrally involved throughout the programme. This is comparable to year-one results. Outside of these, no other service was regularly perceived as being consistently involved, although the local authority, housing and a lesser degree drugs misuse services became more involved over time - something that should be seen as a real positive. It would appear that most services were 7 DI offenders present higher needs in all areas except alcohol misuse when compared to the overall national 2008 sample - albeit such a comparison has limitations insights can still be developed. MOJ 2009: Compendium of research and analysis on the offender assessment system. Eligible for DI ETE (education, training employability) 74.1% Lifestyle and Associates 70.7% Relationships 65.3% Attitudes 64.6% Thinking and Behaviour 62.6% Accommodation 54.4% Drug Misuse 59.4% Alcohol Misuse 27.9% Finance n/a Emotional well-being n/a Average number of needs per offender 4.8 25 ‘involved when required’ for example mental health, health and education (see appendix 12). As shown earlier these three services have a key role to play in addressing the problems facing persistent offenders. One could argue that given the high proportion of needs identified via the OASys data in education, training and employability, thinking and behaviour, accommodation and drug misuse (amongst others) that this support should be consistently involved in some way in the Diamond Initiative in order to effectively meet need. To illustrate, it was only in wave four of the survey that housing was consistently involved at a reasonable level (67%), before this, the picture was very different. It is unlikely that this client need would have been effectively met in the early phase of Diamond. Staff also highlighted the need for greater links and involvement from mental health, immigration and benefit services. ‘Housing is always an issue as we have very limited resources. Thus it is very difficult (if not impossible) for our clients to source private accommodation as they cannot raise a deposit. Few qualify for supported housing’. (Staff member) ‘It’s difficult to access support from the local community mental health teams, particularly for dual diagnosis client’. (Staff member) 26 Table 6: Staff perceptions of services involved in the Diamond over the course of the Initiative8 Furthermore, given the importance of the prison service within the Diamond Initiative - it should be noted that prison service involvement throughout the scheme appears piecemeal without real improvement over the course of the scheme (wave 1, 28% to wave 4, 26% consistently involved). This is an issue previous evaluations have identified as a major challenge within offender management (see Dawson 2007, in terms of prisons sharing information or offender release dates) - it would appear especially pertinent given the aims and clientele of the DI to foster better and more effective prison links. Overall, there were improvements in multi-agency working over the duration of the scheme, although even in year-two it would appear that support gaps remained. The key question is whether the improvements identified would have been adequate to be able to effectively meet offender need. This is a question that may resonate when examining reoffending, given that the evidence base is clear that addressing need is a key ingredient in reducing crime. 8 Those services in green indicate a good progression, whilst those which are amber could be improved. Red indicates a poor improvement on services which appear to be areas of need amongst the clients. % Consistently Involved W1 W2 W3 W4 Police 98 97 97 97 Probation 84 97 91 86 Prison 28 18 16 26 Community Safety Partnership 4 9 9 10 Local Authority 45 60 70 70 Voluntary Groups 4 19 19 12 Housing 18 34 44 67 Education 2 15 14 5 Youth Offending Teams 0 2 3 0 Drug Intervention/ Drug Action Team 24 19 41 39 Mental health 2 2 11 9 Health 4 2 3 3 27 3) The challenge of engaging with offenders Before supportive intervention could be delivered, teams faced the challenge of identifying, locating and engaging with eligible offenders in their area. As discussed later (chapter 4), the referral systems used by teams to target eligible offenders were often problematic and left teams with extensive and time consuming work in terms of researching and tracking down potential clients, searching databases, making enquiries with other agencies and making repeat visits to potential addresses - work which only sometimes resulted in the Diamond offer being made. Overall, throughout the initiative, case loads were generally felt by the teams to be lighter than expected, which had a number of implications. On the one hand, staff appreciated the time and resource this afforded to dedicate to clients on the books, on the other, teams felt a degree of pressure to get the numbers up, which in turn led to:  Taking on referrals from other agencies (who tended to see Diamond as ‘resource rich’), including difficult cases which they occasionally felt were ‘palmed off’ on them.  Increasing further the amount of time and effort spent on researching and locating referrals, to make sure no potential clients got missed.  Pressuring the initial Diamond model to expand to other wards and other groups of offenders - which on occasions, lead to some discomfort in teams, for instance in taking on sexual offenders from PPO teams.  Perhaps, in some cases, leading to ‘selling’ the scheme to offenders at the expense of emphasising the behaviour change required on their part. Key learning: The mix of help offered may not have been effectively matched to client need. Provision should be designed based on local analysis of the needs of the client group, but anticipation of standard problems, based on previous research should be built in at the planning stage. Unmet client need is as much a strategic problem as it is a practical problem. 28 Owing to this pressure and because of the relatively prescriptive eligibility criteria, teams rarely reported disagreements over selection or de-selection decisions. As a voluntary scheme, one of the main challenges faced by teams was persuading offenders to engage. Overall (as discussed later), engagement rates were relatively encouraging, however it is clear that staff came across clients with a range of attitudes and motivations - from those apparently genuinely committed to changing their ways: ‘They realise that they don’t want to continue (down) the road they were on. They want change in their lives. Most of them are in their twenties and realise that where they were heading goes to nowhere’. (Staff survey respondent W4) ‘To be quite honest I was sick and tired of the way I was going, every sort of eight months. I was going into prison for six months. Doing six months there. Coming out. Going back to work and doing exactly the same thing. It is the cycle I just kept doing’. (Offender) To those with perhaps more self interested motivations: ‘The majority are after what they can get from us, often not keeping their side of the agreement’. (Staff survey respondent W4) ‘I signed up to get re-housed and then start putting my life back together again…you know within a couple of months I was looking to colleges’. (Diamond offender) To those, perceived by teams as unready or unable to recognise the benefits of the offer made: ‘Offenders with mental health and drugs issues often have problems that make it difficult to see that Diamond is a genuine chance to move forward in a good way’. (Staff survey respondent W4) 29 It is clear that as a public service offer with a strong police element, Diamond raised suspicions among many offenders; this has the potential to suppress engagement rates and should be addressed in the work. Even offenders who engaged with the scheme reported initial apprehensions, ‘I was a bit suspicious…they’re coppers and, you know…they’re trying to help you and I’m not used to coppers trying to help me, I’m used to ‘em beating me up and chucking me in gaol… yeah I was a little bit suspicious’. (Offender) Although it is equally clear that through a gradual engagement process, the appeal of the support on offer and by Diamond staff demonstrating commitment and ‘good faith’ these apprehensions could be overcome: ‘[The Diamond caseworker] came across as a person that really cares about that individual person…and maybe that’s what it is I need, you know someone who seems genuine and genuinely wants to help…I felt I could talk to him you know and really feel comfortable’. (Offender) ‘At the time I didn’t have any other option. I needed the support. I like the security of dealing with the police and being inside the law. I have a criminal record so now I need help to find jobs’. (Offender) Staff frequently reported that engaging offenders was facilitated by initial contact in prison, prior to release. This allowed team members to address suspicions and start to break down barriers at an early stage, made it more difficult for offenders to avoid contact and encouraged them to consider the future and begin to make plans for release; this is of particular note given the scope for greater prison involvement reported in the previous section. 30 ‘Attending legal visits within the HMP environment. This enables the Diamond team member to convince the individual (away from any distractions) of the wholesale benefits of engaging with Diamond. The prison environment provides a 'captive audience’’. (Staff survey respondent, W4) 4) Working across organisational cultures One of the core themes to emerge from the staff survey and interviews was the challenge in effectively integrating police and non-police staff within the DI. This seems to have resulted in a number of working tensions or culture clashes, especially during the second year of the scheme. The survey indicated that the majority of staff perceived that generally police and non- police worked together well, although even this saw a decrease over time and an increase in mixed responses toward the end of the programme (see figure 5). It was only when we were able to delve deeper that it becomes apparent that there were working tensions running through the scheme; these tensions reflect themselves in a number of areas. Key learning: Staff reported positively on the practice of engaging with offenders in prison. Given the evident problems with the flow of information between custody and community and the importance of the first weeks and months following release, an effective link up with local prisons is a key to success. The challenge of increasing numbers was felt throughout the DI; the scale of provision needs to be matched to the likely supply of eligible clients for the defined catchment area. Diamond’s (original) focus on certain wards was found to be restrictive. 31 Figure 5: To what extent do you agree that Police officers/PCSOs and non-police staff generally work well together? 82.4 79.1 85.9 75.7 15.8 20.9 24.3 12.5 W1 W2 W3 W4 Agree Mixed views / no For the majority of staff (75% in wave 4) working on DI was what they expected; working with partners to support offenders. However, some police staff felt they had been ‘mis- sold’ the DI and were expecting more enforcement activities as opposed to a predominantly supportive role. Some went further stating that, as a police officer, they felt they had ‘their wings clipped’. Some police officers also pointed out that they did not want to be seen as ‘protecting’ the client, were sceptical of rehabilitation and were surprised that the offenders appeared to have more ‘power’ than they would expect them to have within typical engagement with the police. Indeed, Police staff noted that some of their colleagues saw the DI as more ‘social work’ and ‘not real policing’. The police were seen as ‘task orientated’ and ‘authoritarian’, providing the ‘get up and go’ compared to the ‘liberal’ and ‘laid back’ nature of the probation and Local Authority staff, where there was also the perception (among police) that a series of meetings was required before carrying out a task. ‘MPS officers and staff tend to have a robust approach to activity based working; with efforts to undertake tasks to achieve outcomes. Some of our professional partners do not display the same level of urgency’. (Police staff) % 32 As previously raised, the majority of staff perceived DI as a being police led, although many non-police staff questioned this over the duration of the programme. Interviews with the Central DI Strategic Team emphasised that the police numbers allowed for staff to be available if and when needed, providing an ‘out of office hours’ service, that would not necessarily be provided by non-police staff. Partner agency staff felt the probation service would have been better placed to lead because of their skills and prior experience of working with these client groups, both of which were perceived to be lacking amongst police officers. Another issue between staff that was never fully resolved was that of the police uniform. Many non-police staff, perceived that wearing of the police uniform, could alienate and anger clients, making trusting relationships harder to build. ‘One of the PCs was working with a case … at the start, he met him in plain clothes, but then he started meeting him in uniform. And, he was okay with that, up until his family started saying, “Why are you working with a copper?” And then, from that stage, he’s really withdrawn from working with the PC, but he will still work with me, so I think that’s quite telling’. (Probation staff) One police staff member commented that the uniform is not routinely worn, except during compliance visits (for Community Payback offenders) or undertaking a task where it is deemed appropriate to be in uniform. Police staff did also comment that sometimes the uniform would be used to reinforce boundaries or if they suspect illegal behaviour of a client. This was not seen to be problematic by police staff. Interviews with the offenders indicated that the majority had no real issue with police staff wearing a uniform. ‘When they get to know you, they don’t see the uniform.’ (Police staff) ‘…she’s always in uniform… that’s fine.’ (Offender) In summary, it is apparent that there were some working/ cultural tensions across agencies within the DI - some non-police staff going as far as to argue that police officers should undergo an interview process which assesses their skills and experiences 33 necessary for the DI before joining the team. The feasibility of such interviews is not the issue here in itself, the consideration being that many staff felt that such action was required illustrating the unresolved divisions between staff working on the DI. 5) The offenders’ experience of Diamond Almost without exception, the Diamond service users interviewed reported positively on their experience of the scheme. It is important to note however, that these offenders were, by necessity, those who had agreed to participate in DI, had not ‘dropped-out’ and, in all likelihood, were those that teams (who facilitated access) felt were most suitable for interview. With this caveat acknowledged, it is encouraging to report that those interviewed frequently gave positive and optimistic recent life-narratives; charting a trajectory from the anxiety and stress of the period immediately after prison-release to a more stable and fulfilled current state and projecting toward future progress. In many cases offenders saw the support they received from Diamond as integral to this progression, as one interviewee noted; ‘I don’t know what would have happened if I hadn’t come to Diamond team…they’ve turned my life around basically and for all I know I could be back inside now doing a very long stretch…now I’m about to…start voluntary work, my life possibly has a future’. (Offender) Some interviewees reported negative previous experiences of statutory and non- statutory support and contrasted this with Diamond’s joined-up working and needs- based approach. Key learning: Although teams generally worked well across disciplinary boundaries, tensions linked working culture between the police and other agencies were never fully reconciled. There is still work required to establish an integrated approach. 34 ‘…back then…I was offered help, you know, probation and that but…I think the help that was available wasn’t really what I needed, wasn’t specific to me and… the left hand didn’t know what the right hand was doing. So I was really disillusioned with them, had no faith in them and just had an attitude of ‘well… I’m gonna do things my way’, you know?...and that always led me to using drugs and ending up back in prison’. (Offender) The Diamond ethos was seen as collaborative and non-directive, with the voluntary model and the approach taken by Diamond caseworkers (Lead Liaison Officers (LLOs)) allowing participants to feel in control of addressing their own needs. ‘The big difference is probation; you are given (as)… a punishment not as something to help you. Whereas Diamond is voluntary, you are here because you want the help, no one is making you come here and I am actually getting the help that you need through Diamond, whereas probation you are in for five, ten minutes and you don’t get the help you need, it is more a chore…here I can come and say right I have got a drink problem and they will sort that out…they will do anything for you whereas the other ones… you got to go there or you go to prison whereas here you either come here or you don’t, it is entirely up to yourself’. (Offender) Respondents reflected particularly positively on the quality of the relationships developed with Lead Liaison Officers. It was frequently clear that caseworkers had won the trust of their clients by demonstrating commitment and remaining non-judgemental. ‘…he’d kick me up the ass if I do anything wrong… but no he doesn’t criticise me he just tells me where I’m going wrong…‘cos I have a lot of things like my bad drug and alcohol issues since I was very young… (he) has known me quite a while, sometimes… he says, ‘have you got enough food?’… he helped me out getting a hostel…they’re not judgemental… they ain’t gonna criticise me for what I’ve done in my past, everything is a clean slate and we start from now…(they) ain’t gonna pick you up for what you’ve done in the past…’ (Offender) 35 As noted elsewhere and reported in year one, Diamond offenders often reported apprehensions about getting involved in a scheme with strong police involvement. While some retained reservations, for others the change in attitude towards the police, as a result of the positive interactions with Diamond police officers was striking. ‘These guys (Diamond) have given me a different outlook…I never would have phoned the police. Two weeks ago I rang them due to a disturbance outside of (my) flat. Because of Diamond I realise there is care there and they understand me and I can be honest with them - feel like they'll help me now and not get me into trouble’. (Offender) Overall, and in line with the desistance literature, it is evident that Diamond teams managed to engage with individuals who appear committed to life-change but who present a range of needs and face a variety of obstacles in doing so. It is clear that the support provided by Diamond teams, both practical and motivational, was appreciated by participants and that honest and trusting relationships had developed, sometimes from an apprehensive starting point. The extent to which these positive reactions correspond with measurable behaviour changes, in terms of reoffending rates, is explored in detail in chapter 4. 6) Gearing down and looking forward From the outset, the DI had a two year life-span and was always due to reach completion April 2011. None-the-less staff, especially within wave 4 of the survey and the second year interviews, expressed real uncertainly about the future of DI and discussed that they had heard ‘rumours’ about the next steps (or not) of the DI. This uncertainty would have appeared to have had a negative influence on staff morale. ‘I feel that the level of morale within the team has deteriorated due to the uncertainty about the future of the project and the team not being given any answers. This has meant that a number of staff having left and others are looking to leave, thus affecting the atmosphere’. (DI staff). 36 Wave four of the staff survey explicitly asked about team morale, as can be seen in figure 6, there was a fair distribution - although at least one third described their morale as low or very low at this point and a third a potentially apathetic 'neither high or low'. Figure 6: How would you rate the current level of morale in your team? (Wave 4) 10.6 18.2 37.9 18.2 15.1 Very High High Neither high nor low Low Very low Generally, Local Authority workers and probation staff expected to go back to their original employment - although often emphasising their preference of remaining on the DI if there was the possibility. Police staff feared that they may be ‘placed’ somewhere else so many were already looking for attractive positions within the MPS. It is fair to say that most staff and offenders were positive about the need for DI to continue in some manner and described many anecdotal benefits of the scheme, such as perceptions of reduced offending, accommodation, training and the provision of other support. ‘Yes, thank you very much. It’s fantastic and I hope that they keep, the government keep funding the project and I’d love it if people come and use it because I’ve been, I’d recommend it’. (Offender) ‘In my opinion those offenders who we have helped to find work or to get onto training courses have been less likely to re-offend as they are occupying their time and earning’. (Police) % 37 ‘This project has proved really useful to both offenders and the wider community and victims of crime. Something needs to be put in place to carry this work on’. (Local Authority staff) Even though the scheme was viewed positively, it was generally accepted that the Comprehensive Spending Review had altered the overall landscape and that there would not be funding for the DI in the future - but were hopeful that the work they had conducted would not go to waste and that DI may survive in one shape or another. ‘I don’t think it will carry on as expensive as it is now, there is just not the money… but I hope that there is some sort of reduced team, maybe a couple of police, a couple of probation officers and just refer people on the things that could help them’. (Police staff) A recurring theme at this point was the frustration that the central team had given local staff no indication as to whether the DI would continue and would have wanted greater levels of information. However, to be fair, the Central Team themselves were not in a position to be able to predict what the future was for the scheme given of considerable flux occurring throughout the public sector during the final months of the scheme. Key learning: Pilot initiatives suffer from uncertainty. Staff feel this uncertainty. Strategies and contingencies for the end of an initiative should be considered in programme delivery. 38 Chapter 4: The impact of the Diamond Initiative - what do we know about reducing offending? In the year-one evaluation we reported that 28% of Diamond referrals had reoffended within six-months of prison release. A tentative comparison was made against an historic sample of less-than-12-month sentence offenders from London9, which suggested that fewer Diamond referrals had reoffended than might otherwise be expected. Whilst this appeared promising, the comparison group was historic so likely to be affected by overall crime trends. In year-two we are able to examine the impact of the DI on offending with much more rigour, over a full year follow-up period, in more detail and in comparison to a robust and carefully matched control group. In addition to examining impact, this section provides wider insights from the profile of the referral group, assesses the response to the voluntary offer, and highlights potentially promising target groups and approaches. Evaluating the impact of a local offender management initiative Approach In the absence of opportunities for randomisation, a quasi-experimental approach was considered the most robust feasible option for investigating the impact of the Diamond Initiative on offending (Dawson & Williams, 2009). Conviction data, drawn from the Police National Computer (PNC), has been used to compare the ‘proven reoffending’10 of an ‘evaluation cohort’ of less-than-12-month sentence offenders, referred to the Diamond Initiative during the first nine months of the pilot11, against a well matched control group drawn from similar London wards. 9 Those released in Q1 2002 and dealt with by the MPS. 10 The Proven Reoffending rate is the proportion of the cohort for which PNC records one or more criminal offence, committed during the follow-up period, which resulted in a conviction at court. 11 Diamond teams became operational between January and March 2009, cohort offenders have registration and release dates prior to 01/10/2009. 39 The Evaluation Cohort (‘all those referred’) The Diamond Initiative was limited in its geographical coverage to a set of wards (13 at the outset) within six London boroughs. All offenders leaving prison, having served less- than-12-month sentences, and returning to live within these wards were eligible to receive the offer of Diamond assistance12, these individuals were referred to local Diamond teams by prisons and given records on the LISARRT (case management) system. These offender records have been drawn together to form the ‘evaluation cohort’. In order to allow for a full year follow-up period for reoffending to be monitored (plus additional time for offences to result in convictions), only those offenders with prison release dates prior to mid-October 2009 have been included in the cohort13; a total of 368 individuals. The Control Cohort (or who are we comparing DI offenders to?) In order to generate a robust comparison cohort (that, in the absence of Diamond could be expected to exhibit equivalent offending to the referral group), the evaluation set out to identify a group of offenders who lived in the parts of London most similar to the Diamond wards and who would have been referred to Diamond if it had been in operation in their area. A borough and ward level matching process was used to ensure that, in addition to having served similar sentences, within a similar timeframe, the offenders selected for the control group also lived in the parts of London most similar, in theoretically relevant ways, to the Diamond pilot wards (a detailed description of the matching process is provided in appendix 1). Once a control group had been compiled, the quality of the match was verified at an individual level. It is important to note that this approach is both theoretically rigorous14 and that the composition of the control group generated is highly comparable to that of the evaluation cohort. In terms of demographics, offending history, risk and need, the 12 Other eligibility criteria relating to primary management by other agencies, outstanding court cases for serious offending and deportation were also applied by the teams, after the initial referral. 13 This follow up period allows for a minimum of 420 days to have elapsed between each offender being released from prison and the extraction of PNC data in mid December 2010 i.e. a full year follow up period (12 x 30 day months = 360 days), plus a further two months (60 days) to allow for some lag in offences resulting in recorded convictions. For some analyses a sub-cohort of those with 540 days (18 months) of available data has been used. 14 Achieves a level 4 on the Maryland Scale. 40 groups are extremely well matched (see comparative statistics provided in appendix 4). OGRS315 predicts that 44% of the evaluation cohort and 43% of the control would be re- sanctioned within one year of release, providing a clear indication that (in the absence of Diamond) the two groups could be expected to exhibit similar levels of reoffending during the following-up year. ‘Intention to treat’ (ITT) (why compare the referral group?) It is important to note that the evaluation cohort was formed from Diamond referrals not from programme participants; given that Diamond participation was voluntary (and that we know from previous research that willingness to participate can lead to more positive outcomes (Hollin et al., 2008)), forming the cohort from those referred (regardless of whether they agreed to participate or not) allows for motivational impacts on reoffending to be controlled for. Furthermore, while it is possible to form a comparison group of those who would have been referred, had the scheme been available in their area, it is not possible to determine which of those would have agreed to participate in the scheme, thus making it necessary to match at the referral stage in order to preserve the comparability of the match. This ‘intention to treat’ (ITT) approach entails that not all of those included in the evaluation group will necessarily have received the ‘treatment’ (i.e. Diamond support) and therefore that any impact (on reoffending or other outcomes) may be ‘diluted’ amongst non-participant referrals within the cohort. This can be considered a sterner test of impact (i.e. did it have a measurable impact the target population, not just on those who took part? (Colledge et al.,1999)) and does not preclude further analysis on the participating sub-section of the cohort, as explored later in this chapter. Referral outcomes Analysis of LISARRT records reveals that fewer referrals (cohort members) actually received Diamond support than had been anticipated. In addition to those who accepted or refused Diamond support, a substantial proportion of the referral group (45%) never 15 The Offender Group Reoffending Scale, version 3 is a predictor of re-offending based on age, gender and criminal history, its accuracy is validated and it is used widely in offender management. (Howard et al, 2009) 41 received the Diamond offer, either because they could not be located, or because, following research by the teams, they were deemed to be ineligible (a full breakdown of referral outcomes is included in appendix 5) - in light of these findings it is necessary to question whether adequate links were established between Diamond teams and their feeder prisons. In terms of the viability of the voluntary Diamond offer, it is encouraging to note that three in five of those who did receive the offer of support, agreed to participate (this differs from the high acceptance rate quoted in year one and reflects improvements in recording within the LISARRT system allowing a more accurate analysis of referral outcomes). Overall, only one in three of the evaluation cohort actually started on the Diamond Initiative and only about half of those ‘completed’ a period of engagement16, without dropping-out or being removed from the scheme (see Figure 7 and appendix 5 for a full breakdown). Methodologically while lower than expected participation rates, within the cohort, may make impact more difficult to detect, it does not in itself undermine the validity of the quasi-experimental approach. Figure 7: Evaluation cohort referral outcomes Engaged - completed / continuing, 16.8% Engaged - dropped out, 16.0% Refused, 21.7% Unable to trace / contact, 17.1% Ineligible, 28.3% Base: Evaluation Cohort 368 - (all referrals) 16 Those in the ‘completer / continuer’ category had undergone a period of engagement with teams and had agreed for support to be scaled back, or were continuing on the scheme at the time of coding in October 2010. 42 Did the Diamond Initiative reduce reoffending? In order to test the hypothesis that the Diamond Initiative reduced reoffending among those referred to the scheme, PNC data has been used to calculate ‘proven reoffending’ rates over a one-year period following prison release, for both the evaluation and control groups. Comparing reoffending for the evaluation and control cohorts For this comparison, an offender is deemed to have reoffended if PNC data shows that they committed an offence within one year (360 days) of release from prison, which resulted in a conviction at court within 14 months (420 days) of release. Technically, this method differs slightly from that used by the Ministry of Justice to calculate official reoffending rates for the national prison release cohort, where an 18-month conviction follow-up period is permitted17. Given the time-frame for this analysis, reducing the follow- up period to 14 months allows for a larger and better matched sample of offenders to be included in the analysis, and thus for a more robust comparison to be made18. Key result Of the 368 offenders referred to Diamond (the evaluation cohort), 156 (42.4%) reoffended within 12 months of release from prison. From the control group (similar offenders from 17 Reoffending of adults: results from the 2008 cohort, England and Wales. Ministry of Justice Statistics bulletin. March 2010 18 In line with standard practice, breach offences have been excluded from all calculations (however where the breach accompanies a conviction for a new offence, that offence can be counted as a re-offence). Key Learning: The process of identifying and engaging with potential participants was severely hampered by the working of the referral system. Improving the flow of information between custody and community is likely to be crucial for delivering effective offender management. 60% of those offered the scheme agreed to participate - this indicates substantial demand for resettlement help and support. 43 similar localities), 136 out of 327 offenders (41.6%) reoffended within 12 months. This is the headline finding of the Diamond evaluation and would appear to demonstrate that the DI had no impact on offending. Comparisons with national reoffending rates The proven reoffending rate for the national (England and Wales) less-than-12-month sentence cohort is 61.1% (2008, quarter 1 prison releases), this has varied between 58.0% and 64.7% since 2000 (Ministry of Justice, 2010a). In order to compare the Diamond referral group against these national figures, adjustments are required in order to align counting methods used - most importantly the ‘waiting-period’ permitted for offences to result in convictions must be extended to 18 months (with the consequence of reducing the size of the evaluation cohort available for analysis)19. These adjustments produce a proven reoffending rate for the evaluation cohort of 47.0%; which is significantly below the rate for the national cohort20. It is evident that the Diamond referral group (and the control) had substantially less extensive criminal histories than the national less-than-12-month cohort. On average Diamond referrals had been convicted of 20.7 previous offences (control - 20.4), the average for less-than-12- month-sentence offenders in male, local prisons is 48.0 previous offences (Ministry of Justice, 2010b). The reasons why those less-than-12-month sentence offenders who reside in some of London’s most problematic wards (and therefore form the evaluation and control groups in this study), have substantially shorter offending histories and lower reoffending rates than the national cohort, are not fully understood. One potentially relevant observation is that, reflecting the make-up of the population in these wards, 43.8% of the referral group (and 41.9% of the control) were born outside of the UK. While the equivalent proportion for the national, less-than-12-month cohort is not known, given that only 11% of the UK 19 Extending conviction follow up period to 18 months and excluding deportees results in an ‘18-month evaluation cohort’ of 217 offenders. 20 Z=4.108, statistically significant at 99% confidence level - no equivalent measure is available for the control group as it is not possible to identify deportees within the control group 44 population are non-UK born21, it is reasonable to assume that it is substantially below this level. Our data (see appendix 6) suggests that non-UK-born offenders reoffend less and have fewer previous recorded offences than those born in the UK - in light of this, the fact that the Diamond cohort (and its ward matched equivalent) is likely to contain a greater proportion born overseas than the national cohort, may go some way to explain the differences in proven re-offending rates found in this study. Close-up on crime - did the Diamond Initiative have any other effect on offending? The proven reoffending rate, reported above, is a blunt measure of offending behaviour, it does not, for instance, take into account the amount of time offenders ‘survived’ before reoffending, the number of offences committed or the ‘seriousness’ or ‘type’ of those offences. This section delves deeper into the data to examine these more nuanced measures of reoffending. ‘Survival’ - how long did it take offenders to reoffend after prison release? In order to allow valid comparisons to be made with national data, three, six, nine and 12 month reoffending rates have been calculated for the 18-month waiting period sub- cohorts of the evaluation and control groups22. As can be seen in figure 8 (below), the 21 Population by Country of Birth and Nationality (April 2009 to March 2010). Office of National Statistics 22 In order to make like-for-like comparisons, deportees (who cannot be identified within the control group) have not been excluded. Reoffending rates therefore differ slightly from those reported elsewhere. Reoffending rates are Key Learning: The level of reoffending within the Diamond referral group was below that of the national less-than-12-month sentence cohort. This should be understood in terms of differences in the make-up of the referral group compared with the national sample. This suggests the need for regional / local benchmarking in order to provide a framework for assessing the differences and effectiveness of local offender management approaches. 45 proportion of the referral and control groups who had reoffended after each period was generally consistent23, providing no indication of a Diamond impact on the amount of time offenders remained ‘offence-free’. Figure 8: Proven reoffending after 3, 6, 9 and 12 months 44.2 21.5 28.8 39.3 29.6 44.0 52.1 58.0 39.9 30.0 17.6 45.0 3 Month 6 month 9 month 12 month Evaluation Control National less than 12 month cohort (Q1 2000) Within all cohorts, those who reoffended during the follow-up year were most likely to first do so soon after release - for instance, 17.6% of the evaluation group first reoffended within three months of release compared with just 4.3% during months 10 to 12. Again, both the evaluation and control cohorts were markedly below national levels after each of the follow-up periods. The amount of reoffending The (full) Diamond referral group committed a total of 556 offences during the one year follow-up period (excluding breaches) at an average of 1.5 offences each. The control group committed 446 offences - an average of 1.4 offences each. This difference is not statistically significant, indicating that the Diamond Initiative does not appear to have an based on offences committed during that follow-up period, which resulted in a conviction within the period or within the subsequent six months. 23 4% difference between evaluation and control after 3 months - not statistically significant % Key Learning: Those who reoffend tend to do so quickly after prison release; this indicates a ‘window of opportunity’ for supportive engagement immediately following (or commencing before) prison release. 46 impact on the number of offences committed by the referral cohort during the year following release. The average number of offences committed by the two cohorts in each of the 12 months following, and 24 months prior to prison release, is shown in figure 9. The similarity of the offending pattern between the two groups, both before and after release is evident - reinforcing both the degree of similarity between the two groups and further illustrating the ‘no-impact’ finding. Figure 9: Mean offences per month during two years prior to, and year following, prison release24 0 0.1 0.2 0.3 0.4 0.5 Month -24 Month -23 Month -22 Month -21 Month -20 Month -19 Month -18 Month -17 Month -16 Month -15 Month -14 Month -13 Month -12 Month -11 Month -10 Month -9 Month -8 Month -7 Month -6 Month -5 Month -4 Month -3 Month -2 Month -1 Month 1 Month 2 Month 3 Month 4 Month 5 Month 6 Month 7 Month 8 Month 9 Month 10 Month 11 Month 12 Evaluation Control ‘Seriousness’ of reoffending Mason et al. (2007) provide a structure for categorising criminal offences into 10 ‘seriousness’ groups based on the average severity of sentences handed down across England and Wales. Figure 10 shows the proportions of each cohort whose most serious offence, post-release, fell into each of five (combined) seriousness bands. As shown, the seriousness profile of the two groups’ post-release offending is very similar, once again indicating no impact of Diamond referral. 24 Greater monthly fluctuations for the control group may be a product of multi-counting. Prison release Mean offences 47 Figure 10: Most serious post-release offence category 2.6 13.5 34.6 42.9 2.6 3.8 0.7 13.2 33.8 41.2 2.2 8.8 1st or 2nd 3rd or 4th 5th or 6th 7th or 8th 9th or 10th No categorised offence Evaluation Control Base: All those who reoffended E=156, C=136 As shown, more than 40% of those who reoffended committed offences which fell into the 7th and 8th seriousness categories, these bands include crimes such as drug possession, common assault and shoplifting25, providing insights into the kinds of issues offenders may be facing (e.g. substance abuse) but also illustrating the relatively ‘low-level’ nature of the groups’ offending. Given that this type of offending tends to be less ‘costly’ than more serious crime there are implications for the cost-effectiveness of targeting resources at this group. As can be seen in figure 11, the change in seriousness of offending (comparing periods before and after release), is also relatively comparable between the two groups. There is therefore no indication that referral to Diamond impacted on the severity of the offences committed. 25 Examples of offences within each category are provided in appendix 8. Most serious Least serious % 48 Figure 11: Seriousness of most serious offence post-release, compared to most serious offence in two years prior to release 50.0 29.5 14.7 5.8 52.2 24.3 12.5 11.0 Decrease Same Increase No categorised offence Evaluation Control Base: All those who reoffended E=156, C=136 Reoffending by crime type In order to investigate whether Diamond brought about a reduction in any type of reoffending, comparisons have been made between the sub-sections of the two cohorts that had committed offences, within various crime-type categories, during the two years prior to prison release. % Key Learning: No difference has been found between the Diamond referral group and the matched control in terms of the proportion who reoffended, the speed at which they reoffended, the number of offences they committed or the seriousness of those offences. The two groups of offenders exhibited highly similar offending careers, which continued to be comparable after prison release, despite the fact that the evaluation group were referred to the Diamond Initiative. 49 Figure 12 shows the proportion of each previous offending-type subset that reoffended within that offence category, in the year following release. As illustrated, within the evaluation group, fewer recent violent offenders and those who had recently committed offences against property (criminal damage) reoffended, within those crime types, compared to the control. The reverse is true for recent acquisitive and drugs offenders, where more from the evaluation group reoffended. It should be noted however, that owing to the comparatively small base sizes none of these differences are statistically significant. It may be of note, that the types of offences in which the evaluation group performed ‘better’ than the control tended to cluster around ‘disorderly’ / or ‘impulsive’ types of offending (criminal damage, violence, public order and assaults on police26). This provides a tentative indication that Diamond may have had a positive effect on disrupting the types of behaviours that result in these types of violent / disorderly offending, although the mechanisms through which this may have occurred are not currently clear. Conversely the evaluation group performed worse within drug and theft offences. 26 The ‘Offences relating to Police, Courts and Prisons’ category is principally made up of breach offences (within which reoffending was similar for the evaluation and control groups) and assaults on police, (where reoffending was lower for the evaluation group). Key Learning: There is tentative evidence to suggest that the Diamond Initiative may have reduced the degree to which violent or ‘disorderly’ offenders reoffended within these crime types. 50 Figure 12: Percentage of offenders who had committed a recent previous offence within crime-type - that reoffended, within that crime-type, within one year of release Checking for impact among Diamond participants As discussed earlier, one limitation of the ITT approach is that any impact of the initiative on reoffending may be diluted by non-participants within the evaluation group. It is possible however to separate out those who actually engaged with Diamond teams and examine their characteristics and reoffending outcomes. As shown in table 7, those who took part in Diamond (i.e. those ‘engaged’) actually had a higher reoffending rate than those who refused (or did not participate for other reasons)27. This does not imply however that Diamond had a negative effect on their offending behaviour. Analysis reveals that the engaged group were not a random selection of those referred to (or offered) Diamond, they had higher levels of recent previous offending, more identified (OASys) needs, and higher predicted reoffending (OGRS) than those who refused (see appendix 8). Actual reoffending among the participant group was accurately 27 They also committed more offences and tended to commit offences in more serious offence categories - ‘Engagers’ committed an average of 2.1 offences in year following released compared with 1.7 for ‘refusers’. 29.8% of ‘engagers’ committed a category 1 to 6 offence compared with 17.5% of ‘refusers’. 12.5 12.4 26.3 30.0 7.5 20.3 50.3 29.1 23.8 21.8 29.0 32.6 9.7 14.5 42.1 16.4 Offences against property (criminal damage etc) Offences against the person (violence) Offences relating to police / courts / prisons (breach offences etc) Public disorder offences Fraud and kindred Miscellaneous offences (inc. driving offences) Theft and kindred (acquisitive offences) Drugs offences Evaluation Control % 51 predicted by OGRS, suggesting that nothing occurred during the follow-up year to alter the participant group’s offending (either positively or negatively) from what would be expected for a group of this profile. Figure 7: Actual and predicted reoffending by cohort sub-group Engaged Refused Not Eligible Unable to locate Full Evaluation Cohort N 121 80 104 63 368 Proven reoffending (actual) % 50.4 40.0 32.7 46.0 42.4 Predicted reoffending % 49.5 45.7 35.6 46.2 44.2 In designing the ITT approach it was anticipated that agreement to participate and reduced reoffending rates would correspond, based on underlying factors linked to offender motivation. Given that this assumption has not been borne out (voluntary participation was in fact linked to higher, not lower, reoffending rates) it is appropriate to supplement the results reported so far with a second test for impact, which compares the proven reoffending rate of scheme participants (only) against a control group matched at the offender-level (drawn from the original geographically matched control sample). As described in detail in appendix 9, a new control group was formed by extracting the most similar counterpart for each Diamond participant, from within the original control sample.28 28 on a range of criteria including age, OGRS score and criminal career Key Learning: Those who accepted the offer of support were at greater risk of reoffending and presented more needs associated with criminal behaviour, than those who refused. The demand for resettlement support therefore appears to be need-driven and the voluntary model appears well suited to engaging the most problematic offenders. 52 During the one year follow-up period 47.9% of this new participant-matched control group reoffended compared with 50.4% of Diamond participants. There is therefore no indication that Diamond participation had any impact on reoffending during the year following prison release. Both proportions were again accurately predicted by OGRS3. Insights into promising aspects of the Diamond Initiative As a flexible and needs-based approach, Diamond provided a range of different types of assistance to a broad and diverse group of offenders. From the outset of the evaluation the intention was to examine impact in relation to this variation, by identifying any sub-groups of participants (defined either demographically or in terms of the types of ‘outcomes’ they achieved through Diamond) who exhibited particularly positive reoffending rates and might therefore be indicative of promising target groups or programme elements. A series of methods have been used to identify sub-groups that may have been positively impacted by Diamond participation, including:  Comparing actual against predicted reoffending rates for sub-groups of participants  Comparing the actual reoffending rate of sub-groups with the reoffending rate of their ‘most similar matched counterparts’ from the control group.  Comparing the average number of offences committed by sub-groups during the year following release with offending during the previous year. Each of these approaches has its drawbacks, and across them all, analysis has been hampered by small base sizes resulting from lower than expected total participant numbers, however, using a ‘triangulation approach’, one group of participants shows consistent indications of reduced reoffending. Key Learning: A second test for impact at the participant level supports the referral level findings, that the Diamond Initiative had no impact on reoffending. 53 42 of the 121 Diamond participants had recorded outcomes relating to Finance, Debt and Benefits, 17 of these (40.5%) reoffended during the follow-up year; this is 8.2% fewer than predicted by OGRS and 9.5% below the rate of their matched counterparts from the control group. As all participants had been convicted of offences leading to short prison sentences during the year before the follow-up period, it is unsurprising that the average number of recorded offences committed by all participants fell from an average of 3.8 during the year prior to release, to 2.1 during the follow-up year (see previous figure 9). Those with Finance, Debt and Benefits outcomes however had a larger reduction from 3.6 offences each during the previous years to 1.2 offences after release. While these findings are consistently positive, the scale of the reductions and the base sizes are relatively modest and the indication of a positive impact on reoffending remains tentative. It is however of note that recent studies have also indicated the importance of financial barriers to desistance (Bottoms and Shapland, 2011) and the potential impact of practical assistance in this area should not be overlooked29. Did better implementation lead to better reoffending outcomes? Owing to the constraints imposed by the timeframe for this research, impact analysis has (so far) been limited to a group of offenders referred to Diamond during the early months of the scheme, during which time, implementation issues (such as accommodation, IT and equipment) were being resolved and multi-agency working was still developing (as we saw in chapter 3). In order to investigate whether these improvements resulted in 29 The most frequent types of Finance, Debt and Benefits outcomes achieved by the cohort were Benefits Maximised (31), legally acceptable identification (obtained) (11) and Debt reduction scheme in place (8). Key Learning: Tentative but consistent indications of a positive impact of finance, benefits and debt assistance on reoffending were observed and are worthy of both further research and consideration in designing future provision. 54 more positive outcomes, a second cohort of referrals, with later prison release dates (during January, February and March 2010), has also been analysed. As later referrals, it has only been possible to calculate a six-month proven reoffending rate for this group30. This shows that, of the 137 Diamond referrals released during this later period, 27 (19.0%) had reoffended within six-months of release. The equivalent six- month rate for the main referral cohort (pre-October 2009 releases) was 27.5%31, with the original control group at 24.2%32. While it must be emphasised that there is no direct control for the later cohort, a comparable OGRS3 score of 0.46 suggests the groups compared here are broadly similar. Comparing between sub-groups of the cohorts indicates that although lower rates were observed across the later referral group (including those who did not receive Diamond support) those engaged by Diamond showed the greatest reduction in the later cohort compared with main group33. While this is broadly supportive of an impact of Diamond engagement on reoffending during the later implementation phase, the fact that those who did not receive any DI assistance also showed marginally improved reoffending rates suggests that factors other than Diamond (such as general crime trends) may be contributing to the observed differences and this finding should therefore be treated with caution. 30 Offence with 180 days of release, conviction within 213 days 31 The difference between the later cohort and the original evaluation cohort is statistically significant (One tail Z-test, Z=1.84) 32 These are the proportions of the full evaluation and control cohorts who committed an offence within 180 days of release which resulted in a conviction within 213 days. They differ from those reported in figure 8, which are based on the proportion of the cohort for which a full 18 months of data are available and count offences committed within 180 days which resulted in a conviction within 360 days. 33 Engaged portions of both groups had comparable OGRS3 scores - see appendix 12. Key Learning: There are tentative indications that those referred to Diamond later in the pilot, might have achieved better reoffending outcomes. Further analysis, over a longer follow-up period would be required to draw firm conclusions; however if this is sustained in the next analysis, the lesson that early implementation issues should be carefully programme managed, will be clear. 55 Economic Evaluation Assessing the case for reinvestment The Diamond Initiative set out to test the case for a shift in funding emphasis from ‘back- end’ criminal justice and custodial functions to community based intervention and prevention. If Diamond can be shown to have brought about a reduction in cost to the public purse (or to ‘society’), and if this reduction were to outweigh the costs of implementing the scheme, then a clear economic argument for continuation and expansion, funded (at least in part) from savings in other parts of the criminal justice system, could be made. This was one of the key arguments underpinning Diamond (i.e. Justice Reinvestment). As part of this evaluation, MVA Consultancy were contracted to produce an economic cost-benefit assessment of the Diamond Initiative, with the aim of assessing the extent to which money invested in Diamond provides an economic return in terms of the costs of crime averted. In line with the guidance set out in the H.M. Treasury Green Book (H.M Treasury, 2003), MVA quantified the benefit of the Diamond Initiative in monetary terms. Using a cost of crime model based on Home Office Research (Dubourg et al., 2005), the cost of the crime committed by the referral cohort during the year following prison release was compared to that generated by the control group, representing the counterfactual costs that would have occurred in the absence of Diamond. Based on an early cut of PNC data34, MVA estimate that the 281 cohort members available for evaluation, committed crime costed at a total of £1,109,280 during the year following release from prison, at an average cost of £3,950 per offender. In comparison The 234 available control cohort members committed crime costed at £913,140 during the year following release, at an average cost of £3,900 per offender. 34 PNC data run in September 2010, 281 evaluation and 234 control group members had completed 420 days since prison release 56 Given that no impact was found on the amount or seriousness of offending committed by the evaluation cohort compared with the control, it is unsurprising that no evidence has been found to indicate that the Diamond Initiative recouped any of the costs incurred in its work with the less-than-12-month sentence group, or that a redirection of funding from other Criminal Justice functions to a ‘Diamond style’ resettlement offer would be justified on financial grounds. Targeting resettlement resources Calculating the cost of crime for both the Evaluation Cohort and Control Group will always be difficult and a number of important caveats to these estimates are acknowledged by MVA:  The model includes estimates of costs incurred by individuals and households and costs to the Criminal Justice system, there may however be additional, less easily quantified costs resulting from crime. For instance, neither cost of the fear of crime, nor of the impact of crime on indirect victims, witnesses or the families of victims, are factored into these calculations.  The available research does not allow for each specific classification of crime committed by the cohorts, to be costed - to account for this, MVA have costed uncategorised crimes at 50% the cost of categorised crime, but acknowledge that this may under or over estimate the real value.  The model does not take account of the cost of any undetected crime committed by the group. While acknowledging the difficulties in accurately estimating the cost of crime, the implications of these estimates for any scheme which seeks to target resources at the less-than-12-month group must be considered. An average cost of crime in the region of £4,000 per less-than-12-month offender in the year following release, provides a challenging financial context for any intervention attempting to deliver an effective return on investment. For example, to recoup costs within one year, an investment of £1m would need to deliver a zero reoffending rate in a cohort of 250 less-than-12-month-sentence prison releases. 57 Diamond spent £11m over two years, (the majority of which is likely to have been directed toward the less-than-12-month sentence group) and in its first nine months was only able to work with 121 of this less-than-12-month group. Of course, it is not necessary for all benefits to be recouped within the first year for a scheme to be cost effective, however quite apart from questions of how to reduce reoffending, the (financial) scale of the problem (based on these estimates) and the likely costs of potential solutions appear to be difficult to reconcile. Furthermore, increasing the number of offenders that participated and limiting the time-intensive searching for offenders would also go some way to enable a more effective manner or working. It is necessary to consider therefore, whether blanket targeting of the less-than-12-month- sentence group with resettlement assistance remains a viable option. More than half of the less-than-12-month offenders scrutinised in this study (and around 40% of those across England and Wales) did not reoffend within the year – any (additional) spending to attempt to prevent reoffending amongst this group would therefore (necessarily) have yielded a zero return35. Given that good tools exist to risk assess the likelihood of reoffending (OGRS3) and of more ‘serious’ (and expensive) reoffending (OASys), a smarter direction of resources towards those offenders most likely to reoffend and incur costs to society (within the less-than-12-month sentence group or more broadly) would appear essential to delivering more cost effective resettlement. 35 In terms of crime prevented within the first year. Key Learning: Given that no impact on offending has been found, it is unsurprising that no economic benefit, in terms of cost of crimes averted, has been detected. Cost efficient offender management must be scaled to the size of the target population and target resources effectively, for instance by using risk assessment tools (such as OGRS) to achieve standardised selection. 58 Chapter 5: Discussion The current report describes the results from the SRAU evaluation of the Diamond Initiative - a two-year progressive offender management initiative that operated in six London boroughs. This section will present the key results of the evaluation, but also present the learning from the Diamond Initiative that can inform the development of any subsequent model of offender management in London (and beyond). Even though there was a compelling case for a Justice Reinvestment model of offender management in London and many of the ingredients of successful schemes were built into Diamond (e.g. multi-agency working, co-location): the headline finding of the report is that there was no evidence of reduced reoffending. Comparing the Diamond cohort to a statistically matched control group of similar people in similar areas, revealed no differences in offending behaviour (basic reoffending, survival, amount or seriousness of offending) in the 12 months subsequent to prison release. As a finding this holds important implications for the Diamond Initiative. It should be stated that these results do not demonstrate that Diamond has failed as an approach, rather that careful consideration as to why the results are not as expected is required. One possibility is that closer contact with the police, as a result of Diamond participation increased the likelihood of arrest and prosecution for offences that might otherwise have gone undetected and that this nullified any positive impact of the support received, within the reoffending data. While this type of effect has been suggested elsewhere (Hamilton, 2010) and cannot be ruled out, there is no clear evidence to support this and reflections on delivery issues are likely to generate more helpful insights. The methodology implemented in the evaluation allowed a tracking of staff and offender views over the course of the initiative, and it is through this work that we can generate insights into the reasons Diamond may not have achieved better results. These can be summed up as implementation issues, meeting offender need at a local level and cultural tensions and will be addressed in turn. 59 Implementation issues: One of the key findings from chapter 3 was that the set-up and implementation of the scheme had to overcome a number of obstacles (e.g. IT, accommodation, training, staffing) and whilst these, in the main, improved with time, which was a real positive, such innovation requires time. The lack of transition time is likely to have had an influence on how effective staff were able to be and may have impacted on programme integrity. It is known, through previous research, that delivering without programme integrity can negatively influence any reported outcomes. Meeting offender need at a local level: Whilst staff were generally positive about aligning resources to offender need and this was perceived to have improved over the course of the scheme, when offender needs (as measured by OASys) were examined next to local service involvement, one could argue that gaps were evident and more could have been done to effectively meet the needs of offenders. Many of the key agencies required and working within DI (such as housing, education, mental health, drugs treatment) were only involved when required as opposed to being consistently involved. It is possible that this may have influenced the ability of staff to be able to meet need effectively. This emphasises the need for strong coordination across multi-agency working as we know that omissions in this area will negatively influence the likely impact on re-offending. Multi-agency working: Another key learning point form the scheme and one relevant for all Integrated Offender Management initiatives, was the cultural tension identified between police and non-police staff working on the DI. Although evident in year one, this became more pronounced within year two and was apparent within a number of areas such as understanding the aims of DI, training perceptions, views on the use of police uniform and organisational differences in working. A key question here is the role of the police in the management of offenders and whether the police are the most skilled or appropriate people to lead such initiatives. Non-police staff clearly felt police officers were not suitable to lead such an approach, yet there is a strong (52% of the public according to a recent tracker survey) public perception that police are expected to manage offenders on release (IPSOS MORI, 2011). The role of the different agencies within offender management is something that requires further consideration, ensuring that staff 60 skills and motivation are appropriate, along with work ensuring that organisational cultures do not clash. None-the-less, as previously discussed there are signs that the approach did mature over the two year timeframe and there are tentative results indicating this may have led to better outcomes for year two clients. This should be the subject to further longitudinal analysis to test this finding. If these results are confirmed over a longer period, then this would echo with the research that shows strong implementation and programme integrity are essential in order to achieve desired objectives (Barnoski, 2002; Coulter, 2010). Lessons from the Diamond Initiative Outside of the above issues, there is a range of key learning identified through the Diamond evaluation that can inform discussions as to the next steps of offender management in London. These are: 1) There is a clear need for a robust referral, audit and performance tool in any offender management initiative. The LISARRT system encountered implementation difficulties and though it improved over time (to the benefit of DI staff and the SRAU researchers) difficulties were still encountered with the system towards the end of the DI. A system is required that systematically captures the clients, their needs, the ongoing activities and outcomes on a quick-time basis. This information should be used to benefit performance monitoring, decisions about offender support and any future evaluation. The PPO scheme operates the J-Track performance tool which seeks to combine crime data, demographics and needs records. This system might provide a starting point to stimulate discussions in the management of data. 2) Future offender management initiatives should use offending risk assessment measures (such as OGRS) to ensure that resource allocation is matched to offender need. It is to be noted that within DI, more than half of the less-than-12-month cohort did not (and were predicted not to) re-offend within the year. Given good tools do exist, a smarter direction of resources towards those offenders more likely to re-offend, would 61 seem a wise move - both in terms of reducing reoffending and in delivering cost- effectiveness. On a purely technical level, an initiative that used a risk assessment tool (i.e. OGRS) as the only selection criteria would also open up advanced methodological evaluation techniques such as the Regression Discontinuity Design36 (Hahn, Todd & Van der Klaauw, 2001). 3) Those who reoffend, both in the Diamond and control group, tended to do so quickly after prison release; this suggests a real ‘window of opportunity’ for supportive engagement immediately following (and even prior to) prison release. However, being able to effectively address this would require good multi-agency working and data sharing in particular with the prison service and smooth referral systems to ensure a joined up approach between custody and community. A recent report has identified improved information management and effective needs assessment as crucial to improving offender management in prisons (Criminal Justice Joint Inspection, 2011); the benefits of such improvements are likely to extend beyond release from custody. 4) One valuable finding resulting form the research is the high acceptance of the Diamond Offer by offenders. To illustrate, three in five offenders accepted the DI offer but these offenders were at a greater risk of re-offending and presented more needs associated with criminal behaviour than those who refused. There are parallels to be drawn here to the Sheffield Desistance Study (Bottoms & Shapland, 2011) where even a highly criminal sample reported the sincere desire to desist from crime. The demand for resettlement support therefore appears to be need-driven and the voluntary model appears well suited to engaging the most problematic offenders. However, such offender management needs to be tempered by the desistance research that demonstrates that even with this motivation, criminal behaviours are deeply entrenched in lifestyle and cannot just be turned off. As such it is essential to be able to influence positively an offender’s motivation before and after entry into voluntary schemes. There needs to be a structured approach 36 The Regression Discontinuity Design is a before and after, two group design that evaluates causal effects of an intervention. For the RDD to be a feasible approach, persons are required to be assigned to groups solely based on a cut-off score on a pre-programme method. The difference between groups is then measured as the discontinuity between treatment and control regression lines. 62 by teams, supported by evidence and training, to manage offender’s motivation throughout their contact. 5) The finding that both the London cohorts were re-offending below the expected national less-than-12-month sentence cohort is of interest here. Current thinking indicates this can be understood in terms of differences in the make-up of the London groups compared with the national sample. The large proportion of non-UK born offenders within London (who exhibit different and less prolific offending profiles) is one potential explanatory factor. These findings suggest the need for regional / local benchmarking in order to provide a framework for assessing the effectiveness of local offender management approaches. This is a timely finding given the current discussions on a new re-offending measure that captures local issues (Ministry of Justice, 2011). 6) Diamond attempted to map ‘million dollar blocks’ thinking on to existing (Safer Neighbourhoods) ward level geographies. While it remains true that the effects of crime and deprivation are unevenly spread across London, London’s wards are themselves often internally diverse with both challenging and more affluent areas. In practice the original geographic parameters were found to be restrictive; teams failed to see the rationale for rejecting needy and motivated potential clients from other parts of the borough while case loads were relatively light and over the course of the pilot geographic criteria were relaxed. While geographic clustering of offender populations has been demonstrated in the UK context (Allen et al., 2007), the experience of Diamond suggests that these may be insufficiently distinct to warrant an offender management approach that is restricted to a small geographic area. Managing persistent offenders may require a different geography, in contrast to the work of Weisburd et al. (2009) which suggests a geographical approach to managing crime. There is no clear ‘one size fits all’ model available; it is (or needs to be) an approach where support is individually client based, and which offers challenge to longer term problems (fuelled by disadvantage, substance abuse, violence in the home and so forth). However, there are also insights into who may be most suitable for a Diamond approach to offender management. There are tentative indications that offenders with outcomes 63 linked to finance and debt may be linked to better than expected reoffending. Such a finding is consistent with previous research into offender need (Bottoms & Shapland, 2011). The future of offender management There is substantial momentum behind offender management both within the MPS and Government at present. One recommendation from the research would be to align any subsequent evolving approach to offender management in London to the current Government’s model of Integrated Offending Management (see appendix 13 for a conceptualisation of DI and an IOM approach to offender management). The outcomes focus of IOM sits alongside the language of the Coalition Government in terms of crime prevention and demand reduction, but would also assist progress on the previously reported issues such as agency involvement and multi-agency working identified within Diamond. This evolution should build on a multi-sector, multi-agency collaborative approach and seek to bring together many diverse policing and non-policing processes and activities (e.g. DI, PPO, MAPPA, DIP, MARAC processes; Safer Neighbourhoods and more) in order to align and identify the service gaps for managing offenders with a real potential to yield benefits - both economically and for reducing offending. Internally for the MPS, such an approach would be compatible with the MPS/London Anti-Violence Strategy and an intelligence led analysis of crime (a Victim Offender Location Time (VOLT) approach). There is also the growing emphasis on ‘payment by results’ or a financial incentive model of demand reduction in criminal justice. As outlined in the ‘Breaking the cycle’ Green Paper it is likely that a number of London Boroughs (along with Greater Manchester) will act as a pilot to such a way of working. The objective of the pilot is to pay local statutory partners a proportion of the savings if they are successful in reducing demand for:  Court cases and the associated legal aid costs for all sentences except those that result in a long prison sentence; 64  Community orders and Suspended Sentence Orders; and  Custodial sentences for less than 12 months. The savings for the local statutory partners would be available to reinvest in further crime prevention activity at the local level by agreement of the relevant partners. Such a model, whereby finances are provided either for a reduction in re-offending or wider demand on the CJS, poses challenges that would need to be addressed. For example, the need for a greater degree of coordination across agencies, ensuring a joint approach and a shared understanding of the relevant outcomes of such methods of working (Bowles, 2010), have generally resulted in staff attempting to ‘cherry pick’ lower risk offenders (House of Commons Work and Pensions Committee, 2010). Such a manner of working will also likely require a better approach to prevention work (e.g. to reduce demand). In addition, one of the ramifications for criminal justice in any financial incentive model would be the need for robust research when ultimately deciding upon payment awards. Evaluation is a complex and highly technical area. The Diamond evaluation could be seen as a touch-stone for any other local initiative seeking to examine impact upon crime, and to some degree demonstrates that local initiatives can be subject to a rigorous impact evaluation. It could be argued that the methodology adopted in the DI evaluation was both innovative and original in respect to previous offender management evaluations (Culshaw, 2005; Dawson & Cuppleditch, 2007; Dawson, 2007; Skodbo et al., 2007). Not only has the current work been able to go beyond basic levels of reoffending into issues such as severity, survival, frequency and capture changing perceptions over the course of the DI - it has also been possible to conduct sub-group analysis of impact split by levels of engagement. These analyses clearly have to be caveated by the relatively small sample sizes, none-the-less, the aspiration was to go beyond what had been done before and provide a level of analytical depth not previously seen in offender management evaluations. To illustrate, the recent 2009 PPO refresh again cites strong evidence of 65 success (a 39% decrease in offending) but is not able to examine levels of engagement demonstrating the gaps in the analysis of offender management at a national level (Home Office, 2010). Final thoughts on the evaluation The short term custodial sentence offender group are churning through the criminal justice system at a high cost to society. The Diamond Initiative was a two-year progressive model of working with this group of offenders. This is the second published report based upon the SRAU evaluation of the Diamond Initiative. Whilst it is likely that there will be future reports following the Diamond cohort to understand the long term influence of the scheme - the current report should be seen as a major milestone in the evaluation of offender management programmes generating learning for policy makers, practitioners and researchers. Whilst the headline finding of the report is that there was no evidence of reduced reoffending as a result of the Diamond Initiative within the original cohort - this does not demonstrate that Diamond has failed as an approach: such a conclusion would be erroneous. The report presents a complicated set of findings; the main ones are below:  Qualitative findings in both years show that DI was highly regarded by offenders.  Qualitative findings in both years strongly suggest a positive influence on offenders.  The implementation had to overcome a number of obstacles (equipment, staffing, buildings, IT systems) which could be anticipated in the set-up of future offender management initiatives.  One could argue that more could have been done to effectively meet the needs of offenders through consistently involved support agencies.  Teams generally worked well across disciplinary boundaries, although tensions between police and other agencies were apparent and never fully reconciled.  Year 1 findings suggested that those participating in DI had much lower 6-month reconviction rates than a (roughly matched) comparison group. 66  Year 2 experimental and control groups were very well-matched, according to OGRS scores and demographic criteria.  Only one in six of the experimental group (those originally thought to be eligible for DI) had sustained engagement with the programme.  Overall the experimental and control groups had remarkably similar 1-year reconviction rates.  Both experimental and control groups had much lower reconviction rates than a national cohort of prisoners serving under 12 months, perhaps reflecting the high proportion of people in each group who were born abroad. In addition, there are insights that tackling an offenders’ finance, debt and benefit needs may have been effective and that better implementation over the two years may have led to lower reoffending: such results will be the focus on continued longitudinal research to further explore the impact of the Diamond Initiative. This emphasises the key learning of meeting offender need and programme maturation. From an evaluation point of view, the concept of maturation also raises the vital consideration of what can be called the ‘point-zero’ of an initiative. This refers to the point by which an initiative was fully operational, as opposed to still being developed or merely announced. Evaluations have not taken this into account and the question for evaluation research is at which point to commence analysis so to avoid many of an initiative’s teething problems so that we are only assessing the true effect of the programme. From a practical perspective this raises the interesting question as to when clients should be ‘taken on’ to an initiative: if this is done too early within implementation there is the danger that clients will not receive an appropriate service. Furthermore, there are real strategic and practical implications in the design and implementation of offender management programmes that have been generated from the research. It is important that the next phase of offender management within the Metropolitan Police Service does not start from a blank slate but rather builds upon the lessons and implications so to ensure an evidence-based, effective and targeted offender management model. These main such implications are: 67 1) The clear need for a robust referral, audit and performance tool for offender management initiatives; 2) The use of formal risk assessment measures (such as OGRS) to ensure that resource allocation is matched to offender need; 3) The identified ‘window of opportunity’ that demonstrated that offenders offender quickly after prison release highlighting the need for timely engagement; 4) The relatively high acceptance of the Diamond offer by offenders demonstrates the initial willingness to attempt behavioural change. 5) The original geographic parameters of Diamond were found to be restrictive and the wider remit may be required. From a research perspective the aspiration from the outset was to conduct an original and innovative evaluation that would stand up against the firmest of critiques. This has been achieved. The range of methods used, the capture of change over time and the remarkable strength of the counterfactual - all demonstrate the quality of the evaluation. Indeed, the evaluation should be viewed as a touch-stone for other local initiatives seeking to conduct evaluation and further demonstrates that local initiatives can be subject to rigorous evaluation. This latter point is crucial considering the move into payment by results, decentralisation and localism as initiated by the coalition Government. Perhaps the final thought of the evaluation should lay with the offenders and the recognition that evaluating ‘people change’ programmes is really very hard. Offender’s present complex and chaotic lives entrenched in social needs with long-term criminal routines and habits that are not easily changed. This needs to be acknowledged by both policy and research in the guise of realistic expectations when attempting to change criminal behaviour. 68 References Allen, R., Jalleb, K., & Snaith, E. (2007). Justice Reinvestment in Gateshead - the story so far. In R. Allen and V. Stern (Eds.), Justice Reinvestment - A New Approach to Crime and Justice. London: International Centre for Prison Studies. Allen, R., & Stern, V. (Eds.). (2007). Justice Reinvestment - A New Approach to Crime and Justice. London: International Centre for Prison Studies. Aos, S., Miller, M., & Drake, E. (2006). Evidence-based adult correction programmes: What works and what does not. Olympia, WA: Washington State Institute for Public Policy. Baker, K., Jones, S., Roberts, C., & Merrington, S. (2002). Validity and Reliability of ASSET: Findings from the first two years of the use of ASSET. Oxford: Probation Studies Unit, Centre for Criminological Research, University of Oxford. Barnoski, R. (2004). Outcome Evaluation of Washington State’s Research-Based Programme for Juvenile Offenders. Retrieved March 23, 2011, from http://www.wsipp.wa.gov/rptfiles/04-01-1201.pdf Bonta, J., Law, M., & Hanson, K. (1998). The Prediction of Violent and Criminal Recidivism among Mentally Disordered Offenders: A meta-analysis. Psychological Bulletin, 123, 123-142. Bottoms, A., Shapland, J., Costello, A., Holmes, D., & Muir, G. (2004). Towards Desistance: Theoretical Underpinnings for an Empirical Study. The Howard Journal, 43(4), 368-389. Bottoms, A., & Shapland, J. (2011). Steps Towards Desistance Among Male Young Adult Recidivists. In S. Farrall, R. Sparks, M. Hough and S. Maruna (Eds.), Escape Routes: Contemporary Perspectives on Life after Punishment. London: Routledge Bowles, R. (2010). Diamond Initiative: Economic Evaluation. Unpublished manuscript. Cabinet Office. (2010). The Coalition: Our Programme for Government. Retrieved March 23, 2011, from http://www.cabinetoffice.gov.uk/sites/default/files/resources/coalition_programme_for_go vernment.pdf Colledge, M., Collier, P., & Brand, S. (1999). Crime reduction programme and constructive regimes in prison. Research, Development and Statistics Directorate. London: Home Office. Copas, J. B., Ditchfield, J., & Marshall, P. (1994). Development of a New Risk Predictor Score. Research Bulletin 36. London: Home Office. 69 Coulter, S. (2010). Systemic Family Therapy for Families who have Experienced Trauma: A Randomised Controlled Trial. British Journal of Social Work. Doi: 10.1093/bjsw/bcq132. Criminal Justice Joint Inspection. (2011) A Joined Up Sentence? Offender Management in Prisons 2009/2010. HM Inspectorate of Prisons and HM Inspectorate of Probation Culshaw, A. (2005). Prolific and other Priority Offender (PPO) Scheme - Salford case study. Internal report produced for GONW. Dawson, P. (2007). The National PPO Evaluation - Research to Inform and Guide Practice. Home Office Report 09/07. London: Home Office. http://www.justice.gov.uk/inspectorates/hmi-probation/docs/A_Joined_Up_Sentence- rps.pdf Dawson, P., & Cuppleditch, L. (2007). Programmes for Offenders: Guidance for Evaluators. Research, Development and Statistics Directorate. London: Home Office. Dawson, P., & Williams, E. (2009). Reflections from a Police Research Unit - An Inside Job. Policing: A Journal of Policy and Practice, 3(4), 373-380. Dawson, P., & Stanko, B. (2010). An Evaluation of the Diamond Initiative: Year One Findings. London: Metropolitan Police Service; London Criminal Justice Partnership. Day, A., & Howells, K. (2002). Psychological Treatments for Rehabilitating Offenders: Evidence-Based Practice Come of Age. Australian Psychologist, 37, 39-47. Dubourg, R., Hamed, J., & Thorns, J. (2005). The Economic and Social Costs of Crime against Individuals and Households 2003/04. London: Home Office. Farrall, S., Sparks, R., Hough, M., & Maruna, S. (Eds.). (2011). Escape Routes: Contemporary Perspectives on Life after Punishment. London: Routledge. Farrington, D. P. (1990). Implications of Criminal Career Research for the Prevention of Offending. Journal of Adolescence, 13, 93-113. Farrington, D. P. (1997). Human Development and Criminal Careers. In M. Maguire, R. Morgan and R. Reiner (Eds.), The Oxford Handbook of Criminology (2nd ed.). Oxford: Clarendon Press. Farrington, D. P. (2002). Developmental Criminology and Risk-Focussed Prevention. In M. Maguire, R. Morgan and R. Reiner (Eds.), The Oxford Handbook of Criminology (3rd ed.). Oxford: Clarendon Press. Gendreau, P., Little, T., & Goggin, C. (1996). A meta-analysis of the Predictors of Adult Offender Recidivism: What works! Criminology, 34, 575-607. 70 Hahn, J., Todd, P., & Van der Klaauw, W. (2001). Identification and Estimation of Treatment Effects with a Regression-Discontinuity Design. Econometrica, 69(1), 201- 209. Hamilton, Z. (2010). Do Re-entry courts Reduce Reoffending? Results from the Harlem Parole Reoffending Courts. New York: Centre for Court Innovation Harper, G., & Chitty, C. (Eds.). (2005). The Impact of Corrections on Reoffending: A Review of ‘What Works’. (2nd ed.). Home Office Research Study 291. London: Home Office. Herbst, J. R., & Gunn, J. (1991). The Mentally Disordered Offender. London: Butterworth-Heinemann/Mental Health Foundation. HM Prison Service (2011) Population Bulletin - Weekly 18 March 2011 http://www.hmprisonservice.gov.uk/resourcecentre/publicationsdocuments/index.asp?ca t=85 HM Treasury. (2003). The Green Book: Appraisal and Evaluation in Central Government. Retrieved March 23, 2011, from http://www.hm- treasury.gov.uk/d/green_book_complete.pdf Hollin, C. R., & Palmer, E. J. (2006). The Level of Service Inventory - Revised Profile of English Prisoners: Risk and Reconviction Analysis. Criminal Justice and Behaviour, 33, 347-366. Hollin, C. (2008). Evaluating Offending Behaviour Programmes: Does Only Randomization Glister? Criminology and Criminal Justice, 8(1), 89-106. Hollin, C., Maguire, J., Hounsome, J., Hatcher, R., Bilby, C., & Palmer, E. (2008). Cognitive Skills Behaviour Programs for Offenders in the Community: A Reconviction Analysis. Criminal Justice and Behaviour, 35, 269-284 Home Office. (2010). Reducing Demand, Restricting Supply, Building Recovery: Supporting People to Lead a Drug-Free Life. Retrieved March 23, 2011, from http://www.homeoffice.gov.uk/publications/drugs/drug-strategy/drug-strategy- 2010?view=Binary House of Commons Work and Pensions Committee. (2010). Management and Administration of Contract Employment Programmes. London: The Stationary Office. Howard, P., Francis, B., Soothill, K., & Humphreys, L. (2009). OGRS 3: The Revised Offender Group Reconviction Scale. Research Summary 7/09. London: Ministry of Justice. Ipsos Mori. (2011). Police Federation of England and Wales: Police Priorities. Retrieved March 22, 2011, from http://www.ipsos- 71 mori.com/researchpublications/publications/1408/Police-Federation-of-England-and- Wales-Policing-Priorities.aspx King, R. D., Massoglia, M., & MacMillan, R. (2007). The Context of Marriage and Crime: Gender, the Propensity to Marry and Offending in Early Adulthood. Criminology, 45, 33- 65. Kurlycheck, M. C., Brame, R., & Bushway, S. D. (2006). Scarlet Letters and Recidivism: Does an Old Criminal Record Predict Future Offending? Criminology and Public Policy, 5(3), 483-504. Laub, J.H., & Sampson, R. J. (2003). Shared Beginnings, Divergent Lives: Delinquent boys to age 70. Cambridge, MA: Harvard University Press. Lipsey, M. (1995). What do we learn from 400 Research Studies on the Effectiveness of Treatment with Juvenile Delinquents? In J. McGuire (Ed.), What works: reducing reoffending. Chichester: John Wiley & Sons. Maguire, M., & Raynor, P. (2006). How the Resettlement of Prisoners Promotes Desistance from Crime: Or Does It? Criminology and Criminal Justice, 6(1), 17-36. Mason, T., de Silva, N., Sharma, N., Brown, D., & Harper, G. (2007). Local Variation in Sentencing in England and Wales. London: Ministry of Justice. McGuire, J. (2000). What Works in Reducing Criminality. Retrieved March 23, 2011, from http://www.aic.gov.au/events/aic%20upcoming%20events/2000/~/media/conferences/cri minality/mcguire.ashx McGuire, J. (Ed.). (2002). Offender rehabilitation and treatment: Effective Programs and Policies to Reduce Reoffending. Chichester: John Wiley & Sons. McMurran, M. (Ed.). (2002). Motivating Offenders to Change: A Guide to Enhancing Engagement in Therapy. Chichester: John Wiley & Sons. Mednick, S.A., Moffit, T. E., & Stack, S. A. (1987). The Causes of Crime: New Biological Approaches. New York: Cambridge University Press. Ministry of Justice. (2009). Story of the Prison Population 1995-2009, England and Wales. London: Ministry of Justice. Ministry of Justice. (2010a). Breaking the Cycle: Effective Punishment, Rehabilitation and Sentencing of Offenders. London: The Stationary Office. Ministry of Justice. (2010b). Compendium of Reoffending Statistics and Analysis. London: Ministry of Justice 72 Ministry of Justice. (2011). Improvements to MoJ Statistics Consultation Questionnaire. Retrieved March 23, 2011, from http://www.justice.gov.uk/consultations/docs/stats- consultation-questionnaire.doc Morton, S. (2009). Findings from OASys Completion Rates. Research Summary 4/09. London: Ministry of Justice. National Audit Office. (2010). Managing Offenders on Short Custodial Sentences. London: The Stationary Office. National Institute of Drug Abuse. (2003). Crossing the Bridge: An Evaluation of the Drug Treatment Alternative-to-Prison (DTAP) Programme. Retrieved March 23, 2011, from http://www.casacolumbia.org/articlefiles/379- Crossing%20the%20Bridge%20DTAP%20Evaluation.pdf Nilsson, A. (2003). Living Conditions, Social Exclusion and Recidivism Among Prison Inmates. Journal of Scandinavian Studies in Criminology and Crime Prevention, 4, 57- 83. Niven, S., & Stewart, D. (2005). Resettlement Outcomes on Release from Prison in 2003. Home Office Findings 248. London: Home Office. Oldfield, M. (1996). The Kent Reconviction Survey: A 5 year Survey of Reconvictions amongst Offenders made subject to Probation Orders in 1991. Maidstone: Kent Probation Service. Palmer, E. J., & Hollin, C. R. (2004). The use of the Psychological Inventory of Criminal Thinking Styles with English Young Offenders. Legal and Criminological Psychology, 9(2), 253-263. Prison Reform Trust. (2010). Bromley Briefings Prison Factfile. Retrieved March 23, 2011, from http://www.prisonreformtrust.org.uk/Portals/0/Documents/FactfileDec10small.pdf Raynor, P., Kynch, J., Roberts, C., & Merrington, S. (2000). Risk and Need Assessment in Probation Services: An Evaluation. Home Office Research Study 211. London: Home Office. Rutter, M., Giller, H., & Hagell, A. (1998). Antisocial Behaviour by Young People. Cambridge: Cambridge University Press. Sampson, R.J., & Laub, J. H. (1993). Crime in the Making: Pathways and Turning Points through Life. Cambridge, MA: Harvard University Press. Sherman, L., Gottfredson, D., MacKenzie, D., Eck, J., Reuter, P., & Bushway, S. (1997). Preventing Crime: What Works, What Doesn’t, What's Promising. A Report to the United States Congress Department of Criminology and Criminal Justice University of Maryland. 73 Skodbo, S., Brown, G., Deacon, S., Cooper, A., Hall, A., Millar, T., … Whitham, K. (2007). The Drug Interventions Programme (DIP): Addressing Drug Use and Offending through ‘Tough Choice’. London: Home Office. Social Exclusion Unit. (2002). Reducing Re-offending by Ex-Prisoners: Summary of the Social Exclusion Unit Report. London: Social Exclusion Unit. Spivak, A., & Damphousse, K. R. (2006). Who Returns to Prison? A Survival Analysis of Recidivism Among Adult Offenders Released in Oklahoma 1985-2004. Justice Research and Policy, 8(2), 57-88. Spohn, C., & Holleran, D. (2002). The Effect of Imprisonment on Recidivism Rates of Felony Offenders: A Focus on Drug Offenders. Criminology, 40, 329-357. Stewart, D. (2008). The Problems and Needs of Newly Sentenced Prisoners: Results from a National Survey. Ministry of Justice Research Series 16/08. London: Ministry of Justice Swyer, B., & Lart, R. (1996). Prisoners' Mental Health Problems: Screening Needs and Assessing Services. Probation Journal, 43, 205-210. Tucker, S. B., & Cadora, E. (2003). Justice Reinvestment - To Invest in Public Safety by Reallocating Justice Dollars to Refinance Education, Housing, Healthcare and Jobs. Ideas for an Open Society, 3(1), 95-133. Van Mastrigt, S. B., & Farrington, D. P. (2009). Co-offending, Age, Gender and Crime Type: Implications for Criminal Justice Policy. British Journal of Criminology, 49, 552- 573. Weaver, B., & McNeill, F. (2007). Giving up Crime: Directions for Policy. Edinburgh: Scottish Consortium on Crime and Criminal Justice. Weisburd, D., Morris, N. A., & Groff, E. R. (2009). Hot Spots in Juvenile Crime: A Longitudinal Study of Arrest Incidents at Street Segments in Seattle, Washington. Journal of Quantitative Criminology, 24, 443-467. 74 Appendix 1: Forming the control group - borough and ward level matching In forming the main control group a two-stage matching process was used to identify the boroughs and wards within London which were best matched, in theoretically relevant ways, to each of the Diamond boroughs and wards. In stage one, each of the Diamond boroughs was compared against all other (non-Diamond) London boroughs in terms of:  Population size  BME population  Drug crime level  Acquisitive crime level  Violent crime level  Indices of deprivation  Age of population (20-29 and 30-44)  Size of population without education  Number of prison releases For each of the above variables, where the comparison borough was found to be comparable to the Diamond borough to within 1%, a score of 3 was applied; where the difference was between 1% and 3%, a score of 2 was given and where the difference was between 3% and 5%, a score of 1 was allocated; any greater difference received no score. Summing across the criteria resulted in an overall score for each non-Diamond borough, which represented the strength of its match to the Diamond borough in question. The scored comparison boroughs were then ranked and the best (ten) matches for each Diamond borough, taken forward to the second stage of the process. In the second stage, each Diamond ward was matched, using the same process, against every ward located within the best-matched boroughs for that Diamond ward’s ‘parent’ 75 borough. In this way, it was possible to rank the best-matched London wards for each Diamond ward, taking into account both ward level and borough level characteristics (the results of the matching process are shown in appendices 2 and 3). Once a set of best matched wards had been established, the control group was formed by identifying the equivalent group of offenders to the evaluation cohort (less-than-12- month sentence offenders, with similar prison release dates) who had release addresses within the matched wards. In order to identify a sufficient number of offenders, it was necessary to extend the number of best-matched wards for each Diamond ward, to six. In several cases a ward was found to fall within the best six matches for two (or more) Diamond wards, in these instances each offender resident in that ward was counted two (or more) times within the control group37. This resulted in the identification of 232 unique offenders which, when multi-counting rules are applied, formed a control cohort comprising 327 offender records. 37 Multi-counting rules were agreed by the evaluation’s academic reference group and allow for the context of each Diamond ward to be given equal weighted in forming the control. 76 Appendix 2: Top 10 best matched boroughs, for each Diamond borough Croydon Score (max 30) Southwark Score (max 30) Ealing 23 Lambeth 25 Hillingdon 22 Hackney 23 Hounslow 20 Haringey 22 Bromley 18 Lewisham 21 Harrow 18 Brent 21 Redbridge 18 Hounslow 20 Brent 17 Ealing 18 Haringey 17 Islington 18 Merton 17 Tower Hamlets 18 Waltham Forest 17 Greenwich 17 Waltham Forest 17 Newham 15 Croydon 14 Merton 14 Hillingdon 13 Lambeth Score (max 30) Lewisham Score (max 30) Southwark 25 Southwark 21 Haringey 19 Haringey 20 Hackney 18 Hounslow 20 Ealing 18 Brent 19 Brent 17 Ealing 19 Camden 17 Waltham Forest 19 Islington 17 Redbridge 17 Tower Hamlets 17 Lambeth 16 Waltham Forest 17 Barking and Dagenham 16 Lewisham 16 Greenwich 15 Newham 16 Islington 15 Greenwich 16 Hackney 14 Hammersmith and Fulham 16 Merton 14 Newham Score (max 30) Hackney Score (max 30) Hackney 19 Southwark 23 Greenwich 19 Greenwich 20 TowerHamlets 19 Newham 19 WalthamForest 19 Waltham Forest 19 Haringey 18 Lambeth 18 Hounslow 16 Haringey 17 Lambeth 15 Ealing 16 Southwark 15 Islington 16 Brent 15 Tower Hamlets 16 Croydon 14 Bexley 15 Bexley 14 Brent 15 Bromley 14 Lewisham 14 Enfield 14 Croydon 13 Harrow 14 Camden 13 Harrow 13 77 Appendix 3: Top 6 best matched wards, for each original Diamond ward Diamond Borough Diamond Ward Matched Ward Croydon Bensham Manor Heston Central (Hounslow) Preston (Brent) Longthornton (Merton) Heston East (Hounslow) Roxbourne (Harrow) Chapel End (WalthamForest) Selhurst Kensal Green (Brent) Marlborough (Harrow) Roxbourne (Harrow) Cann Hall (WalthamForest) Grove Green (WalthamForest) Lady Margaret (Ealing) Thornton Heath Roxbourne (Harrow) Marlborough (Harrow) Roxeth (Harrow) Preston (Brent) Tottenham Hale (Haringey) Northumberland Park (Haringey) West Thornton Seven Kings (Redbridge) Goodmayes (Redbridge) Heston West (Hounslow) Greenford Green (Ealing) Northolt Mandeville (Ealing) Harefield (Hillingdon) Lewisham Coldharbour Camden Town with Primrose Hill (Camden) Thamesmead Moorings (Greenwich) Bruce Grove (Haringey) William Morris (WalthamForest) Hounslow West (Hounslow) Hornsey (Haringey) Newham Canning Town N Woolwich Riverside (Greenwich) Roxeth (Harrow) Kenton East (Harrow) Bethnal Green North (TowerHamlets) East India and Lansbury (TowerHamlets) White Hart Lane (Haringey) Canning Town S Upper Edmonton (Enfield) Noel Park (Haringey) White Hart Lane (Haringey) Turkey Street (Enfield) Leytonstone (WalthamForest) Valley (WalthamForest) Custom House Upper Edmonton (Enfield) Kidbrooke with Hornfair (Greenwich) Ponders End (Enfield) Cray Valley East (Bromley) White Hart Lane (Haringey) Valley (WalthamForest) 78 Diamond Borough Diamond Ward Matched Ward Southwark Farraday Harlesden (Brent) Feltham West (Hounslow) East India and Lansbury (TowerHamlets) Plumstead (Greenwich) Woolwich Common (Greenwich) Glyndon (Greenwich) Lewisham Evelyn Woolwich Common (Greenwich) Bruce Grove (Haringey) Northumberland Park (Haringey) Norwood Green (Ealing) Cathall (WalthamForest) Abbey (BarkingandDagenham) New Cross Cathall (WalthamForest) Tollington (Islington) Lavender Fields (Merton) Bruce Grove (Haringey) Kilburn (Brent) Woolwich Common (Greenwich) Hackney Leabridge Cann Hall (WalthamForest) Tottenham Green (Haringey) Bruce Grove (Haringey) Kensal Green (Brent) Welsh Harp (Brent) Cathall (WalthamForest) Springfield Greenford Broadway (Ealing) Cranford (Hounslow) Higham Hill (WalthamForest) St Pancras and Somers Town (Camden) Wealdstone (Harrow) Glyndon (Greenwich) 79 Appendix 4: Evaluation cohort and control group comparison statistics Evaluation Control Evaluation Control Number 368 327 233 191 Gender % Female 13.9 15.9 13.7 17.8 % Male 86.1 84.1 86.3 82.2 Ethnicity % BME 50.0 47.1 50.2 48.2 % White 49.2 52.6 48.9 51.8 % Not known 0.8 0.3 0.9 0.0 % White 49.2 52.6 48.9 51.8 % Black 37.0 29.7 38.2 28.3 % Asian 4.6 10.7 4.3 14.1 % Other 4.6 3.7 4.3 3.1 % Mixed 3.8 3.1 3.4 2.6 % Not known 0.8 0.3 0.9 0.0 Age (at prison release) Mean 32.3 32.4 31.5 33.1 % 19 and below 4.1 3.4 4.7 1.6 % 20 to 24 20.7 16.5 21.9 14.7 % 25 to 29 21.2 23.5 21.5 30.4 % 30 to 34 14.9 24.5 16.3 19.9 % 35 to 39 16.6 11.3 15.9 9.4 % 40 to 44 12.0 9.5 11.6 11.5 % 45 to 49 5.2 6.1 4.7 6.3 % 50 and over 5.4 5.2 3.4 6.3 Place of Birth % UK born 55.2 57.8 49.8 57.1 % Non-UK born 43.8 41.9 48.5 42.4 % Not known 1.1 0.3 1.7 0.5 Length of Criminal Career Mean (years) - All 9.4 9.9 9.2 10.3 Mean (years) - UK born 13.4 14.0 13.9 14.0 Mean (years) - Non UK Born 4.5 4.2 4.7 5.4 Age at First Conviction Mean age (years old) - All 22.4 22.0 21.9 22.3 Mean age (years old) - UK born 18.5 18.6 17.5 19.1 Mean age (years old) - Non UK Born 27.3 26.7 26.1 26.6 420 Days of data available 540 Days of data available 80 Evaluation Control Evaluation Control Predicted one year reoffending Mean OGRS3 score 0.44 0.43 0.44 0.43 % High risk 12.5 8.6 13.7 7.9 % Medium risk 51.1 55.4 49.8 58.6 % Low risk 33.2 30.3 33.0 29.3 Not available 3.3 5.8 3.4 4.2 Identified Needs (OASys) % OASys Available 59.2 48.3 56.7 46.6 Of which Mean number of identified needs 4.8 4.6 4.7 4.8 % Accomodation need 54.1 49.4 53.8 53.9 % ETE need 72.5 79.7 72.0 79.8 % Relationships need 63.3 52.5 64.4 58.4 % Lifestyle & Associations need 73.4 69.0 73.5 68.5 % Drugs need 57.3 54.4 56.8 62.9 % Alcohol need 28.4 29.7 26.5 33.7 % Thinking & Behaviour 62.4 57.0 62.9 52.8 % Attitudes 66.5 64.6 64.4 68.5 420 Days of data available 540 Days of data available Evaluation Control Evaluation Control Previous offences (excluding breach offences) Mean 20.7 20.2 20.4 22.8 Range 1 - 201 1 - 145 1 - 201 1 - 145 % 1 to 5 previous offences 29.3 25.4 30.5 25.7 % 6 to 10 18.8 19.0 21.5 16.2 % 11 to 20 18.5 25.1 15.5 27.2 % 21 to 50 23.1 19.3 23.6 16.2 % More than 50 previous offences 10.3 11.3 9.0 14.7 Mean offences in year prior to release (360 days) 3.4 3.2 3.2 3.4 Mean offences in two years prior to release (720 days) 5.2 5.3 4.9 5.6 Mean offences in five years prior to release (720 days) 9.2 9.2 8.8 9.8 Conviction Occasions Mean number of previous conviction occasions 11.4 10.7 11.2 11.6 Mean number of conviction occasions in two years (720 days) prior to release 3.1 3.0 2.9 3.1 420 Days of data available 540 Days of data available 81 Appendix 5: Diamond referral outcomes Completed (or continuing) 62 (16.8%) Dropped out 59 (10.0%) Moved / resides outside DI area 40 (10.9%) Deported / awaiting deportation 28 (7.6%) Magaged by another agency 21 (5.7%) Outstanding case for serious offence / subsequent over-12- month sentence 10 (2.7%) Other 5 (1.4%) Unable to trace 37 (10.1%) Contact attempted, unable to confirm residence / reciept of offer 26 (7.1%) 167 (45.4%) Ineligible / did not receive offer (reciept not confirmed) 104 (28.3%) Not eligible 63 (17.1%) Unable to trace / contact 80 (21.7%) Refused 201 (54.6%) Eligible / received offer Diamond Offer Referral Outcome Referral Sub-Outcome 121 (32.9%) Engaged (started) 82 Appendix 6: Reoffending and previous offending for UK born and non-UK born subsections of the Evaluation and control groups UK born Non-UK born UK born Non-UK born Evaluation Cohort 52.7% 30.4% 29.7 9.8 Control Group 49.7% 30.7% 27.2 10.5 Evaluation cohort - 'engaged group' (known to be in the UK) 55.0% 42.5% 29.4 14.2 Proven Reoffending Average number of previous offences The non-UK-born portion of the evaluation (and control) groups exhibit markedly different (and generally less prolific) offending characteristics to those born in the UK. On average those born overseas have about a third of the recorded previous offences of those born in the UK38 and most importantly, fewer of those born outside of the UK reoffended during the year following release. This can be accounted for, in part, by deportations and voluntary departures from the UK amongst the overseas-born group, however, given that the difference between the groups is also evident among those ‘engaged’ by Diamond (and therefore known to be in the community), it is evident that other factors are also relevant. The data for this study do not enable this issue to be probed more fully. 38 This may be due to portions of offending careers spent outside of the UK 83 Appendix 7: Examples of categorised offences committed by the evaluation cohort post-release O ffenc e categ or y E xam ple offence s 1 No category 1 o ffences (Murde rs) ap pear in criminal histories of an y cohort m emb er Rob bery S exu al A ssa ult S upplyin g class A d ru gs P ose ssion with in tent to supply cla ss B drug s W o unding A ssault occasion ing ABH B urglary A ffray Taking m otor vehicle withou t consent The ft from pe rson Usin g racially threate nin g... word s o r be haviour A ssault on a constable O utraging public decen cy Drivin g while disqua lif ied The ft from m otor vehicle Com m on a ssault P ossessing controlled drug (class A) S hoplifting Drivin g… with excess a lcoho l P ossessing controlled drug (class C) B eing Drunk an d Disord erly Usin g ve hicle while u nin sured B egging in pu blic p lace 10 Travelling on railwa y without paying fa re 6 7 8 9 2 3 4 5 84 Appendix 8: Proportion of those offered Diamond who accepted by sub groups 64.7 73.3 48.9 70.8 47.8 60.6 66.7 66.7 60.0 51.5 56.1 70.5 71.7 73.2 60.9 65.4 68.8 67.0 72.5 80.0 65.7 63.6 45.5 54.9 61.7 60.0 58.0 61.1 63.6 58.9 64.4 54.0 60.0 58.5 54.8 64.9 61.5 51.9 Southw ark New ham Lew isham Lambeth Hackney Croydon 10 or more 7 to 9 4 to 6 2 or 3 0 or 1 Attitudes Thinking & Behaviour Alcohol Drugs Lifestyle & Associations Relationships ETE Accomodation 7 to 8 5 to 6 3 to 4 0 to 2 Low Medium High Non UK born UK born 45 and over 35 to 44 25 to 34 24 and below Asian Black White Other White British Female Male % Gender Number of OASys needs Reoffending risk group (OGRS3) UK born Age group Ethnicity Team Number of offence in previous two years OASYs need 85 Appendix 9: Forming the participant matched control group The participant-matched control group was extracted from the original, geographically matched group of offenders, according to the following procedure:  For each Diamond participant, select all members of the original control group of the same gender  Compare each of these against the Diamond participant in terms of: o Age o OGRS3 score o Length of criminal career o Age at first conviction o Number of previous offences  For each of these criteria assign a score between 0 and 3, reflecting the level of similarity to the Diamond participant (as set out below) Age (years old) OGRS Score Length of criminal career (years) Age at first conviction (years old) No. of previous offences Score 0 or 1 less than 0.03 0 or 1 0 or 1 0 or 2 3 2 or 3 0.03 to 0.06 2 or 3 2 or 3 3 to 5 2 4 or 5 0.06 to 0.09 4 or 5 4 or 5 6 to 9 1 more than 5 more than 0.09 more than 5 more than 5 more than 9 0 Difference between Diamond participant and control group offender  Sum across the five criteria to produce a score representing the offender’s overall similarity to the Diamond participant.  Select the offender with the highest score as the most similar counterpart for that participant and add to the control group. o If two or more offenders have the same score, select the offender with the closest OGRS score to the Diamond participant39. 39 Duplication was permitted where two or more Diamond participants were matched to the same control group offender. 86 Appendix 10: Diamond participants and participant-matched control group comparison statistics Diamond Participants Offender Matched Control Number 121 121 Gender % Female 11.6 11.6 % Male 88.4 88.4 Ethnicity % BME 42.1 44.6 % White 57.9 54.6 % Not known % White 57.9 54.6 % Black 33.1 24.8 % Asian 5.0 8.3 % Other 0.0 6.6 % Mixed 4.1 5.8 Age (at prison release) Mean 32.5 31.7 % 19 and below 0.8 1.7 % 20 to 24 21.5 20.7 % 25 to 29 26.4 27.3 % 30 to 34 12.4 18.2 % 35 to 39 16.5 13.2 % 40 to 44 10.7 9.1 % 45 to 49 5.8 6.6 % 50 and over 5.8 3.3 Place of Birth % UK born 66.1 65.3 % Non-UK born 33.1 34.7 % Not known 0.8 0.0 Length of Criminal Career Mean (years) - All 10.8 10.0 Mean (years) - UK born 13.0 13.3 Mean (years) - Non UK Born 6.5 3.9 Age at First Conviction Mean age (years old) - All 21.3 21.1 Mean age (years old) - UK born 19.8 18.4 Mean age (years old) - Non UK Born 24.4 26.3 87 Diamond Participants Offender Matched Control Predicted one year reoffending Mean OGRS3 score 0.50 0.48 % High risk 12.4 9.1 % Medium risk 61.2 61.8 % Low risk 23.1 21.5 Not available 3.3 1.7 Identified Needs (OASys) % OASys Available 78.5 52.9 Of which Mean number of identified needs 5.1 5.1 % Accomodation need 61.1 49.1 % ETE need 76.8 77.2 % Relationships need 69.5 61.4 % Lifestyle & Associations need 71.6 82.5 % Drugs need 55.8 64.9 % Alcohol need 31.6 36.8 % Thinking & Behaviour 69.5 64.9 % Attitudes 70.5 68.4 Previous offences (excluding breach offences) Mean 24.2 21.8 Range 1 to 189 1 to 103 % 1 to 5 previous offences 18.2 17.4 % 6 to 10 19.0 19.8 % 11 to 20 24.0 24.8 % 21 to 50 26.4 27.3 % More than 50 previous offences 12.4 10.7 Mean offences in year prior to release (360 days) 3.8 3.3 Mean offences in two years prior to release (720 days) 6.2 5.0 Mean offences in five years prior to release (1800 days) 10.9 9.2 Conviction Occasions Mean number of previous conviction occasions 13.5 12.1 Mean number of conviction occasions in two years (720 days) prior to release 3.7 3.1 88 Appendix 11: Main and Later cohort, six month and predicted reoffending by referral outcome group Actual (6 month) Predicted (year) Actual (6 month) Predicted (year) Actual (6 month) Predicted (year) Actual (6 month) Predicted (year) Later cohort (Jan - Mar 2010) 19.1% N=42 47.6% 21.3% N=47 45.7% 16.7% N=48 46.1% 19.0% N=137 46.4% Main cohort (pre mid Oct 2009) 35.5% N=121 49.5% 27.5% N=80 45.7% 21.6% N=167 39.5% 27.5% N=368 44.2% Full cohort Standardised six-month proven reoffending rate Engaged (received Diamond) Eligible not engaged (refused Diamond) Not eligible / not traceable (did not receive offer) 89 Appendix 12: Staff survey results (selected questions) To what extent are the following partners involved in the DI on your borough? Consistently Involved Involved when required Not involved W1 W2 W3 W4 W1 W2 W3 W4 W1 W2 W3 W4 Police 98% 97% 97% 97% 0% 3% 2% 3% 0% 0% 0% 0% Probation 84% 97% 91% 86% 8% 3% 8% 14% 6% 0% 0% 0% Prison 28% 18% 16% 26% 67% 81% 78% 70% 4% 1% 2% 4% Community Safety Partnership 4% 9% 9% 10% 53% 63% 78% 53% 41% 28% 9% 36% Local Authority 45% 60% 70% 70% 49% 36% 28% 27% 4% 4% 0% 3% Voluntary Groups 4% 19% 19% 12% 78% 81% 73% 86% 16% 0% 6% 2% Housing 18% 34% 44% 67% 75% 61% 53% 33% 6% 4% 0% 0% Education 2% 15% 14% 5% 82% 72% 78% 92% 14% 12% 3% 3% Youth Offending Teams 0% 2% 3% 0% 57% 72% 66% 70% 41% 27% 27% 30% Drug Intervention/ Drug Action Team 24% 19% 41% 39% 71% 76% 55% 61% 4% 3% 0% 0% Mental health 2% 2% 11% 9% 82% 82% 75% 86% 14% 16% 9% 5% Health 4% 2% 3% 3% 80% 90% 84% 89% 14% 9% 6% 8% How confident are you that you understand your role on the DI? Wave 1 Wave 2 Wave 3 Wave 4 Very confident 45.1% 62.7% 60.9% 75.8% Fairly confident 51.0% 32.8% 32.8% 24.2% Not very confident 3.9% 3.0% 3.1% 0.0% Not at all confident 0.0% 1.5% 0.5% 0.0% To what extent do you agree that you have received sufficient training to do your job on the DI team effectively? Was training sufficient? Wave 1 Wave 2 Wave 3 Wave 4 Strongly agree 6.0% 9.1% 17.7% 21.2% Agree 40.0% 50.0% 38.7% 39.4% Mixed views 44.0% 34.8% 30.6% 34.9% Disagree 10.0% 4.5% 8.1% 3.0% Strongly disagree 0.0% 1.5% 4.8% 1.5% Is your DI team fully staffed? Wave 1 Wave 2 Wave 3 Wave 4 Yes 7.8% 11.9% 10.9% 28.8% No 88.2% 88.1% 87.5% 71.2% 90 Are all these staff dedicated to the DI team? Are all staff dedicated? Wave 1 Wave 2 Wave 3 Wave 4 Yes 74.5% 83.6% 82.8% 63.6% No 25.5% 16.4% 14.1% 36.4% To what extent do you agree that you have received sufficient training to do your job on the DI team effectively? Was training sufficient? Wave 1 Wave 2 Wave 3 Strongly agree 6.0% 9.1% 17.7% Agree 40.0% 50.0% 38.7% Mixed views 44.0% 34.8% 30.6% Disagree 10.0% 4.5% 8.1% Strongly disagree 0.0% 1.5% 4.8% How did you get involved in the DI? How did you get involved in the DI? W4 I applied / volunteered for the role 84.9% I was posted onto Diamond 15.1% Has working on the DI been what you expected? Has working on the DI been what you expected? W4 Yes 75% No 25% How would you rate the current level of morale in your team? W4 Very High 10.6% High 18.2% Neither high nor low 37.9% Low 18.2% Very low 15.1% Has this changed over time? W4 Yes 77.3% No 22.7% 91 Appendix 13: Conceptual comparison of DI and an IOM approach Knife Crime GPS Tagging on licence programme – Suitability checklist This form is used to confirm the eligibility for the MOPAC (Buddi) GPS tagging for knife crime programme. When deciding on licence conditions, please remember that all licence conditions should be necessary and proportionate to the assessed risk in each case. GPS monitoring should also be tailored to the individual. If you have any questions or queries on individual cases or need advice on these decisions, please contact londonPS.buddigps@justice.gov.uk Next steps for suitable cases • Use the Create and Vary Licence (CVL) process to propose GPS using approved GPS licence condition wording. • Contact Buddi to arrange tagging (use Buddi tag notification form). • Copy in londonPS.buddigps@justice.gov.uk to ensure arrangements are confirmed. • Inform the offender as soon as the referral has been accepted that they will be subject to GPS on licence – do this by sharing the “Wearer Introduction” document – either in a pre-release meeting/contact or via the OMU. 1 Eligible releasing prisons - Belmarsh, Brixton, Bronzefield, Coldingley, Downview, Feltham, High Down, Highpoint, Isis, Onley, Pentonville, Send, Thameside, Wandsworth, Wormwood Scrubs Essential criteria for the offender: Aged 18+ Tag fitting date is between Feb 2019 – 31/03/2026 Serving a standard determinate sentence for either a knife possession offence or another offence which involved the use of a knife or bladed object Being released to a London borough Being released from a pilot prison1 Being released to some type of viable accommodation (to enable tag charging) Under which licence condition will you propose GPS monitoring: Attend drug assessment Reporting to Approved Premises or Police station Attend drug test Exclusion from area/zone or address To comply with any requirements to ensure that you address your alcohol/ drug/violent/anger offending behaviour problems at [COURSE / CENTRE]. Trail monitoring (heat maps, interest and inclusion zones, monitoring of doorstep curfews, data on request) Youth to Adult (Y2A) Hub: Transitions to Adulthood pilot evaluation June 2023 Interim Performance Report Zivile Jakaite, Gemma Hennessey & Thomas Wilkinson Y2A hub: Transitions to Adulthood pilot evaluation summary Key findings from the initial funding period (October 2021 to February 2023): • 415 service users accessed the hub, with an average of 218 service users supported each month. • They are predominantly male, have an average age of 22 & belong to a minority ethnic background. • 40% of service users served a custodial sentence, 30% a community order & 30% a suspended sentence. • Over half (54%) of service users most recent offence was violence or drug-related. • Nearly half (48%) of service users are required to participate in a rehabilitative activity as part of their sentence & a third (35%) to take part in unpaid work. • Most hub service users have ‘Thinking & Behaviour’; ‘Attitudes’; & Education, Training, & Employment (ETE) needs. • Average OGRS scores show hub service users had a 44% chance of reoffending in the first 2 years on release or start of community sentence. • More than half of service users were referred to at least one hub-commissioned service (& nearly a third to two or more). • Overall, just over a quarter of cases referred to commissioned services were closed due to drop-put or lack of engagement. • Service users received a range of support, but some needs may not have been met. • The characteristics of those not referred to any hub services were not statistically different to those that were. • Mentoring & Speech/Language Therapy (SLT) services have the largest caseloads – supporting a quarter of hub service users. • 2,240 hub-commissioned service sessions were delivered in the initial funding period, half of these by the mentoring service. • In general, only a small proportion of service users are reported as ‘not at all motivated’. • Data on outcomes is being captured that will eventually be able to compare service users across a range of progress measures. Illustrations of this are included in the current summary. • Breach & recall numbers were low & most breaches were withdrawn. MOPAC’s Evidence & Insight Unit are conducting a performance and impact evaluation of the Y2A hub coinciding with a process evaluation conducted by a team of externally- commissioned academics from Middlesex and Sheffield Hallam University. This report analyses performance data from probation and the eight hub-commissioned services for the initial funding period of 16 months, from October 2021 to end of February 2023. The Youth to Adult (Y2A) hub in Newham seeks to break the reoffending cycle of young adults through tailored, wrap-around support services. The hub co-locates varied commissioned services and statutory support agencies alongside Probation services for 18-25 year olds on probation in Newham. Project background Funding & governance: Pilot was initially co-funded until March 2023 by His Majesty’s Treasury via the MoJ, to a sum of £3.19 million. The hub has been extended by another year, with funding until March 2024 from MOPAC, London Probation, Barrow Cadbury Trust, DLUC, London Borough of Newham and the MoJ. The hub is governed by the Transitions Programme Board, made up of: MoJ MOPAC Probation service NHS London Borough of Newham The Y2A hub in Newham sits within this evidence base to test if intervening early in the lives of young adults & providing dedicated provision in the form of the hub model offers a significant opportunity to prevent individuals from becoming entrenched in criminality. Reducing reoffending among young adults on probation is a priority for the Mayor’s Office for Policing and Crime (MOPAC) & the Ministry of Justice (MoJ). • Several reports highlight the failures of existing support for this group, especially in facilitating transitions (HMIP, 2016) and recognising maturity (Justice Select Committee, 2016), with significant consequences for desistance. • Literature suggests effective community support for young adults on probation requires: tailored, trauma, and maturity-informed services that address multiple needs (Livingston et al., 2015); specialist support for women (Young Women’s Justice Project, 2021); mental health and communication screening (Justice Select Committee, 2016); structural support of accommodation and employment. • Intervening at the young adulthood stage is critical and is a window of opportunity for impact: • 18-25 year olds are disproportionately involved in the criminal justice system - over a quarter (27%) of adult offenders on probation are aged 18-25 (MoJ, 2019), with a quarter of these expected to reoffend within a year (MoJ, 2023). • Research suggests that many young people can become trapped in a reoffending cycle once in the criminal justice system, with convictions often for minor, non-violent crimes (e.g. Borysik, 2020). However, young adults are the most likely age group to ‘grow out’ of crime and desist long-term (Livingstone et al., 2015). • Literature suggests 18-25 year old offenders need to be treated differently with specialised support services (e.g. Mason and Prior, 2008): • Low maturity levels: It is widely accepted that young people’s brains are still developing until the age of 25; behavioural consequences include emotional regulation and propensity to take risks (Harris and Edwards, 2023). Maturation can be hindered by adverse childhood experiences, substance use, and psychiatric or neurodevelopmental disorders (ibid.). • Difficulties transitioning from child to adult support services: Turning 18 can represent a ‘cliff edge’ to young people as support teams change and availability reduces, meaning young people can fall through the cracks (Bennett and Corry-Roake, 2021). [EMBARGOED UNTIL MOPAC SIGNOFF FOR PUBLICATION] The Hub operating model November 2021 October 2021 Mentoring service (Spark 2 Life) Speech and Language Therapy service (ELFT) Housing service (Step Ahead) Young Women’s support (Advance Charity) March 2022 Restorative Justice and Mediation service (Belong) April 2022 Mental health service (ELFT) Substance misuse service (Change Grow Live) Meaningful activities (OMG Education) Note: Meaningful activities service was paused between September 2022 till the end of December 2022 due to contract- related matters. Y2A hub includes 8 commissioned services: • The Y2A hub in Newham co-locates commissioned services and statutory support agencies alongside Probation. • It provides wrap-around support, tailored to young adults’ distinct needs through a trauma and maturity informed approach. • The Y2A hub ultimately seeks to tackle young adults reoffending by breaking the reoffending cycle (e.g. Clay et al., 2021) through its tailored and expansive wrap-around support services. • The hub aims to: Encourage compliance Promote desistance Improve life outcomes Make communities safer Reduce impact of reoffending on the public sector December 2021 Source: Phillips et al. (2022) The hub offers: • A welcoming trauma-informed environment where young adults engage with co-located support services. • A multi-disciplinary team of staff, trained to work effectively with young adults. • Wrap-around support for young adults. • Support for probation staff for tailored case formulation. • Supported transitions and a bespoke approach for transitions cases. • Bespoke gender-specific offer for young women. • User engagement service. • Integration of existing probation-commissioned ETE ‘Maximus’ service. • On-site support from: • DWP job centre link worker – at hub once a week to provide advice on benefits and employment • Local food bank • Sexual health clinic [EMBARGOED UNTIL MOPAC SIGNOFF FOR PUBLICATION] Existing service commissioned by probation available to Y2A hub service users Hub service flow DWP Work Coach job centre link worker Young adults aged 18-25 years old on probation in Newham • Males and females • Community and custody cases (6 months pre- release) • Individuals aged between 17 years and 6 months - 18 years • Transitioning from Newham YOS to adult probation Case formulation (by PO) Case input forum Screening for speech and language needs Mental health team review Y2A HUB Women’s PO Probation Officer (PO) Probation Support Officer (PSO) Young women’s support Restorative justice and mediation Meaningful activities Housing service Speech and language therapy Mental health Substance misuse Mentoring and coaching Case closure (end of order/ sentence) Appointments with probation officer Referrals to hub services Training for all HUB staff (Interface Enterprises) Appointments with hub services User engagement service Case allocation (based on risk levels/ gender) Education, Training and Employment service Pilot evaluation Performance analytics Impact evaluation Process evaluation MOPAC E&I are conducting an evaluation of the Y2A hub alongside a team of academics (commissioned by the MoJ) from Middlesex & Sheffield Hallam University. The team of academics will conduct the ‘process’ evaluation, using qualitative methods to understand how the hub is being delivered & context. Y2A Hub evaluation MOPAC will conduct the performance and impact evaluation. Key objectives for MOPAC's evaluation are: • To monitor the throughput of the programme in terms of number of individuals; their demographics; needs; services received etc. • To assess impact of the initiative – particularly on reoffending (breaches; recalls & police contact). This report analyses performance data from probation & the 8 commissioned services for the initial funding period of 16 months - between October 2021 to end of February 2023. This report covers the following three areas: – Limitations to analysis, please note this report makes it clear where there is missing data. The process evaluation interim report was completed in March 2023 (Philips et al. 2023). Impact analysis will be conducted at a later stage: • November 2023 – looking at reoffending of hub service users up until the end of February 2023 • November 2024 – looking at reoffending of hub service users up until the end of February 2024 Service user demographics, risks and needs Hub throughput Service user outcomes Total cohort size and total number of service users supported at the hub: Initial funding period October 2021 to February 2023 415 service users were supported by the hub in the initial funding period In the first funding period: • 415 service users were supported by the hub. • The Y2A hub supported an average of 218 service users a month (SD=22.5). Looking forward to the end of the next funding period (end of Feb 2024): • Additional 190 service users expected to be supported, meaning approximately 620 service users forecasted to be supported by the hub. • On average, 19 new service users joined the hub each month: From 5 service users (in February 2023) to 33 service users (November 2021) • An average of 16 service users leave the hub each month: From 5 service users (in July 2022) to 31 service users (April 2022) 183 202 201 208 217 207 216 233 236 235 238 248 241 229 231 218 162 195 224 234 246 271 290 308 335 340 347 358 377 391 398 410 415 145 195 245 295 345 395 445 495 545 595 10/21 12/21 02/22 04/22 06/22 08/22 10/22 12/22 02/23 04/23 06/23 08/23 10/23 12/23 02/24 Total cohort size each month Total number of service users that have been supported at the hub to date Estimate of how many will be supported in total Churn of Y2A service users March 2024: approx. 620 Month Left Newly Joined Remained on the caseload Oct-21 N/A N/A N/A Nov-21 12 33 150 Dec-21 10 29 173 Jan-22 11 10 191 Feb-22 8 15 193 Mar-22 16 25 192 Apr-22 31 21 186 May-22 20 29 187 Jun-22 22 39 194 Jul-22 5 8 228 Aug-22 14 13 222 Sep-22 13 16 222 Oct-22 10 20 228 Nov-22 28 21 220 Dec-22 21 9 220 Jan-23 14 16 215 Feb-23 18 5 213 Note on use of probation data This report draws on probation data. Y2A hub service users are extracted from probation data as follows: • 18-25 year olds on probation in Newham • Has a Y2A probation practitioner allocated • Is in the community or in custody with less than 6 months left until release. This is because individuals near release are likely to already be in contact with the hub. Hub throughput [EMBARGOED UNTIL MOPAC SIGNOFF FOR PUBLICATION] Hub service users are predominantly male, aged 22 & belong to a minority ethnic background Average age 22 years old 73% belong to a minority ethnic background Of the 415 hub service users supported in the initial funding period (until end of February 2023): 16 recorded as care-leavers Comparison of hub service users & 18-25 year olds on probation overall in London Age* Mean SD Overall in London (N=7766) 22.3 2.00 The hub (N=415) 21.7 2.00 Ethnicity Asian or Asian British Black or Black British White Mixed Other ethnic group Missing N % N % N % N % N % N % Overall in London (N=7766) 783 10.1% 2679 34.5% 2021 26.0% 799 10.3% 392 5.1% 1092 14.1% The hub (N=415) 124 29.9% 118 28.4% 76 18.3% 41 9.9% 20 4.8% 36 8.7% Borough of Newham in general*** 351,030 42.2% 148,187 17.5% 107,947 30.8% 16,419 4.7% 7,175 4.9% N/A N/A *For the hub service users age was calculated at the end of the month when the service user first appeared on the probation caseload **Please note pan-London probation data was pulled in March 2023 *** As per data available from ONS 2021, all age groups. Y2A Hub (N=415) How does the Y2A cohort in Newham compare to 18-25 year olds on probation in London?** • Higher proportion of Asian or British Asian (30%) compared to London probation overall (10%). Slightly lower proportion of Black or Black British & White identifying service users in Newham compared to London overall. • Compared to Newham in general (all age groups) there is a higher proportion of Black individuals in the Y2A hub (28%) compared to 18%. 93% are male 33% 31% 20% 11% 5% Asian or Asian British Black or Black British White Mixed Other ethnic group 31, 7% 384, 93% Female Male 424, 5% 7342, 95% Female Male Overall in London (N=7766) Ethnicity breakdown of hub service users Service user demographics, risks & needs [EMBARGOED UNTIL MOPAC SIGNOFF FOR PUBLICATION] Most hub service users have ‘Thinking & Behaviour’, ‘Attitudes’ & ETE needs Over 80% (N=336) of hub service users had at least 3 needs The average number of needs per hub service user was 4 (SD=2.1) Criminogenic needs of hub service users (N=415)* *data was missing for 52 service users Note: this slide includes probation needs and risk assessment data from October 2021 to February 2023. For each service user needs/risk related data was reported as recorded at baseline. Needs/risk related pan-London data was not available for young people on probation, for this reason, we were not able to provide it in this report. Offender Group Reconviction Scale (OGRS): measure of risk of re-offending OGRS score is a general predictor of re-offending ranging from 0 to a 100. A lower score indicates a lower risk of re-offending. OGRS scores for hub service users are typically recorded at baseline when a young person starts working with a Probation Practitioner. The OGRS 2 scale is reported on here – referring to chance of reoffending within a 2 year period (MoJ, 2019). Average OGRS score was 44.1 (SD=24.6): On average service users have a medium chance (44%) of reoffending within two years of release into custody, or start of community sentence. Thinking & Behaviour (83%, N=346), Attitudes (78%, N=324) and Education, Training, and Employment (ETE) (68%, N= 284) were the most commonly recorded criminogenic needs among service users. The average OGRS score of Y2A Hub users indicates a slightly higher likelihood to re- offend within two years of release into custody or start of community sentence compared to 18-25 year olds on probation in London overall – 44 compared to 41 for London overall. Service user demographics, risks & needs Hub throughput Hub throughput DWP Work Coach job centre link worker Young adults aged 18-25 years old on probation in Newham (N=567) Case formulation (by PO) Case input forum Screening for speech & language needs Mental health team review Y2A HUB N=415 Women’s SPOC PO PO PSO Young women’s support N=15 Restorative justice and mediation N=9 Meaningful activities N=32 Housing service N=66 Speech and language therapy N= 108 Mental health N=86 Substance misuse N=15 Mentoring and coaching N=105 Case closures (N=253) Appointments with probation officer Referrals to hub services Training for all HUB staff (Interface Enterprises) N=35 Appointments with hub services User engagement service Case allocation (based on risk levels/ gender) Education, Training and Employment service No referrals to commissioned services N=195 Recall/ breach/ resentenced 170 ‘engagements’ with service users between January 2022 and December 2022. Note: this number refers to case closures, not unique service users, e.g. some service users have left the hub and re-joined – they may be counted twice. When we exclude hub service users unlikely to have been referred to services, the proportion referred to at least one service rises to 58%. These excluded service users are those who: • Joined in February 2023 (N=5) (unlikely to have been referred to services yet). • Were still in custody in February 2023 (less likely to be referred to services). • Left the hub before the end of December 2021 and were not referred to any services (when only two hub service providers were in operation, see timeline here). 195 were not referred to any of the eight hub- commissioned services. Over half of service users were referred to at least one hub-commissioned service * This figure may be higher, as there were some service users who were on the service providers’ caseloads (approx. N=19) but not on the probation caseload. ** Note: we have data on probation practitioner appointment attendance from February 2023 onwards. In February 2023, 80% (N=170) attended an appointment with their probation practitioner.** Nearly a third of service users were referred to two or more hub- commissioned services. No referrals to hub services, 195, 47% 1 referral, 98, 24% 2 referrals, 68, 16% 3 referrals, 38, 9% 4 referrals, 14, 3% 5 referrals, 2, 1% Number of hub service referrals per service user (N=415) Note: alongside the eight hub-commissioned services, hub service users have access to a pre-existing probation-commissioned education, training & employment service for which we do not have data for. It could be that hub service users not referred to one of the eight hub services participated in the ETE service. Note: when further excluding hub service users who left the hub before the end of April 2022 (before the end of the mobilisation phase), the proportion referred to at least one service rises to 64%. 53% were referred to at least one hub- commissioned service (N=220).* Hub throughput Hub service users not referred to commissioned services have similar needs to those who were Hub throughput There were some differences by gender & ethnicity between those referred / not referred to services, but these differences were not statistically significant: • Female service users slightly more likely to be referred to at least one service provider (8.8% to 5.7%) • Asian or Asian British & Black of Black British service users are slightly more likely to be referred to at least one service provider (see here for details on statistical tests used). Age Mean SD Not referred to any hub services 21.8 2.0 Referred to at least one service provider 21.6 2.1 Ethnicity Asian or Asian British Black or Black British White Mixed Unknown Other ethnic group N % N % N % N % N % N % Not referred to any hub services (N=158) 48 30.4% 39 24.7% 32 20.3% 13 8.2% 19 12.0% 7 4.4% Referred to at least one service provider (N=215) 69 32.1% 65 30.2% 39 18.1% 21 9.8% 9 4.2% 12 5.6% Comparison of criminogenic needs within the hub • There were no statistically significant differences between the two groups in the average number of probation- assessed needs, proportion of service users reported to have certain needs, nor risk of reoffending (OGRS). • There was an average of 4.0 (SD=2.0) needs per service user who had been referred to at least one service provider compared to 3.9 for those not referred. • For eight out of the ten need categories, the proportion of service users requiring support for that need was slightly higher for those who accessed support from service providers compared to those who did not. OGRS score Mean SD Not referred to any hub services (N=58) 43.7 24.9 Referred to at least one service provider (N=215) 43.0 24.5 6% 94% Not referred to any hub services (N=158) Female Male 9% 91% Referred to at least one service provider (N=215) Female Male 6% 14% 13% 17% 17% 21% 34% 71% 81% 85% 6% 10% 9% 11% 15% 23% 32% 66% 75% 82% Alcohol Accommodation Emotional Wellbeing Relationships Drugs Finance Lifestyle ETE Attitudes Thinking and Behaviour Not referred to any service providers Referred to at least one service provider Gender This analysis excludes those service users that were unlikely to be referred from the ‘not referred to any hub services’ group (joined in Feb 2023, still in custody in Feb 2023, left hub before Dec 2021). Sentence type Sentence and offence types are similar between those who were and were not referred to services Offence type Theft, burglary, robbery and fraud Motoring Drug-related Violence Sexual Public Order Other Offence N % N % N % N % N % N % N % Not referred to any hub services (N=158) 29 18.4% 22 13.9% 39 24.7% 48 30.4% 6 3.8% 5 3.2% 9 5.7% Referred to at least one service provider (N=215) 33 15.4% 24 11.2% 49 22.8% 67 31.2% 12 5.6% 15 7.0% 15 7.0% There were some differences in sentence type between service users who have not & have been referred to at least one hub service provider, but these were not statistically significant: • Community order (31% not referred compared to 27% referred) • Suspended sentence (27.9% not referred compared to 32.6% referred). Service users not referred to any hub services were very slightly more likely to have committed a theft, burglary, robbery & fraud offence (18.4% compared to 15.4%) as well as drug-related offences (24.7% compared to 22.8%) & less likely to have committed a public order offence (3.2% compared to 7.0%). Again, statistical tests found that offence type was not related to being referred to a hub service or not. 40.0% 41.4% 27.0% 31.0% 32.6% 27.9% 0.5% 0% 50% 100% Referred to at least one service provider (N=215) Not referred to any service providers (N=158) Custodial Community order Suspended sentence Other more investigation is needed into staff decision making, referral processes, thresholds of needs, readiness of service users, and consent among hub service users to understand why service users who were and were not referred to any hub services are similar in terms of demographics, risks, needs and sentence/offence type, Hub throughput 23 30 8 16 10 35 16 8 315 94 90 74 109 96 156 Restorative Justice and Mediation Young Women's Support Substance Misuse Meaningful Activities Accommodation Mental Health Speech and Language Therapy Mentoring Average number of days between referral date and first session Average number of days case is open Mentoring and Speech & Language Therapy have the largest caseloads At the end of February 2023: Mentoring was the largest service with 108 referrals, followed by Speech / language therapy with 105 referrals: * Time from point of referral to first session – calculated as the difference (in days) between the day of referral and the day when service user started receiving support from the service. Some service users were referred while in custody but their support didn’t commence until their release (as not possible) which may have impacted waiting times for some service users. At the end of February 2023: • Speech / Language Therapy service has the highest active caseload (N=52), followed by the Mentoring service (N=38). • Approximately a third of cases are still ongoing. • At least a third of the cases were closed due to completion for five out of eight service providers. • Overall just over a quarter of cases were closed due to drop-put or lack of engagement. Active cases and case closure (end of February 2023) Average time (days) from point of referral to first session, and to case closure First session 9 15 15 32 66 86 105 108 Restorative justice and mediation Young women's support Substance misuse Meaningful activities Housing Mental health Speech and language therapy Mentoring Total number of service user referrals to each commissioned service in initial funding period Restorative Justice and Mediation (N=9) Young Women's Support (N=15) Substance Misuse (N=15) Meaningful Activities (N=32) Housing (N=66) Mental Health (N=86) Speech and Language Therapy (N=105) Mentoring (N=108) Ongoing Closed due to completion Closed due to drop-out/lack of motivation Closed - Other Closed- referral not appropriate Unknown N=38 N=34 N=33 N=52 N=43 N=10 N=25 N=9 N=32 N=16 N=9 N=33 N=24 N=5 N=9 N=10 N=8 N=6 N=2 N=7 N=6 N=6 N=3 N=9 • On average, service users waited 18.2 days to be seen for the first time from the point of referral (some service users are referred while still in custody & only receive support on release). • Mental health service (35 days) & Young women’s support (30 days) had the highest waiting times. • On average, the Young women’s service supported service users for the longest (315 days), before case closure. • All Restorative justice cases remain open. • Please note: these figures include service users whose cases were closed due to completion or drop- out/lack of engagement. Hub throughput 2,240 commissioned service sessions were delivered in the initial funding period • Across all services, 2240 sessions were delivered by the end of February 2023. This includes sessions conducted face-to-face & over the phone. • The mentoring service delivered the highest number of sessions across all service providers (N=1130), potentially because it was the first service to start operating at the hub. • While the Young Women’s Support & the Restorative Justice / Mediation services have a relatively low number of total sessions (114 & 71 sessions respectively), they have a relatively high average number of sessions per service user (7.6 & 7.9 respectively) Mean number of sessions delivered per service user (to date) Mean number of sessions delivered per service user, per month Total number of sessions delivered by the end of February 2023 Note: data is not available for all services in all months, sometimes due to issues like inconsistencies in data entry by service providers 71 114 30 103 371 171 250 1130 Restorative justice and mediation Young women's support Substance misuse Meaningful activities Accommodation Mental health Speech and language therapy Mentoring 10.4 2.4 6.1 5.6 3.2 2 7.6 7.9 0 1 2 3 4 5 6 7 8 May-22 Jun-22 Jul-22 Aug-22 Sep-22 Oct-22 Nov-22 Dec-22 Jan-23 Feb-23 Mentoring Speech and Language Therapy Mental Health Accommodation Meaningful Activities Substance Misuse Young Women's Support Restorative Justice and Mediation Hub throughput Mentoring service (N=108) Speech and language therapy service (N=105) Mental health service (N=86) Service users received a range of support, but some needs may not have been met Meaningful activities service (N=32) 19 8 63 2 64 75 20 81 3 3 3 3 Formal therapy Creative therapy Informal therapy Psychological assessment Therapies received by mental health service users (N=86) Yes No Missing 12 service users attended at least one meaningful activity during their engagement with the service Is support matched to needs? From the performance data: • 50 service users were recorded as having emotional wellbeing needs. Of those, 9 (18%) were referred to the Mental Health service. • 54 service users were identified as having housing- related needs. Of those, 14 (26%) were referred to the accommodation service. • 84 service users were recorded as either having needs related to alcohol or to drugs, but only 4 (5%) of those service users were referred to the substance misuse service. • 346 service users were assessed as having needs related to ‘Thinking & behaviour’, 77 of which were referred to the mentoring service (22.3%). • 324 service users were recorded as having needs related to ‘Attitudes’, & 76 (23.5%) of those service users were referred to the mentoring service. More investigation is needed to understand why some service users do not receive services despite an indication of need in the data. Other factors not measured in the data could include staff decision making, referral processes, thresholds of needs, readiness of service users, and consent among hub service users. * Data was missing for 60 out of 105 service users At least 25* service users were recorded as having received a communications passport 19 attended at least one session of Aspire Higher programme, and 8 service users fully completed it 71 (65%) service users received support related to ETE 34 service users received some type of training related to mentoring Half (16) of referrals to the meaningful activities service did not attend any activities – a third of these dropped out due to non-engagement. On average, engaged services users attended three meaningful activities (M=2.8, SD=1.1). *Out of a total of 32 service users data was missing for 4 Majority of service users (63%) referred to the mental health service received informal therapy: Hub throughput A small proportion of service users tend to be reported as ‘not at all motivated’ Motivation levels for service users across service providers in February 2023 • In February 2023, at least a quarter of service users receiving hub services were recorded by the providers as highly motivated. • Monthly trends in levels of motivation have been relatively stable, with a small proportion of service users tend to be reported as ‘not at all motivated’ How are motivation levels measured? Each month, service providers are asked to assign a motivation score for each service user on the caseload. Motivation levels are scored on a Likert-type scale: ‘Not at all motivated’; ‘Somewhat motivated’ and ‘Highly motivated’. This measure should be used with care due to subjectivity & is most useful for tracking changes within services over time, not comparing across services due to the subjective nature of the measure. 0 20 40 60 80 100 % of service users reported to be ‘highly motivated’ for each service, per month Mentoring Speech and Language Therapy Mental Health Accommodation Meaningful Activities Substance Misuse Young Women's Support Restorative Justice and Mediation 0 20 40 60 80 100 % of service users reported to be ‘not at all motivated’ for each service, per month Mentoring Speech and Language Therapy Mental Health Accommodation Meaningful Activities Substance Misuse Young Women's Support Restorative Justice and Mediation Substance Misuse (N=7) Young Women's Support (N=8) Restorative Justice and Mediation (N=9) Meaningful Activities (N=10) Accommodation (N=12) Mental Health (N=28) Mentoring (N=41) Speech and Language Therapy (N=59) Not at all motivated Somewhat motivated Highly motivated Missing N=3 N=5 N=6 N=10 N=1 N=1 N=14 N=7 N=3 N=6 N=1 N=8 N=13 N=26 N=19 N=2 N=28 N=1 N=10 N=1 N=1 N=3 N=3 Hub throughput Mentoring, Speech & Language Therapy, & Accommodation services had the highest number of case completions N=23 N=12 33 43 9 33 10 2 6 0 34 10 32 24 8 7 3 3 20 5 Closed cases Closed due to other, inc. referral not appropriate Closed due to drop out Closed due to completion • The number of closed cases due to completion range from 0 cases for Restorative Justice & Mediation, to 43 completions for Speech & Language Therapy. • All services record some form of outcome measure for each service user at two time points: on referral & on case completion, in order to measure distance travelled during contact with the service. This service-specific outcome information is reported on the following slides. Service user outcomes Data is being captured on a variety of outcomes which will eventually provide rich insights on service user progress • The table across shows ETE status upon referral and latest assessment for all service users as well as case completions. • These numbers reflect on the achievements of the individual service users, and are not currently indicative of programme success. Service user outcomes ETE status recorded by the mentoring service All service users (N=108) Case completions only (N=33) On referral Latest assessment On referral On completion NEET 76 48 21 11 In education 10 8 5 5 In training 3 9 1 2 Working 17 27 6 13 Apprenticeship 1 N/A N/A N/A Missing 1 16 N/A 2 Commissioned hub services are collecting data on a range of service specific outcomes for service users. The number of service users included in data for each service is too small to provide conclusions at this stage. This will be a focus for later reports on this evaluation. Service user outcomes for the two services with data for the largest number of service users are presented below as an illustration. • The table shows Accommodation service outcomes across referral and latest assessment for all individuals referred, as well as case completions. • These numbers reflect on the achievements of the individual service users, and are not indicative of programme success. Accommodation service outcomes Housing status: All referred service users (N=66) Case completions only (N=33) On referral Latest assessment On referral On case completion Homeless 3 2 1 N/A In an Approved Accommodation 8 5 4 N/A In custody 18 2 8 N/A In transient/short term accommodation 2 4 2 4 With friends/family 28 28 14 10 Supported housing N/A 1 N/A 1 Rented accommodation 7 24 4 18 Breach and recall numbers are low, and most breaches are withdrawn • 87 breaches were recorded at the hub between October 2021 - February 2023. • On average, five service users breached their suspended sentence, community order or post-sentence supervision each month, with no breaches recorded in October 2021 & February 2022 & 14 in December 2022. • More than half (47) of breaches were withdrawn. Notably, 11 out of 14 breaches were withdrawn in December 2022 & all eight breaches recorded in January 2023 were withdrawn. Number of recorded breaches at the hub: December 2021 to end of February 2023 Number of recorded recalls at the hub: October 2021 to end of February 2023 1 0 0 0 0 1 2 2 1 0 3 0 2 0 0 0 0 0 1 2 3 4 6 5 1 0 4 7 5 11 5 3 8 4 1 14 8 5 0 2 4 6 8 10 12 14 16 Breach withdrawn Breach continued • 12 service users were recorded as recalled at the hub between October 2021 - February 2023. • No recalls were recorded in 10 of the 17 months during this reports data collection time-frame. Service user outcomes Appendix Statistical tests used in analysis When comparing service users who have not been referred to any hub services versus those who have been referred to at least one service, statistical tests were run during analysis to determine whether or not any differences found among the two groups were statistically significant or not. Two types of statistical tests were used: • Chi-square test for categorical variables. • Mann Whitney U test for discrete variables (chosen due to data being non-normally distributed). For every test used, a confidence level of 95% (or a p-value of > 0.05) was chosen, as is standard practice. Where significance was found, data is labelled with a *. References • Bennett, L. and Corry-Roake, E. (2021) Evidence review: diverting young adults away from the cycle of crisis and crime. Revolving Doors Agency. • Borysik, B. (2020). New Generation: Preventing young adults being caught in the revolving door. Revolving Doors Agency. • Clay, T., Reynolds, S. and Rose, A. (2021). Breaking reoffending cycles in the criminal justice system: mapping causal factors, leverage points and funding flows. NPC. (Available via: Breaking-reoffending-cycles-in-the-criminal-justice-system-report.pdf (thinknpc.org) • Harris and Edwards, 2023 Young people in transition in the criminal justice system Evidence review • HMIP (2016) Transition Arrangements: a follow-up inspection: An Inspection by HM Inspectorate of Probation. HMIP: Manchester • Justice Select Committee (2016) The Treatment of Young Adults in the Criminal Justice System: Seventh Report of Session 2016-17, HC169.UK: House of Commons. • Livingston, I., Amad S., and Clark L. (2015). Effective Approaches With Young Adults: A Guide for Probation Services. Barrow Cadbury Trust: London. • Ministry of Justice (2019) Probation caseload 30th September 2019 FOI_200326008_probation_caseload_by_age_group_and_ethnicity__tables_.ods (live.com) • Ministry of Justice (2019) Guide to proven reoffending statistics. MoJ: London. guide-to-proven-reoffending-statistics-Apr19.pdf (publishing.service.gov.uk) • Ministry of Justice (2023) Proven reoffending statistics: April to June 2021 Proven reoffending statistics: April to June 2021 - GOV.UK (www.gov.uk) • Phillips, J., Ward, J., Albertson, K., Duke, K., Fowler, A., Cracknell, M and Riley, L. (2022). Evaluation of the Newham Transition to Adulthood Hub Pilot: Interim report August 2022. Middlesex University London & Sheffield Hallam University. • Phillips, J., Ward, J., Albertson, K., Duke, K., Fowler, A., Cracknell, M and Riley, L. (2023). Evaluation of the Newham Transition to Adulthood Hub Pilot: Second Interim report March 2023. Middlesex University London & Sheffield Hallam University. • Wong, K. K. S., Kinsella, R., Bamonte, J., and Meadows, L. (2017). T2A Final process evaluation report. (Available via: T2A-Final-Process- Report-OCTOBER-2017.pdf) • Young Women’s Justice Project (2021) FALLING THROUGH THE GAPS. Falling-through-the-gaps-YWJP-transitions-briefing-paper-2.pdf (t2a.org.uk) 1 Report Author: Barry Charleton Contributors: Adele Harrison, Paul Dawson 2 Executive Summary The Mayors Office for Policing And Crime (MOPAC) and the Metropolitan Police Service (MPS) are working in partnership to support an evidence based approach to engagement with young people in London and to understand better their concerns and priorities. To this end, a survey of young people aged 11-18 years across London’s schools has been conducted, building upon previous research reported in the Youth Talk report in 20131. In addition to issues concerning safety, victimisation and perceptions of police, the current study particularly seeks to understand young people’s views on what may be deemed as intrusive police tactics, such as the use of stop and search, firearms and Taser. Furthermore, it seeks to gain an insight into the thoughts of a group of vulnerable youth not surveyed before - those attending Pupil Referral Units (PRU). A great deal of data has been obtained by the survey. In total 9,492 youths responded providing a good capture of their voices. This report provides some top level findings, and focuses mainly on groups where it is deemed there is an opportunity and/or need for engagement: the youngest (Year 7, ages 11-12); those attending PRU; and a potentially vulnerable group that report safety concerns within their family. Key Findings: Overview The presence of known police officers can make young Londoners feel safer  Overall 41% (n=3,881) of young Londoners never or rarely worry about their safety.  The police make a positive impact upon feelings of safety: 76% (n=7,166) of young Londoners feel safer seeing known police officers patrolling; 69% (n=6,513) feel safer seeing police on public transport and 67% (n=6,332) feel safer having a dedicated Safer Schools Officer.  Over a quarter (28%, n=2,657) of young Londoners worry about their safety on a daily or regular basis. This is higher for the youngest surveyed (school Year 7, 11-12 year olds).  Young Londoners feel least safe in parks 36% (n=3,418) and outside spaces, whilst 33% (n=3,092) said on public transporti. Respondents felt most unsafe about groups of people hanging around/gangs (49%, n=4,680)ii. i On public transport (36%) and in parks and public spaces (32%) were the most frequent responses given by Youth Talk respondents as to where they felt least safe. ii Groups of people hanging around /gangs (51%) was the most frequently given answer as to what worried Youth Talk respondents the last time that they felt unsafe. 3  A minority of young Londoners feel least safe at home (8%, n=774), or else have been made to feel unsafe by a family member (4%, n=347), suggesting they are vulnerable. The levels reported are greater than those documented in the previous Youth Talk (our equivalent survey conducted in 2013). This starkly demonstrates the need to engage and safeguard this vulnerable group. The majority of young Londoners are not affected by gang or safety issues in school  The top three issues to affect young Londoners in school daily or a few times a week are violence/fights (27%), bullying (24%) and people stealing (21%). In addition, to violence/fights and people stealing, drugs use/dealing is also a daily or regular experience for 19% of young Londoners in the areas where they live.  Cyber-bullying/trolling is an emerging issue for young Londoners, with 17% (n=1,644) affected daily or regularly.  The majority of young Londoners are not affected by gang issues at school (80%, n=7,580), or their home neighbourhood (73%, n=6,925).  However, an important minority (11%, n=999) feel under pressure to join a gang. This pressure is more marked for the youngest surveyed (school Year 7, 15%, n=351). Given the greater concerns over safety reported by this age group, and the onus on preventing gang crime, this emphasises the need for effective engagement.  Perceived pressure to join a gang is also more marked for those who have been victims of crime (21%, n=383) and attendees/former attendees of PRU (31%, n=129) in comparison to the overall youth sample, further highlighting the need for effective engagement and preventative education. The opinions of young Londoners suggest that this needs to be undertaken as part of a multi-agency approach, which aligns with MOPAC Policy2. The majority of young Londoners have never been victims of crime  81% (n=7,649) of young Londoners have never been victims of crime.  However, the proportion of young Londoners that have been a victim of crime (19%, n=1,843) is higher than that for the general population (11% as measured by the Public Attitude Survey, 7% as measured by the Crime Survey of England and Wales).  Under half (43%, n=323) of young Londoners who report something stolen to the police, and half (50%, n=173) of those who report an assault/threat to the police, are happy with the police response. This places a renewed focus on the police contact experience with young victims of crime as this is over 30% below that of the adult population, as recorded by the MOPAC User Satisfaction Survey. 4 Young Londoners are interested in the police and what they do  Three out of four young Londoners (73%, n=6,961) are interested in receiving information about the police and what they do. 23% (n=2,154) are interested in joining the police cadets, providing an evidence base for the Mayoral ambition to increase cadet membership.  Under half of young Londoners believe that their local police do an excellent/good job near where they go to school (42%, n=3,944) or where they live (39%, n=3,665). This is over 20% below public confidence as measured via the Public Attitude Survey. Whilst some of this difference can be accounted for by methodological differences, this echoes previous research which indicates that young Londoners are less confident in the police.  Levels of confidence decrease for young Londoners after school Year 7. Important minorities who have been victims of crime, have attended PRU and who feel unsafe with a family member, show greater levels of disagreement that the police listen to young peoples concerns, deal with issues that matter to them, are friendly and approachable or would treat them with respect, as compared to the overall survey population (see Table 2 in section 4). Effective engagement and contact experiences with these groups is essential.  61% (n=5,769) of youth respondents would be willing to meet with/talk to the police about issues which are important to them, predominantly in a school environment, placing a renewed focus upon the role of the Safer Schools Officer. Young Londoners are supportive of Stop and Search as a tactic  The use of Stop and Search is relevant to young Londoners: in comparison to other tactics they are more aware of its use, more likely to have seen or experienced it in real life and 55% (n=5,202) claim to know their rights. Encouragingly, young Londoners are largely supportive of Stop and Search as a tactic. In comparison to other tactics asked about, more young Londoners say that Stop and Search makes them feel safe, and more say that the police should use it whenever they feel necessary.  The majority of young Londoners have never seen or experienced police using firearms, CS spray or Taser in real life. However, an important minority have (firearms 18%, n=1,724; CS spray 17%, n=1,565; Taser 13%, n=1,267). In comparison to a range of tactics asked about, more young Londoners (17%, n=1,581) said that the police should never use firearms, and more felt unsafe (40%, n=3,750) than safe (38%, n=3,605) with regard to their use. 5 Introduction There is established evidence concerning the importance of researching the views of London’s young population and for police to engage positively with them. Not only are 11% of the capitals population3 aged between 10 and 19, but young people are also more likely to come into contact with the police as both victim and offender4. When they do come into contact, it is important for the Police and Criminal Justice System to handle that contact well for a number of reasons. There is evidence that youth reoffending can be reduced with good quality resettlement5, confidence in the police and support of tactics such as stop and search is lower amongst young people6. A recent Mayors Office Strategic Ambitions paper7 also states that work around gangs and serious youth violence must “be shaped by and with young people”. Given that gangs and the police use of stop and search are major media and political issues, this young voice is needed to engage effectively and supportively. Gathering the opinions of London’s youth is therefore vital. The Public Attitude Survey (PAS) is only completed by those aged 16 and over, so a survey delivered to youth aged 11-18 is capable of providing valuable insight into what this vital age group (comprising 11% of the capitals population8) feel about their safety, their experience of crime, and their perceptions of the police and the tactics they use. The views of marginalised or vulnerable youth are particularly valuable, and in an addition to the previous ‘Youth Talk’ survey in 2013, this survey was also distributed to Pupil Referral Units (PRU). In total there were 9,492 respondents to the survey. There was higher proportion of younger respondents - 68% were in school years 7-9 (aged 11-14 years), and only 8% from school years 12-13 (aged 16-18). The male/female split was 54%/46%. See Appendix A for survey methodology. There was not an even distribution of where respondents go to school – please refer to the map in Appendix A. 6 Findings The results of the survey are broken down and discussed under the following main themes: Safety; Issues that Affect Young Londoners; Victimisation, Satisfaction and Under-reporting; Young Londoners Perceptions of the police; and a case study about vulnerable young Londoners who either attend/have attended PRU, or have been made to feel unsafe by a family member. 1. Safety It is a positive finding that a large proportion of young Londoners (41%, n=3,881) never or rarely worry about their safetyiii, although an important minority do. Over a quarter (28%, n=2,657) worry on a daily (11%, n=1,050) or regular basisiv (17%, n=1,607). The Police can make an impact upon safety Sizeable majorities of young Londoners said that they would feel more safe to see police officers that they knew patrolling (76%, n=7,166), police on public transport (69%, n=6,513) and having a dedicated safer schools police officer (67%, n=6,332). This last point is particularly encouraging - in the Youth Talk 2013 report, a lower proportion (58%) said such an officer would make them feel safer. There are currently 266 safer schools officers throughout the MPS, an increase from 176 in financial year 2013-2014. As discussed later, in the section on Perceptions of Police, there is evidence for a renewed impetus on the role of safer schools officers. However, only 40% (n=3,823) said that they would feel safer to see police officers that they did not know patrolling. Furthermore, 32% (n=3,053) said this would make them feel less safe (only 8% said this of officers they did know patrolling). There is some additional evidence from the survey to suggest that some enforcement activity can make the young Londoners feel unsafe: 50% or more of them feel less safe by the possibility of knife arches/wands at school, or more people being arrested in their neighbourhood. This stresses the importance of a local context to policing and the ability to develop relationships with officers. Important minorities of young Londoners feel unsafe Two key findings are: the youngest cohort (Year 7, 11-12 year olds) worry more than those in subsequent school years (see Appendix B for data table). Secondly, those who attend/have attended PRU are more polarised in their opinions, compared to those who have not attended PRU. For example, 35% (n=805) of respondents in school year 7 report worrying about safety on a daily or regular basis. Of those who attend/have attended a PRU, 15% (n=63) worry about their safety on a daily basis, yet a significant 26% (n=107) never worry. Those who have been a victim of crime worry about safety on a daily basis slightly more often (13%, n=247). iii Rarely (a few times a year) iv Regularly (a few times a week) 7 An important finding from the 2013 Youth Talk survey was that a small number (2%, n=279) of respondents felt least safe at home. A much larger proportion of respondents (8%, n=774) reported the same concern this year. Furthermore, 4% (n=347) of respondents were made to feel least safe by a family memberv, and 3% (n=256) said that they would be made to feel less safe by having family they could talk to. Clearly this is a serious issue, and there is great importance to engage with these youth for their own safety. One of the recommendations from the 2013 Youth Talk was that the role of schools in the wider safeguarding needs to be considered and our findings this year would continue to support this. People and places influence Young Londoners feelings of safety In terms of where the youth feel least safe, and what makes them feel least safe bears similarity to the findings reported in Youth Talk: 36% (n=3,418) said in parks and outside spaces , and 33% (n=3,092) said on public transportvi. Regarding this second point, there are opportunities to engage with Transport for London and British Transport Police to further develop understanding how safe commuters feel, and ways of addressing the issue. Despite this, it is worth noting that only 9% (n=880) reported feeling unsafe on their journey to and from school. Respondents felt most unsafe about groups of people hanging around/gangs (49%, n=4,680)vii. With much media attention and police focus around gang crime, this is an important issue to note - however, equally important is to note some degree of ambiguity. A ‘group of people/gangs’ does not necessarily mean the same as ‘gangs’ in the sense as defined by the police. Whilst acknowledging the importance of the issue for the youth, caution should therefore be used in this interpretation. v Q5: The last time you felt unsafe in general, who or what was it that worried you? vi On public transport (36%) and in parks and public spaces (32%) were the most frequent responses given by Youth Talk respondents as to where they felt least safe vii Groups of people hanging around / gangs (51%) was the most frequently given answer as to what worried Youth Talk respondents the last time that they felt unsafe. 8 2. Issues that affect young Londoners and their priorities Most young Londoners are not affected by gangs The majority of young Londoners say that they are never or rarely affected by gang- related activity at school (80%, n=7,580) and in their home neighbourhood (73%, n=6,925). What is particularly encouraging is that the proportion saying that they are never affected is higher than reported in the 2013 Youth Talk survey.viii An important minority do feel under pressure to join gangs There are 11% (n=999) of young Londoners who feel under pressure to join a gang. The proportion of those attending/previously attending PRU (31%, n=129) and those who have been victims of crime (21%, n=383) who feel such pressure is higher still, perhaps indicative of a relationship between those with complex social needs, victimisation and gang involvement. Interestingly a larger proportion of Year 7’s, 15% (n=351) feel pressure to join gangs than any other yearix. This fully supports the prevention approach outlined in the MOPAC gang strategy9. According to the respondents to the 2013 Youth Talk survey, the people who should be most responsible for preventing young people joining a gang were the policex. In the current survey, the question was asked differently: who did the respondents feel can help if they feel under pressure to join a gang, and choose as many from a given list as applicable. Interestingly, while 44% (n=4,192) said the police, this was the fifth most popular choice, falling some way behind four others: the community (76%, n=7,179) and then young people/peers (68%), youth groups/social workers (64%) and teachers/school counsellors (63%). It may be argued that the youth do not trust the police or feel confident that they can help. Or it could be that the youth themselves recognise the gang issue as a complex social one, not just a police one. Either way the results should be used as evidence to support multi-agency gang interventions and prevention strategies which provides further evidence to support MOPAC’s serious youth violence strategic approach. Other issues at school and near home Regarding drug use, 19% (n=1,263) of young Londoners say they are affected on a daily or regular basis in the neighbourhood where they live, while a lower 14% (n=663) say that this about school, which shows similarities to previous research conducted in 2013. Nearly a quarter are affected by bullying on a daily or regular basis at school (24%) and in their neighbourhood (14%), while 17% (n=1,644) are affected daily or regularly at viii 61%, n=5,813 and 54%, n=5,092 (2014) said that gang activity never affected them at school or in their home neighbourhood respectively, compared to 57%, n=6,638 and 47%, n=5408 (2012) ix Please see Appendix B, Table 1 for comparison between school year groups x 36%, n=4,221 said selected the police option in answer to ‘who do you think should be responsible for preventing young people from joining gangs (please tick one)’ from a list of six options 9 school by cyber-bullying/trolling. This second issue shows the emergence of cyber crime/trolling; an area requiring a greater evidence base. 3. Victimisation, Satisfaction and Under Reporting The majority of Young Londoners have never been victims of crime The majority (81%, n=7,649) of young Londoners have not been a victim of crime. However, compared to the general population, young Londoners are more likely to come into contact with the police as a victim of crime (19%, n=1,843 compared to 11%, as reported by the PAS, or 7% as reported by the Crime Survey for England and Wales). This is consistent with other evidence.10 The proportion of victims increases with age (see Appendix B, Table 1, for data table). An important minority of Young Londoners are even more likely to be victims There is a relationship between vulnerable young people and victimisation. Once again, those who attend/have attended a PRU stand-out, more than a third, 39% (n=162) have been victims of crime; while the percentage of young people made to feel least safe by a family member was also higher than the overall sample (31%, n=107). This further highlights that vulnerable groups warrant attention as the police are more likely to come into contact with them. How young Londoners perceive crime may be different to how police do so In terms of specific crimes, 34% (n=3,237) say they have had something valuable stolen from them, and 24% (n=2,311) have been victims of physical injury or threat. Interestingly, both of these figures are higher than the proportion saying that they have been victims of crime. This may be that they do not always consider the incidents an actual crime, and that they are including trivial incidents in response to the question. An examination of free text responses providedxi indicate that this is sometimes the case, revealing instances of friends, classmates or siblings taking low value things, which they may get back, or of fights between them which are not very serious. There is, perhaps some insight here into what the youth actually perceive as being ‘crime’, which in turn may have some consequence regarding both their own behaviour and how they interact with police. This should be borne in mind the police communicate with young people over crime prevention issues. Young Londoners are often repeat victims, but do not often report crime to the police Young Londoners experience repeat victimisation: 43% (n=1,387) of theft victims and 56% (1,276) of assault/threat victims say they have been so more than once. Comparison with Crime Survey for England and Wales figures suggests that this is far higher than the degree of repeat victimisation seen in the adult population.xii11 Mostly, xi In response to the question ‘You say you didn’t report this to the police, was this because…’ and then having selected the reason ‘other’ xii For the period April 2013-March 2014, CSEW figures presented by the Office for National Statistics shows that repeat victimisation for theft from person, other theft, other household theft, bicycle theft, 10 they do tell someone, however only a quarter (23%, n=758) of theft victims and 15% (n=348) of assault/threat victims tell police. The main reasons cited as not reporting a crime to the police were that they did not think it was important or did not think that the police would help. Of those who did report theft to the police, less than half (43% n=323) were very happy or happy with the way police treated them; while 50% (n=173) of assault/threat victims were either very happy or happy. This second figure is lower than the general population as reported in the User Satisfaction Survey (USS), which shows 76% of violent crime victims being completely, very or fairly satisfied with police12 (the USS provides no comparable figure for theft victims). The most frequently cited things that could have improved the opinions of young Londoners was for the matter to be taken seriously, and for someone to be arrested and charged, which aligns with wider victim research13. 4. Young Londoners perceptions of police There is a difference between young Londoners confidence in police and that of London as a whole There is a 26% confidence gap between young London and London as a whole as measured via the PAS (42% vs 68%). Whilst some of this difference can be accounted for by methodological differences, this echoes previous research which indicates that young people are less confident in the police. This research reveals that 42% of young people are confident in the policing around their school, and 39% in their home neighbourhood (see Appendix B, data table 2). There is some overlap between the results of this survey and the PAS in terms of how the boroughs rank for confidence. There are four boroughs that rank below average for confidence in both surveys: Croydon, Lewisham, Newham and Tower Hamlets. Important minorities of young Londoners are less confident Subsequent to Year 7 (11-12 year olds), levels of confidence in the police declines. Young people who have been victims of crime, have experienced pressure to join a gang, and seen or experienced Taser being used also have less confidence in the police (see Appendix B, data table 2). Previous research into public confidence shows that despite these individual differences, confidence in the police can be improved by focussing on four key elements: perceptions of police effectiveness, fairness of personal treatment, the level of police engagement with the community, and local people’s concerns about local disorder. Taken together these four elements indicate that public confidence can be influenced by ‘what police do’14 . domestic burglary and robbery ranged from 0.7% (for robbery) to 8.7% (for domestic burglary); while repeat victimisation for violence with injury was 4.8% and without injury 2.1%. 11 Good quality engagement improves confidence in policing Previous research has shown that good quality engagement is critical for success, and delivering improvements in public confidence. Assessments of community engagement are made from a combination of measures asking respondents if they agree that the local police listen to their concerns, are dealing with them, are friendly and approachable and would treat them with respect if they came into contact with them for any reason. Over half of all respondents agreed that the police listen to concerns, deal with issues that matter to them, are friendly and approachable and would treat them with respect. The youngest Londoners surveyed (Year 7) hold the best opinions of each age group, as shown in table 1. Table 1: confidence in the police by school year As shown in table 2, a greater proportion of current/previous PRU attendees, victims of crime and those who have felt unsafe at home disagree with each of the listed statements, reinforcing again that those coming into contact with the police are more likely to have a lower opinion. However, previous research by Evidence and Insight indicates that not all contact has to break confidence; positive contact with the police can improve confidence15. Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Listen to the concerns of young people Strongly Agree/Agree 1667 72% 1254 59% 1074 52% 602 45% 290 39% 217 41% 147 40% Deal with issues that matter to young people Strongly Agree/Agree 1678 72% 1299 62% 1125 55% 642 48% 314 42% 214 41% 155 42% Are friendly and approachable Strongly Agree/Agree 1748 75% 1404 67% 1207 59% 737 55% 386 51% 274 52% 196 53% Would treat you with respect if you came into contact with them for any reason Strongly Agree/Agree 1800 78% 1476 70% 1336 65% 815 60% 444 59% 315 60% 214 58% Year 12 (age 16-17) Age:Year 13 (age17-18) Table 1. Confidence in the Police by School Year To what extent do you agree that the Police.. Age Year 7 (age 11-12) Year 8 (age 12-13) Year 9 (age13-14) Year 10 (age14-15) Year 11 (age 15-16) 12 Table 2: minority groups and confidence in the police Young Londoners prefer information about local issues, and are keen to talk to police about issues that matter to them Good quality information provision is essential to improving confidence in policing16. The results from this year’s cohort align with previous studies, young people are mostly interested in what is relevant to them locally, and to have it delivered without them having to seek engagement themselves. However, many more young Londoners are interested in receiving information from the police. Three out of four young people would like to receive Information from the police about what they do (73%, n=6,961). Information is mainly received via the television (53%, n=5,030), and that is how most young people would like to receive it (40%, n=3,807), however 28% would like to receive a newsletter and 23% would like to receive information via a safer schools officer. The type of information that they are most interested in receiving concerns local crime issues (58%, n=5,526), local arrests (44%) and crime prevention advice (35%). Only a quarter of the respondents 27% (n=2,531) say that they do not wish to receive information about the police. It is important that this information is presented in a way which does not heighten their fears about their own safety issues. Information should Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Strongly Agree/Agree 5251 55% 224 54% 855 46% 189 55% Strongly Disagree/Disagree 1656 18% 121 29% 480 26% 82 24% Strongly Agree/Agree 5427 57% 219 53% 868 47% 183 53% Strongly Disagree/Disagree 1658 18% 119 29% 491 27% 81 23% Strongly Agree/Agree 5952 63% 235 56% 1006 55% 199 57% Strongly Disagree/Disagree 1395 15% 116 28% 410 22% 79 23% Strongly Agree/Agree 6400 68% 232 56% 1118 61% 210 61% Strongly Disagree/Disagree 1170 12% 119 29% 368 20% 69 20% Would treat you with respect if you came into contact with them for any reason Overall Sample To what extent do you agree that the Police.. Listen to the concerns of young people Deal with issues that matter to young people Are friendly and approachable PRU attendees (current or previous) Victim of crime Made to feel unsafe by family member Table 2. Minority Groups and Confidence in the Police 13 be locally relevant but must be presented as good news stories or as examples of police community engagement. It should not be presented as crime figures alone17. Regarding their own concerns, 61% (n=5,769) said that they would meet/talk to the police to help them understand issues important to young people. Particularly encouraging is that this proportion is similar even for those that attend/have attended PRU (60%, n=249), and for those who have felt unsafe in their own home (59%, n=206). When asked where/how they would most like to do this, the largest proportion (45%, n=4,243) said at school. This appears to be a good place to engage the youth, as there is a low interest in engaging with police at a variety of the community events or groupsxiii. Here there is renewed impetus on the role Safer Schools Officers, who can play a key role in positive engagement. As an additional note, despite a lack of interest shown in community events and groups, there is interest in joining Police Cadets for 23%, (n=2,154 respondents). This provides an evidence base for the Mayoral ambition to increase cadet membership18. Young Londoners are aware of police intrusive tactics but some tactics can make them feel unsafe The young Londoners were asked about a range of police tactics, each of which could, in some way, be regarded as intrusive or forcefulxiv. The questions sought to see how aware young people were of each tactic, whether they had seen or experienced them in real life, circumstances under which they felt their use may be justified, and whether the police using each tactic made them feel safe or unsafe. Please see charts 5, 6 and 7 in Appendix B for comparative responses. An important minority of young Londoners have seen or experienced Firearms, Taser or CS Spray in real life It is not surprising that young people think that the police should only use firearms, Taser and CS Spray in violent situations, or only when someone’s life is in danger. What is perhaps surprising is the proportion of respondents who claim to have seen or experienced police using firearms, Taser and CS Spray in real life (18%, n=1,724; 17%, n=1,565; and 13%, n=1,267 respectively). A larger proportion of young Londoners objected to police use of firearms than any other tactic asked about (17%, n=1,581 saying they should never be used). More of them also reported feeling unsafe knowing the police could use firearms (40%, n=3,750) than safe (38%, n=3,605). This is the only tactic featured on the survey where more respondents felt unsafe than safe. xiii The survey asked if the youth belonged to any of the following: Street elite, Urban Slam, Police Cadets, Scouts , Guides or other Uniformed Group, London/Premier Kicks, Hitz, Street Chance, Met Row, Met Track, Youth Council, Youth Panel, or None of these, and then asked if they were interested in joining any. xiv These tactics included the use of police horses, batons, armoured vehicles, crowd control, CS Spray, Tasers, Firearms and Stop and Search 14 Stop and Search matters to young Londoners The use of stop and search in relation to young Londoners is an important issue. Between 2009 and 2013, 25% of all Stops and Searches carried out by the MPS were of under 18 year-olds19, and between 2011 and 2013. This suggests that the findings above are not surprising; and while they are not necessarily negative, they are important. Once again they provide evidence of how likely young Londoner’s are to come into contact with the police, and therefore how well the police need to focus on making that contact as positive as possible. Previous research indicates that the most important aspect for young people during a stop and search is that police have a good reason to use the tactic and treat people politely during the procedure20. It is important for the police to consider this and wider safeguarding issues, when carrying out stop and search procedures on young Londoners. The survey results confirm the importance of the issue to young Londoners. There is a higher awareness of Stop and Search than for any other tactic (77%, n=7,293 were very or fairly aware). Furthermore, 57% (n=5,384) say that they have seen or experienced Stop and Search being used in real life (this is a higher proportion of respondents than for any other tactic, other than police use of horses). In addition to this, 55%, (n=5,202) claim to know their rights if an officer stopped and searched them. Interestingly, the largest proportion of any age group claiming to know these rights were the youngest (60% of Year 7’s and 59% of Year 8’s). In comparison to the overall sample, a particularly high proportion of current or former PRU attendees (65%) also claimed to know their rights. Encouragingly, the use of Stop and Search appears to be reasonably well supported amongst the respondents. Only 6% (n=583) feel that Stop and Search should never be used, which is the lowest proportion for all tactics asked about. 58% (n=5,457) said that the use of Stop and Search made them feel safe, which is higher than for any other tactic, and while 17% (n=1,630) said that its use made them feel less safe, this is a lower proportion than for any other tactic (with the exception of police using horses, 17%). Also, 40% (n=3,760) said that they felt Stop and Search should be used whenever the officer thinks that it is necessary, which is a greater proportion saying this than for any other tactic (again, with the exception of police using horses). This echoes previous research which found that 56% thought that stop and search should be allowed21. It appears to be the 16-24 cohort who are less supportive of this tactic22. 15 5. Case Studies: Vulnerable Young Londoners The results of the survey have drawn attention to two particular groups of young people, who may be deemed as being vulnerable. The first of these is those who currently attend, or have previously attended Pupil Referral Units (PRU). The second are those identified under the section on Safety - youth who say that the last time they felt unsafe in general; it was a family member that worried them. This section provides additional information on these groups. PRU attendees are vulnerable youth Pupil Referral Units (PRU) are Alternative Provision establishments, providing education for young people excluded or unable to attend mainstream school for a variety of reasons. These reasons can include pupils expelled or suspended from school, pregnant school girls or school-age mothers, asylum seekers and refugees who have no school place, pupils with behavioural or anxiety problems, or vulnerable pupils, or those unable to attend school for medical reasons23. As of March 2013 there were 393 in England24 and 48 in London. Not only are the young people who attend these establishments likely to be vulnerable and live in disruptive circumstances, there is evidence that young people excluded from school are more likely be involved in crime, with sources citing that 60% of young people excluded from school nationally having offended in the previous 12 months25, or that for 15-17 year olds, 88% of males and 74% of females in Youth Offenders Institutes have been excluded from school at some point26. Understanding the views of PRU attendees and developing positive engagement with them is therefore vitalxv. A combined 4% of respondents to the survey are either currently attending a PRU (2%, n=225) or have previously attended a PRU and are now back in mainstream school (2%, n=193). For this section, the views of this combined group of 418 youth shall be discussed, and referred to as the PRU group. A total of 62% (n=258) of the PRU cohort are from the younger years, Years 7-9, ages 11- 14. There is a more notable proportion of boys than with the overall sample (63%, n=264), while there are a very similar number of White British and Black youth (22%, n=92 and 21%, n=89 respectively). A larger proportion of the PRU cohort feel worried by a family member the last time they felt unsafe (10%, n=43), and felt less safe by a family member that they could talk xv MOPAC has recently carried out a survey of Youth Offending Team caseload managers in the London Resettlement Consortia, regarding the emotional and mental health needs in regard to traumatic experiences of young Londoners on their caseloads. MOPAC have further commissioned research by the University of Middlesex regarding assessment and identification of support needs amongst victims in from the YOT caseload cohort. Although many young Londoners are not young offenders, this work shows MOPACs drive to understand the complex relationship between needs, offending and victimisation 16 to (9%, n=38). Although these are very small numbers, it provides some indication of the lack of personal support young people who attend PRU can encounter, and an insight into the kind of safety issues that can affect them. Encouragingly, nearly three quarters of the group (74%, n=308) felt either very safe or fairly safe in the PRU, 26% did not, with 15% (n=61) saying that they did not feel very safe at all. The survey also asked the PRU cohort questions about how safe, supported and happy they felt about the move back into mainstream school. The majority of PRU attendees felt positive, ranging from 72% (n=299) being happy about the move to 79% (n=328) being confident. 68% (n=283) said that they had been offered support from either a PRU resettlement officer, Social worker or Youth Offending Team Police Officer. Young Londoners at risk of harm There is a minority of young Londoners who are at most risk within the environment in which they live. A total of 8% (n=774) of the respondents reported feeling least safe at home; 3% (n=256) said that speaking to a family member would make them feel less safe; 6% (n=1,487) said that they are affected by violence at home on a daily or regular basis; and 11% (n=1,007) select domestic violence/violence within the home as one of the three things they most think the police should focus on to help them feel safe in London. This section shall focus upon the 347 young people (4% of respondents) who say that the last time they felt unsafe in general it was a family member that worried them. Over three quarters of this cohort (77%, n=269) are aged 15 or younger, with a large proportion of them (46%, n=161) being amongst the youngest surveyed (Years 7 and 8, 11-13 year olds). The male/female split is very similar to that of the overall population responding to the survey (53% male, n=185); while three ethnic groups - Asian, White British and Black, each represent near 20% each. 30% (n=105) of this cohort worry about safety on a daily or regular basis, and a further 27% (n=92) worry a few times a month and the issue that most worries them is a family member. A quarter (18%, n=63) say that violence at home affects them on a daily or regular basis, and 14% (n=48) select domestic violence among their top three priorities for the police to deal with - whereas 60% (n=208), report they never experience this type of violence. This could indicate that violence (while very serious and important to draw attention to) is not the only factor that makes them feel afraid of a family member. Clearly a provision for support and engagement with this cohort is necessary for their own safety. It is possible that these young people, having reported their concern in response to an anonymous survey, may be more reluctant to do so without other supportive encouragement and engagement. The role that schools and the police play in a multi-agency wider safeguarding approach is critical to reaching and supporting young Londoners. 17 6. Conclusion Many Young Londoners surveyed feel safe, and believe that the police - particularly known police - make them feel safer. Issues such as gang crime do not worry the majority, and over 80% have never been victims of crime. Over half believe that the police listen to their concerns, deal with issues that matter to them, are friendly and approachable and would treat them with respect. They also demonstrate a desire for engagement with the police regarding issues important to them, preferably via a school environment. However, under half of young Londoners believe their local police do an excellent/good job - 20% below that of the Adult Public Attitude Survey sample. Plus, it must be noted that a quarter worry about their safety daily or regularly, and for an important minority there are both concerns about gangs and pressures to become involved in them. These concerns and pressures are more marked for the youngest cohort. Although the youngest (Year 7, 11-12 year olds), in general, also have the best opinions of the police, opinions which deteriorate as they reach late teens, so there is opportunity for early engagement, to be more successful, with this cohort. Equally young people aged 17-18 who have less positive views desire engagement and better engagement strategies are needed to reach this group. Vulnerable young people, such as victims of crime, those who attend/have attended Pupil Referral Units, and those who have safety concerns over a family member, are more likely to come into contact with the police, and have less favourable opinions of the police - hence positive engagement is also very important with these groups. Police can make positive impacts upon feelings of safety, through known officers on patrol and safer schools officers. Although there are concerns amongst the young people over intrusive police tactics such as the use of firearms, stop and search - more widely experienced than other tactics - is generally well supported. 18 Appendix A: Methodology All data has been obtained from responses to an online survey rolled out to youth across London. The questionnaire was developed by MOPAC’s Evidence and Insight team in collaboration with both MOPAC policy staff and Metropolitan Police Service (MPS) engagement officers. The questionnaire built upon a previous version used for the Youth Talk report conducted in 2013. The survey was promoted in London schools by MPS Safer Schools Officers and letters were sent to all schools in boroughs where response rates were low. MOPAC also used social media to promote the survey. The survey was carried out online, either during class time or as homework. School years 7-13 were surveyed (11-18 year olds). It consisted of a total of 52 questions, some of which were rooted in others, and therefore only completed where relevant. There were 9,492 respondents to the questionnaire from youth at school across the 33 London boroughs (including the City of London). This number is 1% of the population of London aged 10-1927. While this is larger than the representation garnered by the Public Attitude Survey, it must be noted that the survey is not as robust as the PAS. There is not an equal representation of youth attending school on all boroughs, and from three boroughs there are less than 10 respondents in each case: Westminster (three), Islington (four) and Kensington and Chelsea (five). This must be kept in mind when assessing how robust the findings are. The distribution of respondents by borough is shown in Appendix A: Map below. There is also higher representation of younger youth. Youth in school years seven, eight and nine accounted for over 20% of respondents each; school years 11 to 13 were represented by under 10% of respondents each (8%, 5% and 4%). 52 youth (0.5%) responded to the question “What type of school do you attend?” by choosing Pupil Referral unit; however, in response to the question ‘Have you ever attended a Pupil Referral Unit?’, 224 (2%) of them said “yes - currently”, and a further 193 said “yes - have attended... but now back in mainstream school”. Throughout this paper, wherever opinions or feelings of PRU respondents are reported on, it is this group of 418 youth that are be referred to. 19 Map 1: Where online respondents go to school Note: There were also 91 respondents who went to school in the City of London (second from bottom range), and 76 respondents who did not know what borough they went to school on. An additional 226 paper responses were received, some of these were not usable as minimal questions had been completed, and some had been returned after the closing date and were therefore not included within the analysis presented in this report. However, analysis of the 139 responses which were useable aligns with the main findings presented here. The majority of respondents who returned the paper questionnaires were from schools in Islington (64%) and Waltham Forest (17%). 20 Appendix B. Tables Table 1: safety and experience of crime Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Never/Rarely 3884 41% 761 33% 810 38% 912 44% 615 46% 366 47% 245 49% 181 49% 198 47% 718 39% 150 43% Regularly/Daily 2657 28% 805 35% 639 30% 528 26% 311 23% 167 22% 126 24% 81 22% 120 29% 566 31% 105 30% Yes 999 11% 351 15% 214 10% 166 8% 121 9% 79 11% 32 6% 36 10% 129 31% 383 21% 73 21% No 8493 91% 1968 85% 1896 90% 1896 92% 1232 91% 673 90% 492 94% 336 90% 289 69% 1460 79% 274 79% Yes 1843 19% 290 13% 324 15% 417 20% 332 25% 197 26% 156 30% 127 34% 162 39% 107 31% No 7649 80% 2029 86% 1786 84% 1645 80% 1021 76% 555 74% 245 66% 7649 81% 256 61% 240 69% Yes 3237 34% 716 31% 700 33% 755 37% 483 36% 244 32% 190 36% 149 40% 217 52% 1172 64% No 6255 66% 1603 69% 1410 67% 1307 63% 870 64% 508 68% 334 64% 223 60% 201 48% 671 36% Have you ever felt under pressure to join a gang? Have you ever been a victim of crime? Have you ever had something valuable stolen? Overall Sample PRU attendees (current or previous) Victim of crime Made to feel unsafe by family member Age How often do you worry about safety? Table 1. Safety and experience of crime Year 7 Age:Year 13 Year 8 Year 9 Year 10 Year 11 Year 12 21 Table 2: confidence in the police Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Listen to the concerns of young people Strongly Agree/Agree 5251 55% 1667 72% 1254 59% 1074 52% 602 45% 290 39% 217 41% 147 40% 224 54% 855 46% 189 55% 899 57% Deal with issues that matter to young people Strongly Agree/Agree 5427 57% 1678 72% 1299 62% 1125 55% 642 48% 314 42% 214 41% 155 42% 219 53% 868 47% 183 53% 912 58% Are friendly and approachable Strongly Agree/Agree 5952 63% 1748 75% 1404 67% 1207 59% 737 55% 386 51% 274 52% 196 53% 235 56% 1006 55% 199 57% 981 63% Would treat you with respect if you came into contact with them for any reason Strongly Agree/Agree 6400 68% 1800 78% 1476 70% 1336 65% 815 60% 444 59% 315 60% 214 58% 232 56% 1118 61% 210 61% 1019 65% How good a job do you think that the police are doing in the area around your school? Excellent/Good 3944 42% 1326 57% 899 43% 758 37% 413 31% 238 32% 171 33% 139 37% 153 37% 559 30% 133 38% 587 38% How good a job do you think that the police are doing in your home neighbourhood? Excellent/Good 3665 39% 1171 51% 846 40% 717 35% 408 30% 249 33% 158 30% 116 31% 142 34% 509 28% 103 30% 555 35% Age:Year 13 Year 9 Year 10 Table 2. Confidence in the Police Year 11 Year 12 To what extent do you agree that the Police.. Overall Sample Age PRU attendees (current or previous) Victim of crime Made to feel unsafe by family member Year 7 Year 8 Seen Taser used in real life 22 Table 3: police tactics Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Base (number) Percentage Safe 3887 41% 997 43% 880 42% 847 41% 500 37% 276 37% 228 44% 159 43% 163 39% 775 42% Less Safe 2865 30% 798 34% 683 32% 599 29% 399 30% 181 24% 120 23% 85 23% 142 34% 544 30% Safe 4298 45% 1117 48% 972 46% 935 45% 560 41% 312 42% 240 46% 162 44% 178 43% 852 46% Less Safe 2979 31% 808 35% 706 34% 620 30% 406 30% 200 27% 142 27% 97 26% 133 32% 556 30% Safe 3605 38% 954 41% 813 39% 743 36% 491 36% 260 34% 200 38% 144 39% 160 38% 711 39% Less Safe 3750 40% 923 40% 876 42% 826 40% 506 37% 293 39% 196 37% 130 35% 151 36% 758 41% Safe 5457 58% 1462 63% 1304 62% 1146 56% 709 52% 366 49% 276 53% 194 52% 200 48% 997 54% Less Safe 1630 17% 404 17% 343 16% 365 18% 244 18% 126 17% 86 16% 62 17% 110 26% 353 19% Yes 5202 55% 1388 60% 1241 59% 1112 54% 681 50% 359 48% 238 45% 183 49% 271 65% 900 54% No 4290 45% 931 40% 869 41% 950 46% 672 50% 393 52% 286 55% 189 51% 147 35% 853 46% PRU attendees (current or previous) Victim of crime Year 7 Year 8 Year 9 Year 10 Year 11 Year 12 Age:Year 13 Overall Sample Age Stop and Search Does knowing that the police can use the following tactics make you feel… If you were stopped and searched by a police officer, whould you know your rights? Table 3. Police tactics CS Srpay Taser Firearms 23 Table 4: police engagement Base (number) Percentage Base (number) Percentage Base (number) Percentage Yes 5769 61% 249 60% 206 59% No 3723 39% 169 40% 141 41% Where would be the best place to do so? School 4243 45% 266 64% 148 43% Made to feel unsafe by family member Table 4. Police engagement If you were given the chance, would you talk to/meet with the police to help them understand what issues are important to young people? Overall Sample PRU attendees (current or previous) 24 Table 5: awareness and experience of police tactics Table 6: use of police tactics and feelings of safety Use of police tactics and feelings of safety 0 10 20 30 40 50 60 70 Stop & Search Armoured Vehicles Batons/Shields Horses Taser Crowd Control CS Srpay Firearms More Safe Less Safe Awareness and Experience of Police Tactics 0 10 20 30 40 50 60 70 80 90 Stop & Search Horses Armoured Vehicles Batons/Shields Taser Crowd Control Firearms CS Srpay Aware Seen/Experienced % % 25 Table 7: when should police tactics be used When should police tactics be used 0 5 10 15 20 25 30 35 40 45 Horses Stop & Search Armoured Vehicles CS Srpay Taser Firearms Batons/Shields Crowd Control Never Whenever necessary The survey asked in what situations each tactic should be used, giving a range of answers: never, violent situations only, crowd control situations only, only when someone’s life is in danger, when police believe people might be involved in crime and whenever officer thinks necessary. The chart shows responses for ‘never’ and ‘whenever officer thinks necessary only’ for contrast % 26 1 Youth Talk: the Voice of Young London, MPS Corporate Development, August 2013 2 Strategic Ambitions for London: Gangs and Serious Youth Violence, London Crime Reduction Board, June 2014 3 http://www.nomisweb.co.uk/census/2011/KS102EW/view/2013265927?cols=measures 4 Fenn, L. (2011)’Time to Talk: Understanding the experiences of victims and witnesses of shootings in London’, MOPAC - Evidence and Insight 5 Renshaw J (2007) ‘The costs and benefits of effective resettlement of young offenders’ Journal of Children’s Services 2(4) pp.18-29 6 MOPAC Public Attitude Survey, r12 month MPS data 7 Strategic Ambitions for London: Gangs and Serious Youth Violence , London Crime Reduction Board, June 2014 8 http://www.nomisweb.co.uk/census/2011/KS102EW/view/2013265927?cols=measures 9 Strategic Ambitions for London: Gangs and Serious Youth Violence, London Crime Reduction Board, June 2014 10 Evidence and Insight, Citizen Landscape 2012 11 http://www.ons.gov.uk/ons/publications/re-reference-tables.html?edition=tcm%3A77-328153; release: Crime in England and Wales, period ending March 2014; 01. Bulletin Tables - Crime in England and Wales, year ending 2014 12 MOPAC User Satisfaction Survey, FY 2014/15 Q4 13 MOPAC User Satisfaction Survey 14 Stanko and Bradford, 2009 Beyond Measuring ‘How Good a Job’ Police Are Doing: The MPS Model of Confidence in Policing 15 Ibid 16 Wunsch and Hohl, 2008 Information Provision to the Public: The impact of newsletters 17 Fenn, L (2015) ‘A ‘nudge’ in the right direction; applying behavioural insights to police-youth engagement’ 18 MOPAC Policing and Crime Plan 2013-17 19 All Party Parliamentary Group for Children inquiry into ‘Children and the Police’, Initial analysis of information request to police forces, July 2014 20 Morrell et al, 2011 ref: Assessing the impact of police-initiated stop powers on individuals and communities, The UK picture 21 Youth Talk: the Voice of Young London, MPS Corporate Development, August 2013 22 MOPAC Public Attitude Survey, r12 month MPS data 23 Pupil Referral Units Briefing Paper, Reference 090208, September 2009 24 https://www.gov.uk/government/policies/improving-education-for-pupils-outside-mainstream- school/supporting-pages/alternative-provision 25 Back on track London: a PRU workbook, London Councils, November 2011, statistic taken from Youth Justice Boards annual youth crime survey, MORI, 2004 26 Transforming Youth Custody: Putting education at the heart of detention, Ministry of Justice, February 2013 27 http://www.nomisweb.co.uk/census/2011/KS102EW/view/2013265927?cols=measures 1 Youth Voice Survey 2018 December 2018 MOPAC Evidence and Insight Nicole Ramshaw, Barry Charleton & Paul Dawson 2 Contents Executive Summary ......................................................................................................... 3 1. Introduction ................................................................................................................ 4 1.1. Background to the Youth Voice Survey 2018 ................................................................... 4 2. Methodology ............................................................................................................... 6 2.1. A London Online Survey ................................................................................................... 6 2.2. Ethical Considerations ...................................................................................................... 6 2.3. Limitations ........................................................................................................................ 7 3. Results ........................................................................................................................ 8 3.1. Young people and crime ....................................................................................... 8 3.1.1. Safety ............................................................................................................... 8 3.1.2. Crime Victimisation ......................................................................................... 10 3.2. Young people and the police .............................................................................. 12 3.2.1. Perceptions of the police ............................................................................. 12 3.2.2. Stop and Search ........................................................................................... 14 3.2.3. Safer Schools Officers .................................................................................. 15 3.3. Young people and serious violence .................................................................... 17 3.3.1. Knife and gang crime ................................................................................... 17 3.3.2. London Needs You Alive .............................................................................. 19 3.3.3. Knife imagery in media campaigns ............................................................... 20 3.4. Young people and Safeguarding ........................................................................ 22 3.4.1. Online safety ................................................................................................ 22 3.4.2. Sexual harassment and healthy relationships ................................................ 24 4. Conclusion ................................................................................................................. 27 References .................................................................................................................... 30 Appendices .................................................................................................................... 32 3 4 1. Introduction 1.1 Background to the Youth Voice Survey 2018 According to Office for National Statistics (ONS) population estimates, over half a million London residents are aged between 11 and 16; representing nearly 7% of London’s total population. More widely, the number of young people aged under 16 living in London is projected to rise by 7% between mid-2016 and mid-2026 (ONS, 2018), emphasising the importance of giving this group the opportunity to voice their opinions about crime and safety issues that affect them. Research has highlighted a potentially complex picture of crime victimisation and offending amongst young people. Several large-scale surveys have concluded that those aged under 16 experience relatively high levels of criminal victimisation (19% as reported in the MOPAC Youth Matter Survey, 2015), often involving low-level violent offences or theft of personal property (e.g. ONS, 2014/2018; Wilson, Sharp & Patterson, 2006). However, research also suggests that some young people in this age group may also be more likely to offend, particularly for young males and once again in relation to theft and violent offences (Cooper & Roe, 2012; Youth Justice Board & Ministry of Justice, 2018). The ‘age-crime curve’ has been well-documented in academic literature, whereby a sharp increase in offending is seen during early adolescence, which peaks during the mid-late teenage years, and then begins to gradually decline (Farrington, 1986). Beyond this, previous research has also highlighted the importance of early engagement between the police and young people to help build positive relationships (All Parliamentary Group for Children, 2014; MOPAC, 2015). The Crime Survey for England and Wales Youth Module (ONS, 2014) reported that only half of young people nationally hold positive views of the police, with these views tending to become more negative with age. More specifically to London, the MOPAC Youth Matter Survey (2015) found that just 42% of young respondents were confident in police, which was 24 percentage points lower than the adult opinion as measured by the Public Attitude Survey at the time. Moreover, research has suggested that a lack of confidence in the police forms a key barrier that prevents young victims from seeking help from the police (ONS, 2014b; Victim Support, 2014), and has identified particularly high levels of under- reporting of crime amongst young people (ONS, 2014b). This picture is also set against the backdrop of changing crime trends that affect young people. A rise in knife offending has been seen across England and Wales since 2014, with London seeing a 22% increase in knife crime1 during FY 2017/18. Moreover, knife crime disproportionately affects young people in the capital, with around four in ten victims of knife crime resulting in injury2 aged under 25, and the number of young victims of knife crime with injury rising over recent years, from 1563 in 2014 to 2134 in 20173. Furthermore, young people are increasingly growing up in a world dependent on the internet and social media, bringing with it a range of additional safeguarding needs to help protect young people from online offences, including grooming and exploitation (HMIC, 2015). It is for all these reasons that one of the key commitments from the Mayor of London in the Police and Crime Plan 2017-2021 is to ‘keep children and young people safe’. This includes tackling the issues that disproportionately affect young people in London, including knife crime, gang-related crime, serious 1 MOPAC Weapon-enabled Crime Dashboard 2 Ibid 3 Data drawn from MetStats and relate to victims of non-DA Knife Crime with Injury offences where the victim was aged under- 25. 5 youth violence, exploitation and abuse. The Mayor of London’s (2017) Knife Crime Strategy recognises that knife crime disproportionately affects young Londoners, and emphasises the importance of including young people as part of the solution. This includes taking steps to divert those most at risk of offending and victimisation away from crime, and encouraging young people to reach their potential. Similarly, the Violence Against Women and Girls Strategy (Mayor of London, 2018) acknowledges that behaviours and beliefs that lead to sexual violence and harassment can often be manifested in early years, and emphasises the importance of working with young people to challenge these views. Together, this highlights the value of gathering young people’s opinions about crime and safety issues in London that affect them. This group represent the next generation to make the transition to adulthood in the city, and it therefore seems important to include their unique voice to help inform crime and policing decisions in the capital. However, despite this, existing surveys provide limited insight into the views of young people growing up in London. Routine surveys conducted by MOPAC (including the Public Attitude Survey and the User Satisfaction Survey) focus exclusively on gathering the views of adult residents (aged 16+). More widely, projects such as the Crime Survey for England and Wales (CSEW) Youth Module and the recent Safer Lives Survey conducted by The Youth Violence Commission (2018) provide interesting insights into young people’s experiences and attitudes towards crime and policing issues at a national level, but are not specific to those growing up in London. To help fill this gap in knowledge, MOPAC and the Metropolitan Police Service (MPS) have developed specific research projects to capture the views of young people growing up in the capital. Two previous waves of this project have been conducted: ‘Youth Talk’ in 2013 (11-16 year olds) and ‘Youth Matter’ in 2015 (11-18 year olds). The Youth Voice Survey 2018 represents the most recent wave of this project and gives young people the chance to have a say in how their city is policed. This report presents the findings from the Youth Voice Survey 2018, and is divided into four sections; each of which explores a different aspect of young people’s views of crime and policing issues. The first section, ‘Young People and Crime’ focuses on young Londoners’ feelings of safety at home and at school, and provides a snapshot of crime victimisation and reporting amongst this group. The second section, ‘Young People and the Police’ explores young people’s perceptions of the police, and examines engagement with Safer Schools Officers. The third section, ‘Young People and Serious Youth Violence’ measures exposure to gangs and knife crime amongst those growing up in the capital, and gathers reactions to the Mayor’s ‘London Needs You Alive’ anti-knife campaign. Finally, the fourth section ‘Young People and Safeguarding’ looks at online safety amongst young people, and explores early attitudes and opinions towards sexual harassment and healthy relationships. 6 2. Methodology 2.1. A London Youth Online Survey The Youth Voice Survey 2018 was conducted by the Mayor’s Office for Policing And Crime (MOPAC) in close collaboration with the Metropolitan Police Service. The survey was hosted online using a secure portal provided by Opinion Research Services (ORS), and was open between 7th March 2018 and 8th May 2018. The survey comprised a total of 50 questions, and took around 10 to 15 minutes to complete. Topics covered by the survey included: • Feelings of safety at home and at school • Experiences of crime victimisation • Satisfaction with the police for young victims of crime • Views and perceptions of the police • Views of Safer Schools Officers • Views and experiences of Stop and Search • Exposure to knife crime and gangs • Attitudes towards anti-knife crime campaigns (including London Needs You Alive) • Experiences of online safety (those in school years 10 and 11 only) • Experiences of sexual harassment (those in school years 10 and 11 only) The survey was distributed to schools in London via the Metropolitan Police Service’s Safer Schools Officers (dedicated police officers that work collaboratively with schools and educational establishments across the city). This method of distribution was selected to build upon existing partnership working between the Metropolitan Police Service and schools in London. A total of 7832 responses were received from young people aged 11 to 16 across London. The final sample contained roughly equal proportions of young males and females, and contained young people from a diverse range of ethnic backgrounds. Demographic breakdowns of these responses are shown in Appendix A Figure A1. Differences quoted in this research report are statistically significant at the P < 0.05 level. 2.2. Ethical Considerations Distributing the survey via schools provided an ethical safeguard for young people, by ensuring that consent from appropriate adults (teachers or parents) was obtained before participation. To retain confidentiality, the survey did not gather data that would identify either individuals or schools. Before taking part, young people were provided with information about the nature of the research, and were informed that their participation was voluntary. To aid this, young people were able to skip any questions they did not want to answer, while questions relating to more sensitive topics (e.g. crime victimisation, serious youth violence, sexual harassment and online behaviours) also contained explicit ‘prefer not to say’ options. Throughout the survey, young people were provided with links to relevant support organisations and charities as a safeguarding measure for any concerns around topics raised. 7 2.2. Limitations Sampling via Safer Schools Officers provided considerable benefits in terms of practicality and ethical considerations, and allowed the Youth Voice Survey 2018 to reach a wide audience of young people across the capital. However, this method may have also resulted in several limitations, which should be borne in mind when considering the results in this report. Although the final sample of 7,832 represents a strong capture of youth opinions in London, response rates do vary at a borough level, and this could affect the representativeness of results (see Appendix A Table A2 for borough breakdowns). Related to this, given the anonymity of the survey, it is difficult to estimate the number of schools which comprise the sample: pupils that attend a single school may be more likely to demonstrate cohesive views, particularly for questions that relate to school experiences (e.g. safety at school, presence of Safer Schools Officers etc.), and once again this may have impacted on results. Finally, as the survey was disseminated via schools, it is important to recognise that the survey may have missed the views of young people not in formal education, including home-schooled children, those who have been excluded, young offenders, and those who regularly truant or go missing. 8 3. Results 3.1. Young People and Crime The first section of this report explores young people’s experiences and exposure to crime issues in London. This includes overall feelings of safety and perceptions of crime problems in young people’s school/local area, alongside personal experiences of crime victimisation. 3.1.1. Perceptions of Safety Positively, results from the Youth Voice Survey 2018 indicate that the majority of young people feel safe both in the local area where they live (74%, 5787 of 7805) and at school (84%, 6502 of 7786). However, feelings of safety in the local area and at school tend to decrease with age. Illustrating this, those aged 11 years old (79%, 775 of 976) are 18 percentage points more likely to say they feel safe in their local area than those aged 16 years old (61%, 369 of 601) (see Figure 1).The trend is the same but less pronounced for feelings of safety at school, with a 7 percentage point difference between 11 year olds (87%, 845 of 972) and 16 year olds (80%, 484 of 602). 9 Figure 1: Feelings of safety in the local area and at school decrease with age. While the majority of all ethnic groups report feeling safe, young people from a Black background (76%, 941 of 1232) are significantly less likely to feel safe at school compared with young people from a White background (89%, 1885 of 2113). This effect is the same but less pronounced for feelings of safety in the local area, with young people from a Black background (72%, 891 of 1232) once again less likely to say they feel safe compared with those from a White background (77%, 1630 of 2116). As well as overall feelings of safety, the Youth Voice Survey 2018 also sought deeper opinions on a range of perceived problems either locally or at school (See Table 1). Table 1. Issues perceived as a ‘big problem’ in the local area and at school. Local Area School % Frequency % Frequency Violence 31% 2353 Violence 28% 2145 People joining gangs 30% 2320 Stealing 21% 1609 People carrying knives 29% 2238 Hate Crime 18% 1418 Stealing 25% 1947 People joining gangs 16% 1232 People using drugs 24% 1853 Sexist bullying 13% 967 Hate Crime 24% 1860 Sexual harassment 12% 945 People dealing drugs 20% 1552 People carrying knives 12% 936 Sexual harassment 19% 1433 People using drugs 11% 860 Sexist bullying 16% 1205 People dealing drugs 9% 710 Approximate base: Local Area (7700), School (7700) Two main observations can be made about the findings. Firstly, violence is the issue most often perceived as being a big problem in both environments, and to a similar degree. Secondly, for most other issues, fewer young people rate them as being a big problem at school in comparison to their local area. 10 While it is interesting to look at young people’s perceptions of safety in isolation, it is also possible to seek richer insights by looking at how feelings of safety are associated with other topics covered by the Youth Voice Survey 2018 (see Section 2.1 for more information about topics covered). Logistic regression analysis4 allows us to look at the risk-factors that may make a young person more likely to feel ‘unsafe’ in the area where they live and at school. Conversely, improvements in these factors are likely to beneficially impact on young people’s feelings of safety. When looking at the risk factors for feeling unsafe, results highlight a strong overlap between safety at school and safety in the local area. Feeling unsafe at school is one of the strongest predictors of feeling unsafe locally, and vice versa. This suggests these two environments are closely connected. Furthermore, results also reinforce the influence of young people’s perceptions of serious violence upon their feelings of safety. Illustrating this, feeling that ‘people joining gangs’ and ‘people carrying knives’ are a problem locally are both strong risk-factors for feeling unsafe in the local area, whilst feeling that ‘violence’ is a problem at school is also an important risk-factor for feeling unsafe at school. In particular, this supports the continued drive to tackle youth violence issues in London – although the vast majority of young Londoners feel safe, steps to reduce gang crime, knife crime and youth violence are likely to have wider beneficial impacts on young people’s feelings of safety. 3.1.2. Crime Victimisation Results from the Youth Voice Survey 2018 show that just over 1 in 10 young people have been the victim of a crime during the last year (12%, 851 of 7381). This suggests a level of self-reported victimisation roughly comparable with that seen amongst adults in London (11%, Public Attitude Survey 2017-18). Furthermore, these results are broadly in line with national levels seen for young people aged 10 to 15 (11%) (ONS, 2018b), supporting the validity of the Youth Voice Survey 2018. The types of crimes experienced by young people are shown in Box 1. Those who have been the victim of a crime in the last year most often said that they had experienced ‘theft or burglary’ (35%, 276 of 783) or ‘violence’ (19%, 146 of 783). This is in line with previous findings that identified notable proportions of young people feeling that ‘people being violent’ and ‘people stealing things’ are problematic issues in their school and local area. Analyses highlight a range of risk-factors that may make a young person more likely to become the victim of crime, including: • Having poorer levels of mental wellbeing5. • Knowing someone who is in a gang. • Having a bad opinion of the police. • Feeling unsafe at home. • Having personally carried a knife. 4 Logistic regression is an analytic method that looks at the effect of several predictor variables together upon an outcome. In this way, it is possible to understand the most important risk-factors that make an outcome more likely. 5 Levels of mental wellbeing were assessed by asking young people how often they felt happy, worried or anxious, stressed, lonely and like they are not good enough. An overall wellbeing score was then calculated by averaging responses across these scales. Box 1. What crime did you experience? (Only asked to victims, N = 783) Theft or Burglary (35%) Violence (19%) Hate Crime (12%) Online Crime (7%) Criminal Damage (4%) Something Else (20%) I don’t know (4%) 11 In particular, this appears to emphasise an overlap between early victimisation and exposure to other aspects of violence amongst young people in London. Illustrating this, those who ‘know someone who is in a gang’ or ‘have personally carried a knife’ are more likely to say they have been the victim of a crime in the last 12 months. Beyond this, it is also notable that young victims are more likely to experience certain other vulnerabilities, including poorer levels of ‘mental wellbeing’ and increased risk of feeling ‘unsafe at home’. This suggests that young victims of crime are also likely to be experiencing a range of other adversities, highlighting unique safeguarding needs amongst this group. In terms of reporting their crime, 44% (348 of 791) of young victims said they told the police. The most commonly selected reason for not reporting a crime to the police was because young people ‘didn’t feel the crime was important or serious enough’ (see box 2). However, results also suggest that poor perceptions or experiences of the police also form key barriers to reporting amongst young people. Illustrating this, feeling the police would not help them, not liking the police, and having bad previous experiences of the police all emerged as common reasons for young victims not reporting their crime. Furthermore, young people also express a hesitancy to talk to the police, with some being concerned about possible repercussions. Moreover, although base numbers are low, results suggest that poor perceptions of the police may influence the decision to report a crime to a greater extent amongst young people from BAME groups. BAME young people are significantly more likely than White young people to say they did not report because they ‘don’t like the police’ (26% (54 of 210) compared with 14% (14 of 102)), because they’ve ‘had a bad experience with the police before’ (17% (36 of 210) compared with 8%, (8 of 102)) and because they ‘didn’t think the police would help’ (37% (78 of 210) compared with 25% (26 of 102)). This is in line with wider findings from the Youth Voice Survey that perceptions of the police tend to be more negative amongst young people from certain minority ethnic groups (see section 3.2.1 of this report). Amongst young victims who did report their crime to the police, 29% (101 of 345) said they were happy with the way the police dealt with it, while 49% (168 of 345) were unhappy (remaining respondents were neutral or said they did not know). Although results are not directly comparable, this is notably below the levels of satisfaction with the police service seen for adult victims of crime as measured in the User Satisfaction Survey, which stood at 70% for FY 2017-18. This reinforces the importance of continuing to drive improvements to the service delivered by the police to ensure that young victims receive the support they need. Box 2. Why didn’t you report this crime to the police? (N = 416) I didn’t think it was important/serious enough (48%) I didn’t think the police would help me (32%) I dealt with it myself (31%) I didn’t want to be a grass or a snitch (27%) I didn’t feel comfortable talking to the police (23%). I don’t like the police (22%). I was worried about what would happen next (21%) I was worried about my friends/family finding out (17%) I didn’t think about reporting it (16%) I had a bad experience with the police before (14%) I felt frightened by the offender (13%) The offender was my friend/family (13%) 12 3.2. Young People and the Police Having looked at young people’s feelings of safety and experiences of crime in London, the second section of this report aims to further explore the relationship between young people and the police - this includes engagement with Safer Schools Officers and experiences of Stop and Search. 3.2.1. Perceptions of the Police When young people were asked about their ‘overall opinion’ of the police, 50% (3874 of 7815) said they have a ‘good opinion of the police’, while around 1 in 7 (14%, 1121 of 7815) said they have a ‘bad opinion of the police’. However, notable proportions of young people said that they either have ‘no opinion’ of the police (23%, 1767 of 7815) or that they ‘don’t know’ (13%, 1053 of 7815) in response to this question, highlighting a sizeable group of young people who may not have yet formed a strong opinion about the police. In line with findings from the CSEW (ONS 2014) and MOPAC’s (2015) Youth Matter Survey, results suggest that opinions of the police in London become more negative with age. For example, those aged 16 are three times more likely to say they have a ‘bad opinion’ of the police (21%, 127 of 600) than those aged 11 (7%, 67 of 978). Beyond their ‘overall opinion’, young people were also asked about their wider attitudes towards different aspects of the police’s role, with mixed opinions once again emerging (see Figure 2). 13 Figure 2: Young people have mixed wider perceptions of the police. Previously, we identified a satisfaction gap between young people and adults in London, and importantly these results highlight similar gaps in public perceptions of the police. Illustrating this, 31% of young people (2421 of 7799) agree the police do a ‘good job in their local area’, which is considerably lower than the result seen for adult residents at 67% (Public Attitude Survey, FY 17-18) and a decline on the result seen in the last MOPAC ‘Youth Matter’ Survey in 2015 (39%6). Analyses suggest that some of the most important risk-factors for young people having a ‘bad overall opinion’ of the police include: • Feeling the police do not do a good job in the local area. • Feeling the police do not treat everyone fairly. • Ethnicity (being from a Black or Mixed Ethnic Group compared with being from a White Ethnic Group). • Feeling the police cannot be relied on to be there when you need them. • Feeling the police are not helpful and friendly. • Having a negative experience of Stop and Search (compared with not being Stopped and Searched). In line with previous research into police legitimacy and procedural justice (e.g. Sunshine & Tyler, 2003; Tyler, 2003), these findings highlight the importance of young people’s views of police fairness and effectiveness in determining their wider opinions of the police. Most notably, perceptions of police unfairness emerge as a particularly strong predictor of young people having bad opinions. However, results 6 Changes to question wording mean that results from the ‘Youth Matter’ Survey 2015 are not directly comparable. Results relate to the proportion of young people saying the police do an ‘excellent’ or ‘good’ job in the local area. 14 from the Youth Voice Survey 2018 also appear to highlight issues with perceived police fairness amongst young people in London. Illustrating this, 37% of young people agree that the police ‘treat everyone fairly regardless of their skin colour or religion’ (2874 of 7740), with results for those from a white ethnic background (50%, 1061 of 2104) notably higher than those from a black ethnic background (20%, 241 of 1216). Together, these findings emphasise the importance of considering opportunities to help improve young people’s perceptions of police fairness in London, perhaps with particular emphasis upon engaging with those from minority ethnic backgrounds. Improvements in this area are also likely to result in more favourable overall opinions of the police amongst young people. Furthermore, these findings highlight the influence of Stop and Search upon young people’s opinions of the police. When young people reported positive Stop and Search encounters7, no adverse effect on young people’s overall opinion of the police was seen. In contrast, when young people reported negative Stop and Search encounters8, this effect was more drastic – young people were then notably more likely to report a bad overall opinion of the police. These findings suggest that it may not be the Stop and Search encounter per-se which can be detrimental to young people’s opinions of the police, but emphasise the importance of the way in which the interaction is carried out. If a young person feels they are treated well, this appears to buffer against potential negative impacts on wider opinions of the police. This is broadly in line with previous research that emphasises the importance of fair procedural interactions upon public confidence and police legitimacy (e.g. Sunshine & Tyler, 2003; Tyler, 2003), and the potentially disproportionate impact of negative interactions (e.g. Stanko et al. 2012, Bradford, Jackson & Stanko, 2009). 3.2.2. Stop and Search Given the influence of Stop and Search upon young people’s perceptions of the police as discussed above, it is an important finding that 44% of the survey cohort (3247 of 7358) agreed that the police should use Stop and Search, compared with 12% (895 of 7358) that disagreed. This demonstrates broad support for Stop and Search amongst young people, although this is lower than the proportion of adults in London who agree with the use of Stop and Search (77%)9. Furthermore, this is also a decrease when compared with MOPAC’s (2013) Youth Talk survey, where 56% of respondents said that Stop and Search should be allowed. While 38% (2834 of 7496) of young people say that they know someone that has been Stopped and Searched, 10% (749 of 7496) report that they have been subject to the procedure themselves. This figure may be higher than expected amongst those aged 11 to 16, and may in part reflect young people’s own understanding of what a ‘Stop and Search’ encounter is, or wider experiences of related procedures (such as ‘Stop and Account’). Despite this, it is still important to recognise that this proportion of young people feel or believe that they have been Stopped and Searched by police, regardless of it how it would be officially recorded. Previous research has shown that people are largely supportive of Stop and Search so long as it is conducted in a procedurally just manner (Singer, 2013), while section 3.2.1 of this report concludes that negative Stop and Search interactions are capable of having a detrimental impact on young people’s overall opinions of the police. The Youth Voice Survey 2018 highlights mixed experiences of Stop and 7 Positive Stop and Search encounters refer to those where the young person agrees that the police ‘were polite’, ‘treated them with respect’ AND ‘told them the reasons why they had been stopped. 8 Negative Stop and Search encounters refer to those where the young person does NOT agree with the above. 9 ‘Disproportionality: Justice Matters’ MOPAC, March 2018 15 Search, with 43% (312 of 719) of young people agreeing the police were polite during the process, 42% (296 of 707) agreeing the police treated them with respect, and 48% (341 of 712) saying the police explained why the procedure was being carried out. These figures bear some similarity to adult opinions – previous research shows that 36% say they were treated with respect, although a higher proportion (61%) say that they were told the reason why they were Stopped and Searched (HMIC and YouGov, 2013). Although base sizes are small, young people from a BAME background are significantly less likely to report positive Stop and Search experiences across all measures than young people from a White background (see Table 2). Table 2: Experiences of Stop and Search Experience, by Ethnicity. White BAME % Frequency % Frequency … the police were polite 58% 80 40% 160 …the police treated you with respect 62% 83 38% 151 …told you the reason you were stopped 61% 82 45% 180 Approximate base: White (135), BAME (400). All differences are statistically significant at the P < 0.05 level. 3.2.3. Safer Schools Officers Having looked at young people’s overall perceptions of the police in London, this next section specifically explores young people’s attitudes towards Safer Schools Officers. These are dedicated police officers from the Metropolitan Police Service that work with schools and pupil referral units. The Youth Voice Survey 2018 reveals mixed levels of awareness of Safer Schools Officers amongst young people. Around 6 in 10 young people say they are aware that their school has a Safer Schools Officer (57%, 4386 of 7706), 13% (977 of 7706) say they are ‘not aware’ and a further 30% (2343 of 7706) say they ‘do not know’. Despite this, for young people who are aware of their Safer Schools Officer, a notable proportion (43%, 1847 of 4327) say this officer makes them feel ‘more safe’ at school, while only a small minority say it makes them feel ‘less safe’ (2%, 85 of 4327). Similarly, the majority of young people say they would feel confident speaking to their Safer Schools Officer if a crime were to happen to them or they were worried about something (56%, 2409 of 4327), although 31% (1363 of 4327) say they would not feel confident doing this. However, analyses also highlight certain groups of young people for whom the positive impact of Safer Schools Officers may be less pronounced. In particular, those from older age groups and young people from Black, Mixed or Other Ethnic Backgrounds were less likely to say that having a Safer Schools Officer makes them feel ‘more safe’ at school, or to say they would feel confident speaking to this officer. Moreover, those who feel ‘unsafe’ at school are notably less likely to feel confident speaking to their Safer Schools Officer (28%, 137 of 491) than those who feel ‘safe’ at school (60%, 2229 of 3704). These findings seem to highlight a particularly vulnerable group of young people who may be experiencing issues at school, but feel unable to seek help from their Safer Schools Officers. This reinforces the importance of identifying and engaging with young people most at risk in school to ensure they feel confident to seek help. 16 Overall, results highlight a willingness amongst young people to have the police working together with their school. When asked to select the top three things the police should do in schools to help keep everyone safe (Box 3), young people tended to prioritise helping those experiencing difficulties and dealing with crime-related issues in school over more supportive tasks, such as giving assemblies or organising activities. However, young people who said they had been personally in a gang were significantly more likely to say they would like Safer Schools Officers to organise activities at school (37%, 81 of 217) than those not in a gang (30%, 1930 of 6509). It is worth noting that the fourth most commonly selected option is for police to ‘search school pupils to make sure they are not carrying drugs or knives’. This is perhaps surprising given the potentially intrusive nature of these searches and the previously highlighted relationship between negative police stop interactions and wider opinions of the police. It is possible this may be influenced by the fact that Safer Schools Officers are more likely to be known to young people. Those choosing searches as a priority also showed greater support for police searches in general, for example they were also more likely to agree that the police should conduct Stop and Search (57%, 1536 of 2714) than those not choosing this as a priority (37%, 1711 of 4644). Furthermore, those feeling ‘unsafe’ at school were more likely to feel the police should prioritise searching young people for drugs and weapons (40%, 410 of 1014)) than those feeling safe at school (35%, 2287 of 6502), with this forming the second-highest priority for this group. Finally, results from the Youth Voice Survey 2018 also suggest that Safer Schools Officers can have additional beneficial impacts beyond the school environment; in particular upon young people’s wider ‘overall opinions’ of the police. However, results suggest that building positive relationships between officers and pupils is critical to this. Illustrating this, those who are aware that their school has a Safer Schools Officer and who feel confident speaking to this officer are significantly less likely to have an overall ‘bad overall opinion’ of the police than those not aware of their Safer Schools Officer. In contrast, if a young person is aware of their Safer Schools Officer but does not feel confident speaking to them then this beneficial effect is no longer seen. This suggests it is not solely the presence of a Safer Schools Officer per se that can benefit young people’s wider opinions of the police, but instead reinforces the importance of engaging with pupils and ensuring this officer is approachable. Box 3. What are the top three things you think police officers should do in your school? (N = 7832) Help find the right support for young people who are having difficulties at home or school (42%). Investigate crimes that happen in school (41%). Deal with people who commit crime in school (40%). Search young people in school to make sure they are not carrying things like drugs or knives (36%). Make it easier for young people to contact the police and talk about problems or worries (35%). Give assemblies or lessons on crime and police issues (30%). Help to organise activities for young people at your school (e.g. school trips, or youth clubs) (29%). Spend time walking around to provide a visible police presence (23%). 17 3.3. Young People and Serious Violence This section is looks at young people’s experiences and exposure to serious violence in London, including knife crime and gangs. The section also explores young people’s reactions to anti-violence campaigns, including the Mayor’s ‘London Needs You Alive’ campaign. 3.3.1 Knife and Gang Crime Within the Youth Voice Survey 2018, young people were asked about their exposure to gangs and knife crime. A quarter of young people say that they know someone who has carried a knife (26%, 1790 of 6856) or who is in a gang (23%, 1583 of 6798). Certain groups of young people appear to show greater vicarious exposure to these issues, including young victims of crime and those who have attended a Pupil Referral Unit (PRU). For example, 54% of young victims of crime (396 of 738) know someone that has carried a knife compared with 21% of non-victims (1082 of 5210). Similarly, 53% of young victims know someone in a gang (382 of 724), compared to 17% of non-victims (904 of 5184). When looking at PRU attendees, 47% (92 of 196) say they know someone who has carried a knife with them, compared with 25% of non-PRU attendees (1188 of 4673). Once again, it is a similar picture for exposure to gangs, with 46% of PRU attendees saying they know someone in a gang (87 of 191) compared with 22% of non-PRU attendees (1022 of 4585). The proportion of young people responding to the survey who claim to have carried a knife (218 of 7033) or to have been in a gang themselves (217 of 7056) is 3% in each case. 38% (76 of 198) of those saying they have personally been in a gang also say that they have carried a knife. The proportion here saying 18 that they have carried a knife is lower than that shown by other research10, but in each case there are substantial methodological differences. Analyses suggest that some of the most important risk-factors in determining whether a young person has ever personally carried a knife include: • Feeling pressured to carry a knife. • Being personally in a gang. • Having been the victim of a crime. • Knowing people who have been ‘Stopped and Searched’. • Feeling people carrying knives is a big problem at school. This suggests that personal involvement in gangs remains an important risk-factor for knife-carrying. Beyond this, however, results appear to highlight the influence of wider vicarious exposure to knife issues upon young people’s behaviour, with feeling ‘pressured to carry a knife’ and believing ‘people carrying knives is a big problem in school’ also emerging as important risk-factors. More widely, it is also worth noting that being the victim of a crime also emerges as an important risk-factor, perhaps reinforcing an overlap between victimisation and offending amongst young people. The Youth Voice Survey 2018 also provided a platform for young people to give their views about knife crime, including around why some young people may carry knives and what the police can do to help tackle knife crime (see Figure 3). Figure 3: Example comments from young people around knife crime in London. Almost half of those asked why young people carried knives referred to doing so for their own protection. This was often linked to wider feelings of safety, with some comments highlighting a general fear about crime in London (including murder, kidnapping, sexual harassment, stealing and violence) among young people. Other themes also emerged, voiced by smaller proportions, including to specifically commit crime, harm and intimidate others, the influence of gangs, the desire to “look cool”, “hard” or “gain respect”, the ineffectiveness of police to protect them, or people’s own poor emotional wellbeing. 10 6% in the CSEW; 30% in Scottish Centre for Crime and Justice Research, 2010 19 For those asked about what can stop people from carrying knives, the most common theme (occurring in about 30% of comments) to emerge regarded police action. This included increased police numbers and patrolling and also the Stop and Search of people suspected of carrying knives. The views expressed about police action are consistent with data gleaned from elsewhere in the survey, with 58% of young people (4233 of 7337) agreeing that Stop and Search would stop people carrying knives. However, deeper analysis revealed significant factors related to those that disagreed with this view: being older, being from a black ethnic group, having a bad opinion of the police, and knowing someone that is in a gang. This in turn echoes previous findings from MOPAC’s Knife Crime Strategy Consultation in June 2017. Another common theme regarded education about the realities and consequences of carrying knives. The nature of the education that young people spoke of was multi-agency, involving schools, teachers, the police, parents, charities, previous offenders and victims. This commonly overlapped with comments about wider communication and understanding on behalf of all the different agencies referred to. The young people expressed a desire and stressed the importance for such agencies to engage with their own situations, circumstances and points of view. This highlights positive opportunities for multi-agency work to help educate young people on the risks and impacts of knife crime. 3.3.2. London Needs You Alive One section of the Youth Voice 2018 survey measured exposure and attitudinal responses to the London Needs You Alive campaign. This campaign was run by the Mayor of London with the aim of reducing the number of young people carrying knives in the capital by reinforcing their sense of worth. The campaign aimed to reach young people who are at risk of knife violence, with a particular focus on those aged 13 to 16 in boroughs with high levels of knife crime. Several strands of communication were used, including social media, online video platforms, street advertising and an educational toolkit. At the time of the Youth Voice Survey 2018 (7th March 2018 to 8th May 2018), around 1 in 6 young people said they had heard about the London Needs You Alive campaign (17%, 1190 of 6975). In general, results support the campaign’s focus on young people aged 13 to 16 and those most affected by knife crime. Illustrating this, awareness of the campaign increased with age, from 12% of 11-year-olds (107 of 869) to 22% of 16-year-olds (114 of 527). Similarly, those saying they had personally carried a knife were also more likely to be aware of the campaign (26%, 53 of 206) than those saying they had not carried a knife (17%, 1081 of 6418). Furthermore, results highlight the success of certain key media channels, including those central to the London Needs You Alive campaign such as Instagram, online video advertising and street advertising (see Box 411). Young people supported the use of incentives to encourage engagement with campaigns: over half said they would be more likely to get involved with the London Needs You Alive campaign if they had the chance to win ‘a holiday or day out’ (62%, 645 of 1033), ‘clothes or vouchers’ (54%, 559 of 1031) or the ‘opportunity to record a music track or spend the day in the studio with a famous artist (52%, 545 of 1042). 11Respondents could choose more than one channel. Only asked to those aware of the campaign. Box 4. Where did you hear about the campaign? (Most commonly selected answers, N =1190) Instagram (42%) Video advert on the internet (34%) Posters or outdoor adverts (27%) Television (24%) School lessons or assemblies (24%) 20 Table 3: Perceptions of the London Needs You Alive campaign. % Frequency The things I saw gave a positive message to young people in London. 59% 640 The things I saw made me think more about what can happen when people carry knives. 53% 580 The things I saw made me think more about what I can achieve and why my life is important. 51% 558 The things I saw worked well for people my age. 39% 425 The things I saw made me feel more worried about knife crime in London 22% 250 Approximate base: 1100 When asked about their wider attitudes to the campaign (see Table 3), over half of young people seeing the campaign felt it gave a positive message, made them think more about the risks of carrying knives, and made them think more about why their life is important. In contrast, a minority of young people disagreed with the attitudinal questions (maximum 11%), with remaining respondents either ‘neutral’ or saying they ‘don’t know’. Findings therefore appear to show general support for the London Needs You Alive campaign, with many young people holding positive attitudes, but nevertheless also highlight a sizeable group of young people who remain either neutral or unsure about their attitudes towards the campaign. Moreover, results suggest that young people’s perceptions of the London Needs You Alive campaign may be less positive for those already exposed to certain crime and safety issues. For example, young people saying they know someone who carries a knife were more likely to disagree that the campaign ‘made them think about why their life is important’ (14%, 44 of 309) than those saying they do not know anyone who carries a knife (8%, 49 of 607). Furthermore, young people already feeling unsafe at school or in their local area were more likely to say the things they saw made them ‘more worried about knife crime in London’ than those feeling safe, perhaps suggesting that in some cases the campaign may have exacerbated concerns amongst those already worried about their safety. 3.3.3. Knife Imagery in Media Campaigns Within the Youth Voice Survey 2018, young people were shown an image taken from an anti-knife campaign developed by the Metropolitan Police Service. The original image, which had been used as part of this campaign, contained a knife covered in blood – the image was subsequently edited to remove this knife, creating two identical campaign images except for the presence or absence of the knife (Appendix B). Young people completing the survey were randomly shown one of these two images and were asked a series of attitudinal statements relating to the picture they saw (Table 4). In this way, it was possible to isolate the effect of including knife imagery within anti-knife media campaigns upon young audiences. 21 Table 4: Proportion agreeing with statements, by whether or not the image shown contained a knife. Knife No Knife % Frequency % Frequency I would remember this picture 52% 820 51% 811 This picture would get my attention* 65% 1027 61% 972 This picture would make me want to know more 35% 539 33% 513 This picture would make me feel scared* 34% 530 30% 473 The picture would make me feel more worried about knife crime* 54% 873 50% 809 *Statistically significant differences are seen in the proportion agreeing at P < 0.05 for questions marked with a (*). Approximate base: Knife (1550), No Knife (1580) The effect of knife imagery upon young people appears to be complex. For example, young people seeing campaign imagery with a knife were more likely to say the picture got their attention (65% compared with 61%), but were also more likely to say the picture made them feel scared (34% compared with 30%) or more worried about knife crime in London (54% compared with 50%). Furthermore, the effect of knife imagery upon young people’s levels of fear and worry appears to be stronger amongst those already concerned about knife crime. For example, when looking at those who feel knife crime is a ‘big problem’ in the area where they live, the proportion saying they would feel scared as a result of seeing the image is notably higher for those seeing the image containing a knife (40%, 174 of 437) compared with those seeing the image without a knife (29%, 117 of 401). In contrast, this effect is not seen amongst those feeling knife crime is ‘not a problem’ in their local area, where the proportion feeling scared remains at 33% (115 of 345) regardless of the condition. For those already worried about knife crime then, viewing knife imagery may further exacerbate this fear. This effect is important to consider when designing anti-knife crime campaigns; perhaps even more so when campaign materials aim to target young people living in areas most affected by knife crime. 22 3.4. Young People and Safeguarding The final section of this report seeks to explore young people’s experiences of online safety and inappropriate content, and to measure early experiences and attitudes towards sexual harassment and healthy relationships. Given the sensitive nature of these topics, these questions were only asked to young people in school years 10 and 11 (aged 14 to 16), with a total base of 1752 respondents. 3.4.1. Online Safety Young people in years 10 and 11 were asked whether they had experienced any of a set of online or social media risks (see Table 5). Over half of respondents (56%, 787 of 1396) said they had experienced someone that they did not know try to ‘add’ or speak to them online, while approximately a quarter had either had been bullied online (25%, 354 of 1389) or been sent rude or sexual content (23%, 304 of 1346). 8% of young people (114 of 1387) said they had been threatened or physically hurt in real life as a result of online activity. 23 Table 5: Proportion experiencing online risks in the last 12 months. % Frequency Someone you didn’t know tried to add you or speak to you online 56% 787 Someone said mean things to you or bullied you online 25% 354 Someone sent you rude or sexual content 23% 304 Someone tried to persuade you to do something that you did not want to do 16% 220 Someone shared embarrassing pictures of you, or used your picture in a way you didn’t want 11% 152 Someone threatened or physically hurt you in real life as a result of something that happened online, or something that was posted on social media 8% 114 Approximate Base: 1300 A larger proportion of female respondents are affected than males in almost every case. With the exception of ‘sharing embarrassing pictures’ or ‘being threatened or hurt in real life as a result of something that happened online’, each of these differences are statistically significant (see Appendix C Table C1). Young people who feel unsafe in their local area (37%, 139 of 376) are more likely to say they have been bullied online than those feeling safe in their local area (22%, 208 of 958). This effect is also seen for feeling safe at school, with 39% of those feeling unsafe (69 of 176) saying they have been bullied online compared with 24% of those feeling safe at school (269 of 1143). As well as things that they have personally experienced online, the survey also asked young people about things they may have seen online (See Table 6). Table 6: Proportion of young people seeing potentially inappropriate content in the last year. % Frequency Violent content (e.g. pictures/videos showing fights/weapons) 53% 734 Racist content, or content that promotes hate or discrimination 47% 650 Gang-related content 38% 518 Approximate base: 1380 47% of young people have seen racist content, or content that promotes hate or discrimination in the last year. There are also safeguarding considerations for those who have seen gang-related content online. Those young people that have seen such content are more likely to feel unsafe, to be victims of crime, and to know people who are in gangs when compared to those who have not (see Table 7). 24 Table 7. Safeguarding Considerations and Online Gang Related Content Seen Gang Content Not Seen Gang Content % Frequency % Frequency Feel unsafe in local area 42% 215 19% 139 Feel unsafe at school 16% 85 10% 72 Victim of crime in last 12 months 25% 121 10% 71 Know someone in a gang 55% 265 17% 117 Know someone that has carried a knife 58% 282 24% 169 Approximate base: Seen Gang Content (500), Not Seen Gang Content (710). All differences are statistically significant at the P < 0.05 level. 3.4.2. Sexual Harassment and Healthy Relationships As part of the Youth Voice Survey 2018, young people in school years 10 and 11 were shown examples of scenarios that could be considered as sexual harassment or domestic abuse, and were asked whether it was ‘always’, ‘sometimes’ or ‘never’ ok for people to do these things (see Table 8). Table 8: Proportion of young people thinking it is acceptable (always ok or sometimes ok) to… Always/ Sometimes OK Frequency Check your partner’s phone or social media to see that they’ve been up to or who they’ve been talking to 47% 642 Stare or wolf-whistle at people you fancy as they walk past. 31% 432 Tell your partner not to hang out with some of their friends because you don’t like them 31% 429 Insult your partner during an argument 24% 331 Make sexual comments or jokes about people 23% 318 Try to dance with someone you fancy, even if they don’t want to dance with you 23% 316 Hit or push your partner during an argument 6% 86 Approximate base: 1370 Around half of young people (47%, 642 of 1372) feel it is ok for someone to check their partner’s phone or social media to see what they’ve been doing or who they’ve been talking to, while a third (31%, 429 of 1367) feel it is ok for someone to tell their partner not to see certain friends. For each of the behaviours listed, young people are more likely to feel they are ‘sometimes acceptable’ than ‘always acceptable’ (see appendix C Table C2). 25 Although far fewer feel it is ok to hit or push a partner (6%, 86 of 1371), this still equates to around 1 in 20 young people who feel physical violence may be acceptable. Analyses suggest that some of the most important risk factors for feeling that ‘hitting or pushing a partner’ is acceptable include: • Being in a gang. • Feeling the police can NOT be relied on to be there when needed. • Feeling violence is a problem in the local area. This appears to highlight an overlap between feeling domestic violence is acceptable and exposure to other aspects of serious youth violence, including ‘being in a gang’ and ‘feeling violence is a problem in the local area’. This could suggest that wider exposure to other violence-related issues may be important in determining young people’s more specific attitudes towards domestic violence. Closely related to this, believing the police cannot be relied on to be there when needed’ also emerges as one of the strongest predictors of feeling it is acceptable to hit or push a partner. Previous research has shown the importance of perceived police legitimacy upon people’s willingness to obey wider laws (Tyler, 2006). In line with this, these findings could suggest that for some young people an underlying lack of trust in the police may be associated with early views around the acceptability of offending behaviours. When looking at the acceptability of domestic abuse and sexual harassment behaviours by gender, young females tend to be more likely than young males to feel these behaviours are ‘never’ ok. In contrast, young males are around twice as likely to answer ‘I don’t know’ than females. For example, while 6% of young females (37 of 651) say they do not know whether it’s acceptable for someone to hit their partner, this figure stands at 14% (86 of 603) for young males. This could highlight a greater level of uncertainty amongst young males as to the extent to which sexual harassment or domestic abuse may be considered acceptable, and reinforces the importance of engaging with young people at an early age to discuss these important issues and to help lay the foundations for healthy relationships into adulthood. Beyond this, young people were also asked about their own personal experiences of sexual harassment, with results suggesting that these experiences are relatively common (See table 9). Table 9: Proportion who have experienced sexual harassment in the last year. % Frequency Unwanted staring 48% 639 Wolf-whistling 24% 321 Unwanted sexual comments 22% 294 Jokes or taunts of a sexual nature 22% 286 Approximate base: 1320. When looking at this in more detail, young females are notably more likely to experience sexual harassment than young males. This means the picture of harassment amongst young females becomes even more stark: over two-thirds of girls in school years 10 and 11 say they have experienced unwanted staring over the last year (68%, 435 of 642), while around a third have experienced jokes or taunts of a sexual nature (30%, 191 of 634) or sexual comments (34%, 218 of 639). 26 22% of young people experiencing sexual harassment said they spoke to someone about it or sought help (149 of 679). Moreover, when help was sought, the clear majority said they spoke to their family or friends (90%, 134 of 149); in contrast, very few young people said they spoke to the police (7%, 10 of 149) or to a support organisation/charity (5%, 7 of 149). For those who chose not to speak to anyone about their experiences, normalisation of sexual harassment behaviours again emerges as a key barrier to seeking help. Many young people said they did not seek help because they didn’t feel the incident was ‘important or serious enough’, ‘it happens too often’, or because ‘it’s normal/just banter’ (see box 5). This was also reflected in young people’s comments: Finally, it is worth noting that around 1 in 10 young people who did not talk about their experiences of sexual harassment said this was because they felt ashamed, or did not know who to speak to. This could reinforce the importance of ensuring young people feel empowered to seek help and are aware of available channels for support. Box 5. Why didn’t you speak to anyone about your experience of sexual harassment? (Most frequently chosen answers N = 482) I didn’t feel I needed to (66%) I didn’t think it was important or serious enough (43%) It’s normal/it was just banter (27%) It happens too often (18%) I didn’t think about talking to anyone (16%) I didn’t know who did it (14%) I felt ashamed, or didn’t feel comfortable talking about it (13%) I didn’t know who to speak to (11%) 27 4. Conclusion In conclusion, the Youth Voice Survey 2018 has gathered a wealth of data, offering valuable insights into young people’s experiences and perceptions of crime and safety issues across the capital. The survey is a rich source of information, providing young people with a voice and a platform for sharing their views. Overall, there are some reassuring findings that have emerged from the survey. Most notably, it is positive that the clear majority of young people say they feel safe both in the area where they live and at school. Furthermore, results suggest potentially beneficial impacts of Safer Schools Officers upon young people’s feelings of safety at school, and it is encouraging that the majority of young Londoners broadly support the use of Stop and Search tactics to help tackle knife crime in the capital. Despite this, results from the survey also highlight some areas for attention. Analyses reveal relatively high levels of under-reporting of crime amongst young people in London, despite victimisation appearing to be associated with multiple vulnerabilities such as exposure to youth violence issues and wider poor wellbeing. Moreover, less than a third of those who do report their crimes to the police say they are satisfied with the service provided to them, which is notably lower than victim satisfaction rates seen amongst adults in London. In addition to this, results highlight the prevalence of exposure to gang involvement and knife possession amongst those growing up in the capital, and support commitments to encourage positive attitudes and behaviours that underlie healthy relationships amongst young people. However, there are some clear opportunities arising from the survey, both for MOPAC and the Metropolitan Police, and for wider partnership working. When looking across all the topics covered by the Youth Voice Survey 201812, results highlight a complex interplay between crime and safety issues (see Table 10). This suggests that steps to improve young people’s perceptions or experiences of these issues are likely to have wider benefits across a range of inter-related outcomes. In particular, topics associated with serious youth violence (‘knife crime’ and ‘gangs’) serve as risk-factors for several outcomes, including wider feelings of safety, young people’s attitudes towards the police, experiences of criminal victimisation, and personal offending behaviours (e.g. knife possession). This could suggest that exposure to aspects of serious youth violence are particularly important in shaping young people’s wider attitudes towards a range of crime and safety issues, and supports the continued drive to tackle knife and gang violence in the capital. Within the Youth Voice Survey 2018, young people expressed a desire for more interventions that raise their awareness of the potential consequences of carrying weapons, and suggested that this could be most effective if delivered in partnership between different organisations, including the police, schools, previous offenders and victims. Furthermore, results highlight the important role that the police play in a range of outcomes, with ‘opinion of the police’ associated with several measures including feelings of safety and experiences of victimisation. When considered alongside the finding that sizeable proportions of young people may not yet have formed strong opinions of the police, this also highlights a valuable opportunity to help build early positive relationships with young Londoners. Finally, results further reinforce the overlap between feelings of safety, crime victimisation, and wider wellbeing issues. This highlights the importance of ensuring that the police and partnership organisations 12 The relationships between different topics covered in the survey were explored using several logistic regression models on a range of outcome variables. 28 across London recognise this complex picture of vulnerability, and provide appropriate support to safeguard young people. The next steps are to share findings from the Youth Voice Survey 2018 more widely to provide an evidence base and catalyst for continuous improvements to policy, commissioning and delivery of services to help keep young people safe in London. 29 *Outcomes are listed vertically in the table, while risk factors are displayed horizontally. Important relationships between these are highlighted in interconnecting cells. Outcome Safety Opinion of police Crime victimisation Safer Schools Officers Knife Crime Gangs Stop and Search Mental Wellbeing Feeling unsafe in local area Those who feel unsafe at school are more likely to feel unsafe in the local area. Those who feel the police do a bad job locally are more likely to feel unsafe in the local area. Those who feel knife crime is a problem locally are more likely to feel unsafe in the local area. Those who feel gangs are a problem locally are more likely to feel unsafe in the local area. Those with poorer mental wellbeing are more likely to feel unsafe in the local area. Feeling unsafe at school Those who feel unsafe at school, or feel violence is a problem at school, are more likely to feel unsafe at school. Those who have a bad opinion of the police are more likely to feel unsafe at school. Young victims of crime are more likely to feel unsafe at school. Those who know someone who carries a knife are more likely to feel unsafe at school. Those who have personally been in a gang are more likely to feel unsafe at school. Those with poorer mental wellbeing are more likely to feel unsafe at school. Having a bad opinion of the police Those who feel unsafe at school are more likely to have a bad opinion of the police. Those with wider negative perceptions of the police (e.g. unfairness) are more likely to have a bad opinion of the police. Young victims of crime are more likely to have a bad opinion of the police. Those without an SSO, or who have an SSO but don’t feel confident talking to them, are more likely to have a bad opinion of the police. Those who know someone who carries a knife are more likely to have a bad opinion of the police. Those who have had a negative ‘Stop and Search’ encounter are more likely to have a bad opinion of the police. Being a young victim of crime Those who feel unsafe at home are more likely to have been a victim of crime. Those who have a bad opinion of the police are more likely to have been a victim of crime. Those who have personally carried a knife, or know someone who has, are more likely to have been a victim of crime. Those who know someone in a gang are more likely to have been a victim of crime. Those who have been ‘Stopped and Searched’ are more likely to have been a victim of crime. Those with poorer mental wellbeing are more likely to have been a victim of crime. Having personally carried a knife Young victims of crime are more likely to have carried a knife. Those who have felt pressured to carry, know others who carry, and feel knives are a problem at school are more likely to have carried a knife. Those in a gang are more likely to have carried a knife. Those who know someone who has been ‘Stopped and Searched’ are more likely to have carried a knife. Being involved in a gang Those who feel unsafe at school are more likely to have been in a gang. Those who have carried a knife are more likely to have been in a gang. Those who know other people in gangs are more likely to have been in a gang. Those who have been ‘Stopped and Searched’ are more likely to have been in a gang. Table 10: Summary of the most important risk-factors from different areas of the Youth Voice Survey 2018, and their relationships with key outcomes. 30 References: All Parliamentary Group for Children (2014). It’s all about trust: building good relationships between children and the police. Retrieved from: https://www.ncb.org.uk/sites/default/files/uploads/documents/Policy_docs/appgc_children_and_pol ice_report_-_final.pdf Bradford, B., Jackson, J., & Stanko, E. (2009). Contact and confidence: revisiting the impact of public encounters with the police. Policing and society, 19(1), pp. 20-46. Cooper, C. & Roe, S. (2012). An estimate of youth crime in England and Wales: police recorded crime committed by young people in 2009/10. Research Report 64. Retrieved from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/ 167982/horr64.pdf Dawson, P. (2018). Key trends and learning about knife crime. Paper presented at the Knife Crime Summit, London, UK. Farrington, D. (1986). Age and Crime. In Tonroy, M. & Morris, N. (Eds). Crime and justice: an annual review of research Vol. 7 pp. 189-250. HMIC (2015). Online and on the edge: real risks in a virtual world. An investigation into how forces deal with the online sexual exploitation of children. Retrieved from: https://www.justiceinspectorates.gov.uk/hmicfrs/wp-content/uploads/online-and-on-the-edge.pdf HMIC & YouGov (2013). Stop and Search Report. Retrieved from: https://www.justiceinspectorates.gov.uk/hmicfrs/media/stop-and-search-survey-summary.pdf. Local Government Association (2018). The relationship between family violence and youth offending. Local Government Association: London. MOPAC (2015). Youth Matter: listening to the voice of young London. Retrieved from: https://www.london.gov.uk/sites/default/files/youth_matter_report_final_version.pdf ONS (2014). Chapter 2: 10 to 15 year olds’ perceptions of the police. Retrieved from: http://webarchive.nationalarchives.gov.uk/20160109023108/http://www.ons.gov.uk/ons/dcp171776 _365182.pdf ONS (2014b). Chapter 3: Experiences of crime among 10 to 15 year olds. Retrieved from: http://webarchive.nationalarchives.gov.uk/20160109023849/http://www.ons.gov.uk/ons/dcp171776 _365203.pdf ONS (2018). Subnational population projections for England: 2016-based. Retrieved from: https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationprojectio ns/bulletins/subnationalpopulationprojectionsforengland/2016based ONS (2018b). Statistical bulletin: Crime in England and Wales: year ending December 2017. Retrieved from: https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/crimeinenglandan dwales/yearendingdecember2017#main-points 31 Singer, L. (2013). London Riots: Searching for a Stop. Policing: a Journal of Policy and Practice, 7(1), 32- 41. Stanko, E., Jackson, J., Bradford, B., and Hohl, K. (2012) A golden thread, a presence amongst uniforms, and a good deal of data: studying public confidence in the London Metropolitan Police. Policing and Society: An International Journal of Research and Policy, 22:3, 317-331. The Mayor of London (2018). A safer city for women and girls: the London tackling violence against women and girls strategy 2018-2021. Retrieved from: https://www.london.gov.uk/sites/default/files/vawg_strategy_2018-21.pdf The Mayor of London (2017). The London knife crime strategy. Retrieved from: https://www.london.gov.uk/sites/default/files/mopac_knife_crime_strategy_june_2017.pdf The Youth Violence Commission (2018). The Youth Violence Commission Interim Report – July 2018. Retrieved from: http://yvcommission.com/literature-reports/ Sunshine, J. & Tyler, T. (2003). The Role of Procedural Justice and Legitimacy in shaping public support for Policing. Law & Society Review, 37(3), pp. 513-548. Tyler, T. (1990). Why do people obey the law? New Haven and London: Yale University Press. Tyler, T. (2003). Procedural justice, legitimacy, and the effective rule of law. Crime and Justice, 30, pp. 283-357. Victim Support (2014). Suffering in silence: children and unreported crime. Retrieved from: https://www.victimsupport.org.uk/sites/default/files/Suffering%20in%20silence%20- %20Children%20and%20unreported%20crime.pdf Youth Justice Board & Ministry of Justice (2018). Youth Justice Statistics 2016/17: England and Wales: Statistics Bulletin. Retrieved from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/ 676072/youth_justice_statistics_2016-17.pdf Wilson, D., Sharp, C. & Patterson, A. (2006). Young people and crime: findings from the 2005 Offending, Crime and Justice Survey. Home Office Statistical Bulletin. Retrieved from: http://dera.ioe.ac.uk/6902/1/hosb1706.pdf 32 Appendix A: Demographics Figure A1: Demographic breakdowns for respondents. 33 Table A2: Borough distributions of respondents. Live Go to School % Frequency % Frequency Barking and Dagenham 13% 1010 13% 1016 Barnet 5% 371 5% 406 Bexley 2% 141 2% 140 Brent 2% 119 0% 11 Bromley 1% 105 1% 99 Camden 5% 423 9% 666 City of London 2% 152 1% 82 Croydon 3% 220 3% 255 Ealing 5% 382 5% 369 Enfield 2% 126 1% 95 Greenwich 5% 391 5% 414 Hackney 3% 224 3% 267 Hammersmith and Fulham 2% 190 3% 272 Haringey 1% 80 0% 32 Harrow 2% 170 2% 145 Havering 7% 550 7% 550 Hillingdon 2% 188 2% 185 Hounslow 2% 120 1% 51 Islington 2% 185 0% 20 Kensington and Chelsea 1% 66 0% 15 Kingston upon Thames 2% 119 1% 106 Lambeth 3% 256 3% 198 Lewisham 1% 107 0% 36 Merton 1% 74 2% 154 Newham 1% 41 0% 5 Redbridge 3% 233 3% 257 Richmond upon Thames 3% 236 3% 249 Southwark 3% 245 4% 312 Sutton 0% 14 0% 1 Tower Hamlets 3% 235 3% 253 Waltham Forest 4% 275 4% 278 Wandsworth 3% 257 3% 265 Westminster 3% 238 4% 337 I don’t know 3% 228 3% 253 Outside of London 1% 61 0% 38 Total 100% 7832 100% 7832 34 Appendix B: Campaign Images With (Picture 1) and Without (Picture 2) Knife Imagery. 35 Appendix C: Safeguarding Table C1: Online safeguarding experiences. *Statistically significant differences are seen between males and females at P < 0.05 for questions marked with a (*). Approximate base: Male (590), Female (630) In the last 12 months has someone… Male Frequency Female Frequency …you didn’t know tried to add you or speak to you online* 49% 295 65% 415 …said mean things to you or bullied you online* 21% 127 28% 180 …sent you rude or sexual content* 16% 91 28% 176 …tried to persuade you to do something that you did not want to do* 13% 75 19% 118 …shared embarrassing pictures of you, or used your picture in a way you didn’t want 10% 61 11% 69 …threatened or physically hurt you in real life as a result of something that happened online, or something that was posted on social media 8% 48 8% 51 36 Table C2. Acceptability of sexual harassment and domestic abuse behaviours: full results. Always OK Frequency Sometimes OK Frequency Never OK Frequency I don’t know Frequency Checking your partner’s phone or social media to see that they’ve been up to or who they’ve been talking to 9% 129 37% 513 38% 522 15% 208 Telling your partner not to hang out with some of their friends because you don’t like them 5% 68 26% 361 54% 732 15% 206 Staring or wolf-whistling at people you fancy as they walk past. 5% 66 27% 366 51% 701 18% 243 Trying to dance with someone you fancy, even if they don’t want to dance with you 4% 57 19% 259 60% 819 17% 238 Insulting your partner during an argument 4% 52 20% 279 61% 834 15% 208 Making sexual comments or jokes about people 4% 50 19% 268 63% 863 14% 195 Hitting or pushing your partner during an argument 2% 31 4% 55 83% 1135 11% 150 Approximate base: 1370 
\ No newline at end of file